From 35748654f39a99c226cd14f3b92822eb64bd7037 Mon Sep 17 00:00:00 2001
From: Matthias Schiffer <mschiffer@universe-factory.net>
Date: Sun, 1 Dec 2013 07:40:27 +0100
Subject: Add fast SSE2 implementations "xmm" for Salsa20 and Salsa20/12

Public Domain implementations by D. J. Bernstein, see
http://cr.yp.to/snuffle.html
---
 CMakeLists.txt                                     |    2 +-
 src/cpuid.h                                        |    1 +
 src/crypto/cipher/salsa20/CMakeLists.txt           |    1 +
 src/crypto/cipher/salsa20/xmm/CMakeLists.txt       |   13 +
 src/crypto/cipher/salsa20/xmm/salsa20_amd64_xmm6.s | 4823 +++++++++++++++++++
 src/crypto/cipher/salsa20/xmm/salsa20_x86_xmm5.s   | 5078 ++++++++++++++++++++
 src/crypto/cipher/salsa20/xmm/salsa20_xmm.c        |   84 +
 src/crypto/cipher/salsa2012/CMakeLists.txt         |    1 +
 src/crypto/cipher/salsa2012/xmm/CMakeLists.txt     |   13 +
 .../cipher/salsa2012/xmm/salsa2012_amd64_xmm6.s    | 4823 +++++++++++++++++++
 .../cipher/salsa2012/xmm/salsa2012_x86_xmm5.s      | 5078 ++++++++++++++++++++
 src/crypto/cipher/salsa2012/xmm/salsa2012_xmm.c    |   84 +
 12 files changed, 20000 insertions(+), 1 deletion(-)
 create mode 100644 src/crypto/cipher/salsa20/xmm/CMakeLists.txt
 create mode 100644 src/crypto/cipher/salsa20/xmm/salsa20_amd64_xmm6.s
 create mode 100644 src/crypto/cipher/salsa20/xmm/salsa20_x86_xmm5.s
 create mode 100644 src/crypto/cipher/salsa20/xmm/salsa20_xmm.c
 create mode 100644 src/crypto/cipher/salsa2012/xmm/CMakeLists.txt
 create mode 100644 src/crypto/cipher/salsa2012/xmm/salsa2012_amd64_xmm6.s
 create mode 100644 src/crypto/cipher/salsa2012/xmm/salsa2012_x86_xmm5.s
 create mode 100644 src/crypto/cipher/salsa2012/xmm/salsa2012_xmm.c

diff --git a/CMakeLists.txt b/CMakeLists.txt
index 4e06cff..d154d34 100644
--- a/CMakeLists.txt
+++ b/CMakeLists.txt
@@ -1,5 +1,5 @@
 cmake_minimum_required(VERSION 2.8.3)
-project(FASTD C)
+project(FASTD C ASM)
 
 set(CMAKE_MODULE_PATH ${FASTD_SOURCE_DIR}/cmake)
 set(FASTD_VERSION "v10+")
diff --git a/src/cpuid.h b/src/cpuid.h
index 310f0c2..d909ad9 100644
--- a/src/cpuid.h
+++ b/src/cpuid.h
@@ -32,6 +32,7 @@
 
 
 #define CPUID_FXSR	((uint64_t)1 << 24)
+#define CPUID_SSE2	((uint64_t)1 << 26)
 #define CPUID_PCLMULQDQ	((uint64_t)1 << 33)
 #define CPUID_SSSE3	((uint64_t)1 << 41)
 
diff --git a/src/crypto/cipher/salsa20/CMakeLists.txt b/src/crypto/cipher/salsa20/CMakeLists.txt
index 3d9654e..9f8f9f3 100644
--- a/src/crypto/cipher/salsa20/CMakeLists.txt
+++ b/src/crypto/cipher/salsa20/CMakeLists.txt
@@ -1,2 +1,3 @@
 fastd_cipher(salsa20 salsa20.c)
+add_subdirectory(xmm)
 add_subdirectory(nacl)
diff --git a/src/crypto/cipher/salsa20/xmm/CMakeLists.txt b/src/crypto/cipher/salsa20/xmm/CMakeLists.txt
new file mode 100644
index 0000000..b31d94d
--- /dev/null
+++ b/src/crypto/cipher/salsa20/xmm/CMakeLists.txt
@@ -0,0 +1,13 @@
+if(ARCH_X86_64)
+  fastd_cipher_impl(salsa20 xmm
+    salsa20_xmm.c
+    salsa20_amd64_xmm6.s
+  )
+endif(ARCH_X86_64)
+
+if(ARCH_X86)
+  fastd_cipher_impl(salsa20 xmm
+    salsa20_xmm.c
+    salsa20_x86_xmm5.s
+  )
+endif(ARCH_X86)
diff --git a/src/crypto/cipher/salsa20/xmm/salsa20_amd64_xmm6.s b/src/crypto/cipher/salsa20/xmm/salsa20_amd64_xmm6.s
new file mode 100644
index 0000000..82a897f
--- /dev/null
+++ b/src/crypto/cipher/salsa20/xmm/salsa20_amd64_xmm6.s
@@ -0,0 +1,4823 @@
+
+# qhasm: int64 r11_caller
+
+# qhasm: int64 r12_caller
+
+# qhasm: int64 r13_caller
+
+# qhasm: int64 r14_caller
+
+# qhasm: int64 r15_caller
+
+# qhasm: int64 rbx_caller
+
+# qhasm: int64 rbp_caller
+
+# qhasm: caller r11_caller
+
+# qhasm: caller r12_caller
+
+# qhasm: caller r13_caller
+
+# qhasm: caller r14_caller
+
+# qhasm: caller r15_caller
+
+# qhasm: caller rbx_caller
+
+# qhasm: caller rbp_caller
+
+# qhasm: stack64 r11_stack
+
+# qhasm: stack64 r12_stack
+
+# qhasm: stack64 r13_stack
+
+# qhasm: stack64 r14_stack
+
+# qhasm: stack64 r15_stack
+
+# qhasm: stack64 rbx_stack
+
+# qhasm: stack64 rbp_stack
+
+# qhasm: int64 a
+
+# qhasm: int64 arg1
+
+# qhasm: int64 arg2
+
+# qhasm: int64 arg3
+
+# qhasm: int64 arg4
+
+# qhasm: int64 arg5
+
+# qhasm: input arg1
+
+# qhasm: input arg2
+
+# qhasm: input arg3
+
+# qhasm: input arg4
+
+# qhasm: input arg5
+
+# qhasm: int64 k
+
+# qhasm: int64 kbits
+
+# qhasm: int64 iv
+
+# qhasm: int64 i
+
+# qhasm: stack128 x0
+
+# qhasm: stack128 x1
+
+# qhasm: stack128 x2
+
+# qhasm: stack128 x3
+
+# qhasm: int64 m
+
+# qhasm: int64 out
+
+# qhasm: int64 bytes
+
+# qhasm: stack32 eax_stack
+
+# qhasm: stack32 ebx_stack
+
+# qhasm: stack32 esi_stack
+
+# qhasm: stack32 edi_stack
+
+# qhasm: stack32 ebp_stack
+
+# qhasm: int6464 diag0
+
+# qhasm: int6464 diag1
+
+# qhasm: int6464 diag2
+
+# qhasm: int6464 diag3
+
+# qhasm: int6464 a0
+
+# qhasm: int6464 a1
+
+# qhasm: int6464 a2
+
+# qhasm: int6464 a3
+
+# qhasm: int6464 a4
+
+# qhasm: int6464 a5
+
+# qhasm: int6464 a6
+
+# qhasm: int6464 a7
+
+# qhasm: int6464 b0
+
+# qhasm: int6464 b1
+
+# qhasm: int6464 b2
+
+# qhasm: int6464 b3
+
+# qhasm: int6464 b4
+
+# qhasm: int6464 b5
+
+# qhasm: int6464 b6
+
+# qhasm: int6464 b7
+
+# qhasm: int6464 z0
+
+# qhasm: int6464 z1
+
+# qhasm: int6464 z2
+
+# qhasm: int6464 z3
+
+# qhasm: int6464 z4
+
+# qhasm: int6464 z5
+
+# qhasm: int6464 z6
+
+# qhasm: int6464 z7
+
+# qhasm: int6464 z8
+
+# qhasm: int6464 z9
+
+# qhasm: int6464 z10
+
+# qhasm: int6464 z11
+
+# qhasm: int6464 z12
+
+# qhasm: int6464 z13
+
+# qhasm: int6464 z14
+
+# qhasm: int6464 z15
+
+# qhasm: stack128 z0_stack
+
+# qhasm: stack128 z1_stack
+
+# qhasm: stack128 z2_stack
+
+# qhasm: stack128 z3_stack
+
+# qhasm: stack128 z4_stack
+
+# qhasm: stack128 z5_stack
+
+# qhasm: stack128 z6_stack
+
+# qhasm: stack128 z7_stack
+
+# qhasm: stack128 z8_stack
+
+# qhasm: stack128 z9_stack
+
+# qhasm: stack128 z10_stack
+
+# qhasm: stack128 z11_stack
+
+# qhasm: stack128 z12_stack
+
+# qhasm: stack128 z13_stack
+
+# qhasm: stack128 z14_stack
+
+# qhasm: stack128 z15_stack
+
+# qhasm: int6464 y0
+
+# qhasm: int6464 y1
+
+# qhasm: int6464 y2
+
+# qhasm: int6464 y3
+
+# qhasm: int6464 y4
+
+# qhasm: int6464 y5
+
+# qhasm: int6464 y6
+
+# qhasm: int6464 y7
+
+# qhasm: int6464 y8
+
+# qhasm: int6464 y9
+
+# qhasm: int6464 y10
+
+# qhasm: int6464 y11
+
+# qhasm: int6464 y12
+
+# qhasm: int6464 y13
+
+# qhasm: int6464 y14
+
+# qhasm: int6464 y15
+
+# qhasm: int6464 r0
+
+# qhasm: int6464 r1
+
+# qhasm: int6464 r2
+
+# qhasm: int6464 r3
+
+# qhasm: int6464 r4
+
+# qhasm: int6464 r5
+
+# qhasm: int6464 r6
+
+# qhasm: int6464 r7
+
+# qhasm: int6464 r8
+
+# qhasm: int6464 r9
+
+# qhasm: int6464 r10
+
+# qhasm: int6464 r11
+
+# qhasm: int6464 r12
+
+# qhasm: int6464 r13
+
+# qhasm: int6464 r14
+
+# qhasm: int6464 r15
+
+# qhasm: stack128 orig0
+
+# qhasm: stack128 orig1
+
+# qhasm: stack128 orig2
+
+# qhasm: stack128 orig3
+
+# qhasm: stack128 orig4
+
+# qhasm: stack128 orig5
+
+# qhasm: stack128 orig6
+
+# qhasm: stack128 orig7
+
+# qhasm: stack128 orig8
+
+# qhasm: stack128 orig9
+
+# qhasm: stack128 orig10
+
+# qhasm: stack128 orig11
+
+# qhasm: stack128 orig12
+
+# qhasm: stack128 orig13
+
+# qhasm: stack128 orig14
+
+# qhasm: stack128 orig15
+
+# qhasm: int64 in0
+
+# qhasm: int64 in1
+
+# qhasm: int64 in2
+
+# qhasm: int64 in3
+
+# qhasm: int64 in4
+
+# qhasm: int64 in5
+
+# qhasm: int64 in6
+
+# qhasm: int64 in7
+
+# qhasm: int64 in8
+
+# qhasm: int64 in9
+
+# qhasm: int64 in10
+
+# qhasm: int64 in11
+
+# qhasm: int64 in12
+
+# qhasm: int64 in13
+
+# qhasm: int64 in14
+
+# qhasm: int64 in15
+
+# qhasm: stack512 tmp
+
+# qhasm: int64 ctarget
+
+# qhasm: stack64 bytes_backup
+
+# qhasm: enter crypto_stream_salsa20_amd64_xmm6
+.text
+.p2align 5
+.globl _crypto_stream_salsa20_amd64_xmm6
+.globl crypto_stream_salsa20_amd64_xmm6
+_crypto_stream_salsa20_amd64_xmm6:
+crypto_stream_salsa20_amd64_xmm6:
+mov %rsp,%r11
+and $31,%r11
+add $480,%r11
+sub %r11,%rsp
+
+# qhasm: r11_stack = r11_caller
+# asm 1: movq <r11_caller=int64#9,>r11_stack=stack64#1
+# asm 2: movq <r11_caller=%r11,>r11_stack=352(%rsp)
+movq %r11,352(%rsp)
+
+# qhasm: r12_stack = r12_caller
+# asm 1: movq <r12_caller=int64#10,>r12_stack=stack64#2
+# asm 2: movq <r12_caller=%r12,>r12_stack=360(%rsp)
+movq %r12,360(%rsp)
+
+# qhasm: r13_stack = r13_caller
+# asm 1: movq <r13_caller=int64#11,>r13_stack=stack64#3
+# asm 2: movq <r13_caller=%r13,>r13_stack=368(%rsp)
+movq %r13,368(%rsp)
+
+# qhasm: r14_stack = r14_caller
+# asm 1: movq <r14_caller=int64#12,>r14_stack=stack64#4
+# asm 2: movq <r14_caller=%r14,>r14_stack=376(%rsp)
+movq %r14,376(%rsp)
+
+# qhasm: r15_stack = r15_caller
+# asm 1: movq <r15_caller=int64#13,>r15_stack=stack64#5
+# asm 2: movq <r15_caller=%r15,>r15_stack=384(%rsp)
+movq %r15,384(%rsp)
+
+# qhasm: rbx_stack = rbx_caller
+# asm 1: movq <rbx_caller=int64#14,>rbx_stack=stack64#6
+# asm 2: movq <rbx_caller=%rbx,>rbx_stack=392(%rsp)
+movq %rbx,392(%rsp)
+
+# qhasm: rbp_stack = rbp_caller
+# asm 1: movq <rbp_caller=int64#15,>rbp_stack=stack64#7
+# asm 2: movq <rbp_caller=%rbp,>rbp_stack=400(%rsp)
+movq %rbp,400(%rsp)
+
+# qhasm: bytes = arg2
+# asm 1: mov  <arg2=int64#2,>bytes=int64#6
+# asm 2: mov  <arg2=%rsi,>bytes=%r9
+mov  %rsi,%r9
+
+# qhasm: out = arg1
+# asm 1: mov  <arg1=int64#1,>out=int64#1
+# asm 2: mov  <arg1=%rdi,>out=%rdi
+mov  %rdi,%rdi
+
+# qhasm: m = out
+# asm 1: mov  <out=int64#1,>m=int64#2
+# asm 2: mov  <out=%rdi,>m=%rsi
+mov  %rdi,%rsi
+
+# qhasm: iv = arg3
+# asm 1: mov  <arg3=int64#3,>iv=int64#3
+# asm 2: mov  <arg3=%rdx,>iv=%rdx
+mov  %rdx,%rdx
+
+# qhasm: k = arg4
+# asm 1: mov  <arg4=int64#4,>k=int64#8
+# asm 2: mov  <arg4=%rcx,>k=%r10
+mov  %rcx,%r10
+
+# qhasm:               unsigned>? bytes - 0
+# asm 1: cmp  $0,<bytes=int64#6
+# asm 2: cmp  $0,<bytes=%r9
+cmp  $0,%r9
+# comment:fp stack unchanged by jump
+
+# qhasm: goto done if !unsigned>
+jbe ._done
+
+# qhasm: a = 0
+# asm 1: mov  $0,>a=int64#7
+# asm 2: mov  $0,>a=%rax
+mov  $0,%rax
+
+# qhasm: i = bytes
+# asm 1: mov  <bytes=int64#6,>i=int64#4
+# asm 2: mov  <bytes=%r9,>i=%rcx
+mov  %r9,%rcx
+
+# qhasm: while (i) { *out++ = a; --i }
+rep stosb
+
+# qhasm: out -= bytes
+# asm 1: sub  <bytes=int64#6,<out=int64#1
+# asm 2: sub  <bytes=%r9,<out=%rdi
+sub  %r9,%rdi
+# comment:fp stack unchanged by jump
+
+# qhasm: goto start
+jmp ._start
+
+# qhasm: enter crypto_stream_salsa20_amd64_xmm6_xor
+.text
+.p2align 5
+.globl _crypto_stream_salsa20_amd64_xmm6_xor
+.globl crypto_stream_salsa20_amd64_xmm6_xor
+_crypto_stream_salsa20_amd64_xmm6_xor:
+crypto_stream_salsa20_amd64_xmm6_xor:
+mov %rsp,%r11
+and $31,%r11
+add $480,%r11
+sub %r11,%rsp
+
+# qhasm: r11_stack = r11_caller
+# asm 1: movq <r11_caller=int64#9,>r11_stack=stack64#1
+# asm 2: movq <r11_caller=%r11,>r11_stack=352(%rsp)
+movq %r11,352(%rsp)
+
+# qhasm: r12_stack = r12_caller
+# asm 1: movq <r12_caller=int64#10,>r12_stack=stack64#2
+# asm 2: movq <r12_caller=%r12,>r12_stack=360(%rsp)
+movq %r12,360(%rsp)
+
+# qhasm: r13_stack = r13_caller
+# asm 1: movq <r13_caller=int64#11,>r13_stack=stack64#3
+# asm 2: movq <r13_caller=%r13,>r13_stack=368(%rsp)
+movq %r13,368(%rsp)
+
+# qhasm: r14_stack = r14_caller
+# asm 1: movq <r14_caller=int64#12,>r14_stack=stack64#4
+# asm 2: movq <r14_caller=%r14,>r14_stack=376(%rsp)
+movq %r14,376(%rsp)
+
+# qhasm: r15_stack = r15_caller
+# asm 1: movq <r15_caller=int64#13,>r15_stack=stack64#5
+# asm 2: movq <r15_caller=%r15,>r15_stack=384(%rsp)
+movq %r15,384(%rsp)
+
+# qhasm: rbx_stack = rbx_caller
+# asm 1: movq <rbx_caller=int64#14,>rbx_stack=stack64#6
+# asm 2: movq <rbx_caller=%rbx,>rbx_stack=392(%rsp)
+movq %rbx,392(%rsp)
+
+# qhasm: rbp_stack = rbp_caller
+# asm 1: movq <rbp_caller=int64#15,>rbp_stack=stack64#7
+# asm 2: movq <rbp_caller=%rbp,>rbp_stack=400(%rsp)
+movq %rbp,400(%rsp)
+
+# qhasm: out = arg1
+# asm 1: mov  <arg1=int64#1,>out=int64#1
+# asm 2: mov  <arg1=%rdi,>out=%rdi
+mov  %rdi,%rdi
+
+# qhasm: m = arg2
+# asm 1: mov  <arg2=int64#2,>m=int64#2
+# asm 2: mov  <arg2=%rsi,>m=%rsi
+mov  %rsi,%rsi
+
+# qhasm: bytes = arg3
+# asm 1: mov  <arg3=int64#3,>bytes=int64#6
+# asm 2: mov  <arg3=%rdx,>bytes=%r9
+mov  %rdx,%r9
+
+# qhasm: iv = arg4
+# asm 1: mov  <arg4=int64#4,>iv=int64#3
+# asm 2: mov  <arg4=%rcx,>iv=%rdx
+mov  %rcx,%rdx
+
+# qhasm: k = arg5
+# asm 1: mov  <arg5=int64#5,>k=int64#8
+# asm 2: mov  <arg5=%r8,>k=%r10
+mov  %r8,%r10
+
+# qhasm:               unsigned>? bytes - 0
+# asm 1: cmp  $0,<bytes=int64#6
+# asm 2: cmp  $0,<bytes=%r9
+cmp  $0,%r9
+# comment:fp stack unchanged by jump
+
+# qhasm: goto done if !unsigned>
+jbe ._done
+# comment:fp stack unchanged by fallthrough
+
+# qhasm: start:
+._start:
+
+# qhasm:   in12 = *(uint32 *) (k + 20)
+# asm 1: movl   20(<k=int64#8),>in12=int64#4d
+# asm 2: movl   20(<k=%r10),>in12=%ecx
+movl   20(%r10),%ecx
+
+# qhasm:   in1 = *(uint32 *) (k + 0)
+# asm 1: movl   0(<k=int64#8),>in1=int64#5d
+# asm 2: movl   0(<k=%r10),>in1=%r8d
+movl   0(%r10),%r8d
+
+# qhasm:   in6 = *(uint32 *) (iv + 0)
+# asm 1: movl   0(<iv=int64#3),>in6=int64#7d
+# asm 2: movl   0(<iv=%rdx),>in6=%eax
+movl   0(%rdx),%eax
+
+# qhasm:   in11 = *(uint32 *) (k + 16)
+# asm 1: movl   16(<k=int64#8),>in11=int64#9d
+# asm 2: movl   16(<k=%r10),>in11=%r11d
+movl   16(%r10),%r11d
+
+# qhasm:   ((uint32 *)&x1)[0] = in12
+# asm 1: movl <in12=int64#4d,>x1=stack128#1
+# asm 2: movl <in12=%ecx,>x1=0(%rsp)
+movl %ecx,0(%rsp)
+
+# qhasm:   ((uint32 *)&x1)[1] = in1
+# asm 1: movl <in1=int64#5d,4+<x1=stack128#1
+# asm 2: movl <in1=%r8d,4+<x1=0(%rsp)
+movl %r8d,4+0(%rsp)
+
+# qhasm:   ((uint32 *)&x1)[2] = in6
+# asm 1: movl <in6=int64#7d,8+<x1=stack128#1
+# asm 2: movl <in6=%eax,8+<x1=0(%rsp)
+movl %eax,8+0(%rsp)
+
+# qhasm:   ((uint32 *)&x1)[3] = in11
+# asm 1: movl <in11=int64#9d,12+<x1=stack128#1
+# asm 2: movl <in11=%r11d,12+<x1=0(%rsp)
+movl %r11d,12+0(%rsp)
+
+# qhasm:   in8 = 0
+# asm 1: mov  $0,>in8=int64#4
+# asm 2: mov  $0,>in8=%rcx
+mov  $0,%rcx
+
+# qhasm:   in13 = *(uint32 *) (k + 24)
+# asm 1: movl   24(<k=int64#8),>in13=int64#5d
+# asm 2: movl   24(<k=%r10),>in13=%r8d
+movl   24(%r10),%r8d
+
+# qhasm:   in2 = *(uint32 *) (k + 4)
+# asm 1: movl   4(<k=int64#8),>in2=int64#7d
+# asm 2: movl   4(<k=%r10),>in2=%eax
+movl   4(%r10),%eax
+
+# qhasm:   in7 = *(uint32 *) (iv + 4)
+# asm 1: movl   4(<iv=int64#3),>in7=int64#3d
+# asm 2: movl   4(<iv=%rdx),>in7=%edx
+movl   4(%rdx),%edx
+
+# qhasm:   ((uint32 *)&x2)[0] = in8
+# asm 1: movl <in8=int64#4d,>x2=stack128#2
+# asm 2: movl <in8=%ecx,>x2=16(%rsp)
+movl %ecx,16(%rsp)
+
+# qhasm:   ((uint32 *)&x2)[1] = in13
+# asm 1: movl <in13=int64#5d,4+<x2=stack128#2
+# asm 2: movl <in13=%r8d,4+<x2=16(%rsp)
+movl %r8d,4+16(%rsp)
+
+# qhasm:   ((uint32 *)&x2)[2] = in2
+# asm 1: movl <in2=int64#7d,8+<x2=stack128#2
+# asm 2: movl <in2=%eax,8+<x2=16(%rsp)
+movl %eax,8+16(%rsp)
+
+# qhasm:   ((uint32 *)&x2)[3] = in7
+# asm 1: movl <in7=int64#3d,12+<x2=stack128#2
+# asm 2: movl <in7=%edx,12+<x2=16(%rsp)
+movl %edx,12+16(%rsp)
+
+# qhasm:   in4 = *(uint32 *) (k + 12)
+# asm 1: movl   12(<k=int64#8),>in4=int64#3d
+# asm 2: movl   12(<k=%r10),>in4=%edx
+movl   12(%r10),%edx
+
+# qhasm:   in9 = 0
+# asm 1: mov  $0,>in9=int64#4
+# asm 2: mov  $0,>in9=%rcx
+mov  $0,%rcx
+
+# qhasm:   in14 = *(uint32 *) (k + 28)
+# asm 1: movl   28(<k=int64#8),>in14=int64#5d
+# asm 2: movl   28(<k=%r10),>in14=%r8d
+movl   28(%r10),%r8d
+
+# qhasm:   in3 = *(uint32 *) (k + 8)
+# asm 1: movl   8(<k=int64#8),>in3=int64#7d
+# asm 2: movl   8(<k=%r10),>in3=%eax
+movl   8(%r10),%eax
+
+# qhasm:   ((uint32 *)&x3)[0] = in4
+# asm 1: movl <in4=int64#3d,>x3=stack128#3
+# asm 2: movl <in4=%edx,>x3=32(%rsp)
+movl %edx,32(%rsp)
+
+# qhasm:   ((uint32 *)&x3)[1] = in9
+# asm 1: movl <in9=int64#4d,4+<x3=stack128#3
+# asm 2: movl <in9=%ecx,4+<x3=32(%rsp)
+movl %ecx,4+32(%rsp)
+
+# qhasm:   ((uint32 *)&x3)[2] = in14
+# asm 1: movl <in14=int64#5d,8+<x3=stack128#3
+# asm 2: movl <in14=%r8d,8+<x3=32(%rsp)
+movl %r8d,8+32(%rsp)
+
+# qhasm:   ((uint32 *)&x3)[3] = in3
+# asm 1: movl <in3=int64#7d,12+<x3=stack128#3
+# asm 2: movl <in3=%eax,12+<x3=32(%rsp)
+movl %eax,12+32(%rsp)
+
+# qhasm:   in0 = 1634760805
+# asm 1: mov  $1634760805,>in0=int64#3
+# asm 2: mov  $1634760805,>in0=%rdx
+mov  $1634760805,%rdx
+
+# qhasm:   in5 = 857760878
+# asm 1: mov  $857760878,>in5=int64#4
+# asm 2: mov  $857760878,>in5=%rcx
+mov  $857760878,%rcx
+
+# qhasm:   in10 = 2036477234
+# asm 1: mov  $2036477234,>in10=int64#5
+# asm 2: mov  $2036477234,>in10=%r8
+mov  $2036477234,%r8
+
+# qhasm:   in15 = 1797285236
+# asm 1: mov  $1797285236,>in15=int64#7
+# asm 2: mov  $1797285236,>in15=%rax
+mov  $1797285236,%rax
+
+# qhasm:   ((uint32 *)&x0)[0] = in0
+# asm 1: movl <in0=int64#3d,>x0=stack128#4
+# asm 2: movl <in0=%edx,>x0=48(%rsp)
+movl %edx,48(%rsp)
+
+# qhasm:   ((uint32 *)&x0)[1] = in5
+# asm 1: movl <in5=int64#4d,4+<x0=stack128#4
+# asm 2: movl <in5=%ecx,4+<x0=48(%rsp)
+movl %ecx,4+48(%rsp)
+
+# qhasm:   ((uint32 *)&x0)[2] = in10
+# asm 1: movl <in10=int64#5d,8+<x0=stack128#4
+# asm 2: movl <in10=%r8d,8+<x0=48(%rsp)
+movl %r8d,8+48(%rsp)
+
+# qhasm:   ((uint32 *)&x0)[3] = in15
+# asm 1: movl <in15=int64#7d,12+<x0=stack128#4
+# asm 2: movl <in15=%eax,12+<x0=48(%rsp)
+movl %eax,12+48(%rsp)
+
+# qhasm:                               unsigned<? bytes - 256
+# asm 1: cmp  $256,<bytes=int64#6
+# asm 2: cmp  $256,<bytes=%r9
+cmp  $256,%r9
+# comment:fp stack unchanged by jump
+
+# qhasm:   goto bytesbetween1and255 if unsigned<
+jb ._bytesbetween1and255
+
+# qhasm:   z0 = x0
+# asm 1: movdqa <x0=stack128#4,>z0=int6464#1
+# asm 2: movdqa <x0=48(%rsp),>z0=%xmm0
+movdqa 48(%rsp),%xmm0
+
+# qhasm:   z5 = z0[1,1,1,1]
+# asm 1: pshufd $0x55,<z0=int6464#1,>z5=int6464#2
+# asm 2: pshufd $0x55,<z0=%xmm0,>z5=%xmm1
+pshufd $0x55,%xmm0,%xmm1
+
+# qhasm:   z10 = z0[2,2,2,2]
+# asm 1: pshufd $0xaa,<z0=int6464#1,>z10=int6464#3
+# asm 2: pshufd $0xaa,<z0=%xmm0,>z10=%xmm2
+pshufd $0xaa,%xmm0,%xmm2
+
+# qhasm:   z15 = z0[3,3,3,3]
+# asm 1: pshufd $0xff,<z0=int6464#1,>z15=int6464#4
+# asm 2: pshufd $0xff,<z0=%xmm0,>z15=%xmm3
+pshufd $0xff,%xmm0,%xmm3
+
+# qhasm:   z0 = z0[0,0,0,0]
+# asm 1: pshufd $0x00,<z0=int6464#1,>z0=int6464#1
+# asm 2: pshufd $0x00,<z0=%xmm0,>z0=%xmm0
+pshufd $0x00,%xmm0,%xmm0
+
+# qhasm:   orig5 = z5
+# asm 1: movdqa <z5=int6464#2,>orig5=stack128#5
+# asm 2: movdqa <z5=%xmm1,>orig5=64(%rsp)
+movdqa %xmm1,64(%rsp)
+
+# qhasm:   orig10 = z10
+# asm 1: movdqa <z10=int6464#3,>orig10=stack128#6
+# asm 2: movdqa <z10=%xmm2,>orig10=80(%rsp)
+movdqa %xmm2,80(%rsp)
+
+# qhasm:   orig15 = z15
+# asm 1: movdqa <z15=int6464#4,>orig15=stack128#7
+# asm 2: movdqa <z15=%xmm3,>orig15=96(%rsp)
+movdqa %xmm3,96(%rsp)
+
+# qhasm:   orig0 = z0
+# asm 1: movdqa <z0=int6464#1,>orig0=stack128#8
+# asm 2: movdqa <z0=%xmm0,>orig0=112(%rsp)
+movdqa %xmm0,112(%rsp)
+
+# qhasm:   z1 = x1
+# asm 1: movdqa <x1=stack128#1,>z1=int6464#1
+# asm 2: movdqa <x1=0(%rsp),>z1=%xmm0
+movdqa 0(%rsp),%xmm0
+
+# qhasm:   z6 = z1[2,2,2,2]
+# asm 1: pshufd $0xaa,<z1=int6464#1,>z6=int6464#2
+# asm 2: pshufd $0xaa,<z1=%xmm0,>z6=%xmm1
+pshufd $0xaa,%xmm0,%xmm1
+
+# qhasm:   z11 = z1[3,3,3,3]
+# asm 1: pshufd $0xff,<z1=int6464#1,>z11=int6464#3
+# asm 2: pshufd $0xff,<z1=%xmm0,>z11=%xmm2
+pshufd $0xff,%xmm0,%xmm2
+
+# qhasm:   z12 = z1[0,0,0,0]
+# asm 1: pshufd $0x00,<z1=int6464#1,>z12=int6464#4
+# asm 2: pshufd $0x00,<z1=%xmm0,>z12=%xmm3
+pshufd $0x00,%xmm0,%xmm3
+
+# qhasm:   z1 = z1[1,1,1,1]
+# asm 1: pshufd $0x55,<z1=int6464#1,>z1=int6464#1
+# asm 2: pshufd $0x55,<z1=%xmm0,>z1=%xmm0
+pshufd $0x55,%xmm0,%xmm0
+
+# qhasm:   orig6 = z6
+# asm 1: movdqa <z6=int6464#2,>orig6=stack128#9
+# asm 2: movdqa <z6=%xmm1,>orig6=128(%rsp)
+movdqa %xmm1,128(%rsp)
+
+# qhasm:   orig11 = z11
+# asm 1: movdqa <z11=int6464#3,>orig11=stack128#10
+# asm 2: movdqa <z11=%xmm2,>orig11=144(%rsp)
+movdqa %xmm2,144(%rsp)
+
+# qhasm:   orig12 = z12
+# asm 1: movdqa <z12=int6464#4,>orig12=stack128#11
+# asm 2: movdqa <z12=%xmm3,>orig12=160(%rsp)
+movdqa %xmm3,160(%rsp)
+
+# qhasm:   orig1 = z1
+# asm 1: movdqa <z1=int6464#1,>orig1=stack128#12
+# asm 2: movdqa <z1=%xmm0,>orig1=176(%rsp)
+movdqa %xmm0,176(%rsp)
+
+# qhasm:   z2 = x2
+# asm 1: movdqa <x2=stack128#2,>z2=int6464#1
+# asm 2: movdqa <x2=16(%rsp),>z2=%xmm0
+movdqa 16(%rsp),%xmm0
+
+# qhasm:   z7 = z2[3,3,3,3]
+# asm 1: pshufd $0xff,<z2=int6464#1,>z7=int6464#2
+# asm 2: pshufd $0xff,<z2=%xmm0,>z7=%xmm1
+pshufd $0xff,%xmm0,%xmm1
+
+# qhasm:   z13 = z2[1,1,1,1]
+# asm 1: pshufd $0x55,<z2=int6464#1,>z13=int6464#3
+# asm 2: pshufd $0x55,<z2=%xmm0,>z13=%xmm2
+pshufd $0x55,%xmm0,%xmm2
+
+# qhasm:   z2 = z2[2,2,2,2]
+# asm 1: pshufd $0xaa,<z2=int6464#1,>z2=int6464#1
+# asm 2: pshufd $0xaa,<z2=%xmm0,>z2=%xmm0
+pshufd $0xaa,%xmm0,%xmm0
+
+# qhasm:   orig7 = z7
+# asm 1: movdqa <z7=int6464#2,>orig7=stack128#13
+# asm 2: movdqa <z7=%xmm1,>orig7=192(%rsp)
+movdqa %xmm1,192(%rsp)
+
+# qhasm:   orig13 = z13
+# asm 1: movdqa <z13=int6464#3,>orig13=stack128#14
+# asm 2: movdqa <z13=%xmm2,>orig13=208(%rsp)
+movdqa %xmm2,208(%rsp)
+
+# qhasm:   orig2 = z2
+# asm 1: movdqa <z2=int6464#1,>orig2=stack128#15
+# asm 2: movdqa <z2=%xmm0,>orig2=224(%rsp)
+movdqa %xmm0,224(%rsp)
+
+# qhasm:   z3 = x3
+# asm 1: movdqa <x3=stack128#3,>z3=int6464#1
+# asm 2: movdqa <x3=32(%rsp),>z3=%xmm0
+movdqa 32(%rsp),%xmm0
+
+# qhasm:   z4 = z3[0,0,0,0]
+# asm 1: pshufd $0x00,<z3=int6464#1,>z4=int6464#2
+# asm 2: pshufd $0x00,<z3=%xmm0,>z4=%xmm1
+pshufd $0x00,%xmm0,%xmm1
+
+# qhasm:   z14 = z3[2,2,2,2]
+# asm 1: pshufd $0xaa,<z3=int6464#1,>z14=int6464#3
+# asm 2: pshufd $0xaa,<z3=%xmm0,>z14=%xmm2
+pshufd $0xaa,%xmm0,%xmm2
+
+# qhasm:   z3 = z3[3,3,3,3]
+# asm 1: pshufd $0xff,<z3=int6464#1,>z3=int6464#1
+# asm 2: pshufd $0xff,<z3=%xmm0,>z3=%xmm0
+pshufd $0xff,%xmm0,%xmm0
+
+# qhasm:   orig4 = z4
+# asm 1: movdqa <z4=int6464#2,>orig4=stack128#16
+# asm 2: movdqa <z4=%xmm1,>orig4=240(%rsp)
+movdqa %xmm1,240(%rsp)
+
+# qhasm:   orig14 = z14
+# asm 1: movdqa <z14=int6464#3,>orig14=stack128#17
+# asm 2: movdqa <z14=%xmm2,>orig14=256(%rsp)
+movdqa %xmm2,256(%rsp)
+
+# qhasm:   orig3 = z3
+# asm 1: movdqa <z3=int6464#1,>orig3=stack128#18
+# asm 2: movdqa <z3=%xmm0,>orig3=272(%rsp)
+movdqa %xmm0,272(%rsp)
+
+# qhasm: bytesatleast256:
+._bytesatleast256:
+
+# qhasm:   in8 = ((uint32 *)&x2)[0]
+# asm 1: movl <x2=stack128#2,>in8=int64#3d
+# asm 2: movl <x2=16(%rsp),>in8=%edx
+movl 16(%rsp),%edx
+
+# qhasm:   in9 = ((uint32 *)&x3)[1]
+# asm 1: movl 4+<x3=stack128#3,>in9=int64#4d
+# asm 2: movl 4+<x3=32(%rsp),>in9=%ecx
+movl 4+32(%rsp),%ecx
+
+# qhasm:   ((uint32 *) &orig8)[0] = in8
+# asm 1: movl <in8=int64#3d,>orig8=stack128#19
+# asm 2: movl <in8=%edx,>orig8=288(%rsp)
+movl %edx,288(%rsp)
+
+# qhasm:   ((uint32 *) &orig9)[0] = in9
+# asm 1: movl <in9=int64#4d,>orig9=stack128#20
+# asm 2: movl <in9=%ecx,>orig9=304(%rsp)
+movl %ecx,304(%rsp)
+
+# qhasm:   in8 += 1
+# asm 1: add  $1,<in8=int64#3
+# asm 2: add  $1,<in8=%rdx
+add  $1,%rdx
+
+# qhasm:   in9 <<= 32
+# asm 1: shl  $32,<in9=int64#4
+# asm 2: shl  $32,<in9=%rcx
+shl  $32,%rcx
+
+# qhasm:   in8 += in9
+# asm 1: add  <in9=int64#4,<in8=int64#3
+# asm 2: add  <in9=%rcx,<in8=%rdx
+add  %rcx,%rdx
+
+# qhasm:   in9 = in8
+# asm 1: mov  <in8=int64#3,>in9=int64#4
+# asm 2: mov  <in8=%rdx,>in9=%rcx
+mov  %rdx,%rcx
+
+# qhasm:   (uint64) in9 >>= 32
+# asm 1: shr  $32,<in9=int64#4
+# asm 2: shr  $32,<in9=%rcx
+shr  $32,%rcx
+
+# qhasm:   ((uint32 *) &orig8)[1] = in8
+# asm 1: movl <in8=int64#3d,4+<orig8=stack128#19
+# asm 2: movl <in8=%edx,4+<orig8=288(%rsp)
+movl %edx,4+288(%rsp)
+
+# qhasm:   ((uint32 *) &orig9)[1] = in9
+# asm 1: movl <in9=int64#4d,4+<orig9=stack128#20
+# asm 2: movl <in9=%ecx,4+<orig9=304(%rsp)
+movl %ecx,4+304(%rsp)
+
+# qhasm:   in8 += 1
+# asm 1: add  $1,<in8=int64#3
+# asm 2: add  $1,<in8=%rdx
+add  $1,%rdx
+
+# qhasm:   in9 <<= 32
+# asm 1: shl  $32,<in9=int64#4
+# asm 2: shl  $32,<in9=%rcx
+shl  $32,%rcx
+
+# qhasm:   in8 += in9
+# asm 1: add  <in9=int64#4,<in8=int64#3
+# asm 2: add  <in9=%rcx,<in8=%rdx
+add  %rcx,%rdx
+
+# qhasm:   in9 = in8
+# asm 1: mov  <in8=int64#3,>in9=int64#4
+# asm 2: mov  <in8=%rdx,>in9=%rcx
+mov  %rdx,%rcx
+
+# qhasm:   (uint64) in9 >>= 32
+# asm 1: shr  $32,<in9=int64#4
+# asm 2: shr  $32,<in9=%rcx
+shr  $32,%rcx
+
+# qhasm:   ((uint32 *) &orig8)[2] = in8
+# asm 1: movl <in8=int64#3d,8+<orig8=stack128#19
+# asm 2: movl <in8=%edx,8+<orig8=288(%rsp)
+movl %edx,8+288(%rsp)
+
+# qhasm:   ((uint32 *) &orig9)[2] = in9
+# asm 1: movl <in9=int64#4d,8+<orig9=stack128#20
+# asm 2: movl <in9=%ecx,8+<orig9=304(%rsp)
+movl %ecx,8+304(%rsp)
+
+# qhasm:   in8 += 1
+# asm 1: add  $1,<in8=int64#3
+# asm 2: add  $1,<in8=%rdx
+add  $1,%rdx
+
+# qhasm:   in9 <<= 32
+# asm 1: shl  $32,<in9=int64#4
+# asm 2: shl  $32,<in9=%rcx
+shl  $32,%rcx
+
+# qhasm:   in8 += in9
+# asm 1: add  <in9=int64#4,<in8=int64#3
+# asm 2: add  <in9=%rcx,<in8=%rdx
+add  %rcx,%rdx
+
+# qhasm:   in9 = in8
+# asm 1: mov  <in8=int64#3,>in9=int64#4
+# asm 2: mov  <in8=%rdx,>in9=%rcx
+mov  %rdx,%rcx
+
+# qhasm:   (uint64) in9 >>= 32
+# asm 1: shr  $32,<in9=int64#4
+# asm 2: shr  $32,<in9=%rcx
+shr  $32,%rcx
+
+# qhasm:   ((uint32 *) &orig8)[3] = in8
+# asm 1: movl <in8=int64#3d,12+<orig8=stack128#19
+# asm 2: movl <in8=%edx,12+<orig8=288(%rsp)
+movl %edx,12+288(%rsp)
+
+# qhasm:   ((uint32 *) &orig9)[3] = in9
+# asm 1: movl <in9=int64#4d,12+<orig9=stack128#20
+# asm 2: movl <in9=%ecx,12+<orig9=304(%rsp)
+movl %ecx,12+304(%rsp)
+
+# qhasm:   in8 += 1
+# asm 1: add  $1,<in8=int64#3
+# asm 2: add  $1,<in8=%rdx
+add  $1,%rdx
+
+# qhasm:   in9 <<= 32
+# asm 1: shl  $32,<in9=int64#4
+# asm 2: shl  $32,<in9=%rcx
+shl  $32,%rcx
+
+# qhasm:   in8 += in9
+# asm 1: add  <in9=int64#4,<in8=int64#3
+# asm 2: add  <in9=%rcx,<in8=%rdx
+add  %rcx,%rdx
+
+# qhasm:   in9 = in8
+# asm 1: mov  <in8=int64#3,>in9=int64#4
+# asm 2: mov  <in8=%rdx,>in9=%rcx
+mov  %rdx,%rcx
+
+# qhasm:   (uint64) in9 >>= 32
+# asm 1: shr  $32,<in9=int64#4
+# asm 2: shr  $32,<in9=%rcx
+shr  $32,%rcx
+
+# qhasm:   ((uint32 *)&x2)[0] = in8
+# asm 1: movl <in8=int64#3d,>x2=stack128#2
+# asm 2: movl <in8=%edx,>x2=16(%rsp)
+movl %edx,16(%rsp)
+
+# qhasm:   ((uint32 *)&x3)[1] = in9
+# asm 1: movl <in9=int64#4d,4+<x3=stack128#3
+# asm 2: movl <in9=%ecx,4+<x3=32(%rsp)
+movl %ecx,4+32(%rsp)
+
+# qhasm:   bytes_backup = bytes
+# asm 1: movq <bytes=int64#6,>bytes_backup=stack64#8
+# asm 2: movq <bytes=%r9,>bytes_backup=408(%rsp)
+movq %r9,408(%rsp)
+
+# qhasm: i = 20
+# asm 1: mov  $20,>i=int64#3
+# asm 2: mov  $20,>i=%rdx
+mov  $20,%rdx
+
+# qhasm:   z5 = orig5
+# asm 1: movdqa <orig5=stack128#5,>z5=int6464#1
+# asm 2: movdqa <orig5=64(%rsp),>z5=%xmm0
+movdqa 64(%rsp),%xmm0
+
+# qhasm:   z10 = orig10
+# asm 1: movdqa <orig10=stack128#6,>z10=int6464#2
+# asm 2: movdqa <orig10=80(%rsp),>z10=%xmm1
+movdqa 80(%rsp),%xmm1
+
+# qhasm:   z15 = orig15
+# asm 1: movdqa <orig15=stack128#7,>z15=int6464#3
+# asm 2: movdqa <orig15=96(%rsp),>z15=%xmm2
+movdqa 96(%rsp),%xmm2
+
+# qhasm:   z14 = orig14
+# asm 1: movdqa <orig14=stack128#17,>z14=int6464#4
+# asm 2: movdqa <orig14=256(%rsp),>z14=%xmm3
+movdqa 256(%rsp),%xmm3
+
+# qhasm:   z3 = orig3
+# asm 1: movdqa <orig3=stack128#18,>z3=int6464#5
+# asm 2: movdqa <orig3=272(%rsp),>z3=%xmm4
+movdqa 272(%rsp),%xmm4
+
+# qhasm:   z6 = orig6
+# asm 1: movdqa <orig6=stack128#9,>z6=int6464#6
+# asm 2: movdqa <orig6=128(%rsp),>z6=%xmm5
+movdqa 128(%rsp),%xmm5
+
+# qhasm:   z11 = orig11
+# asm 1: movdqa <orig11=stack128#10,>z11=int6464#7
+# asm 2: movdqa <orig11=144(%rsp),>z11=%xmm6
+movdqa 144(%rsp),%xmm6
+
+# qhasm:   z1 = orig1
+# asm 1: movdqa <orig1=stack128#12,>z1=int6464#8
+# asm 2: movdqa <orig1=176(%rsp),>z1=%xmm7
+movdqa 176(%rsp),%xmm7
+
+# qhasm:   z7 = orig7
+# asm 1: movdqa <orig7=stack128#13,>z7=int6464#9
+# asm 2: movdqa <orig7=192(%rsp),>z7=%xmm8
+movdqa 192(%rsp),%xmm8
+
+# qhasm:   z13 = orig13
+# asm 1: movdqa <orig13=stack128#14,>z13=int6464#10
+# asm 2: movdqa <orig13=208(%rsp),>z13=%xmm9
+movdqa 208(%rsp),%xmm9
+
+# qhasm:   z2 = orig2
+# asm 1: movdqa <orig2=stack128#15,>z2=int6464#11
+# asm 2: movdqa <orig2=224(%rsp),>z2=%xmm10
+movdqa 224(%rsp),%xmm10
+
+# qhasm:   z9 = orig9
+# asm 1: movdqa <orig9=stack128#20,>z9=int6464#12
+# asm 2: movdqa <orig9=304(%rsp),>z9=%xmm11
+movdqa 304(%rsp),%xmm11
+
+# qhasm:   z0 = orig0
+# asm 1: movdqa <orig0=stack128#8,>z0=int6464#13
+# asm 2: movdqa <orig0=112(%rsp),>z0=%xmm12
+movdqa 112(%rsp),%xmm12
+
+# qhasm:   z12 = orig12
+# asm 1: movdqa <orig12=stack128#11,>z12=int6464#14
+# asm 2: movdqa <orig12=160(%rsp),>z12=%xmm13
+movdqa 160(%rsp),%xmm13
+
+# qhasm:   z4 = orig4
+# asm 1: movdqa <orig4=stack128#16,>z4=int6464#15
+# asm 2: movdqa <orig4=240(%rsp),>z4=%xmm14
+movdqa 240(%rsp),%xmm14
+
+# qhasm:   z8 = orig8
+# asm 1: movdqa <orig8=stack128#19,>z8=int6464#16
+# asm 2: movdqa <orig8=288(%rsp),>z8=%xmm15
+movdqa 288(%rsp),%xmm15
+
+# qhasm: mainloop1:
+._mainloop1:
+
+# qhasm: 						z10_stack = z10
+# asm 1: movdqa <z10=int6464#2,>z10_stack=stack128#21
+# asm 2: movdqa <z10=%xmm1,>z10_stack=320(%rsp)
+movdqa %xmm1,320(%rsp)
+
+# qhasm: 								z15_stack = z15
+# asm 1: movdqa <z15=int6464#3,>z15_stack=stack128#22
+# asm 2: movdqa <z15=%xmm2,>z15_stack=336(%rsp)
+movdqa %xmm2,336(%rsp)
+
+# qhasm: 		y4 = z12
+# asm 1: movdqa <z12=int6464#14,>y4=int6464#2
+# asm 2: movdqa <z12=%xmm13,>y4=%xmm1
+movdqa %xmm13,%xmm1
+
+# qhasm: uint32323232	y4 += z0
+# asm 1: paddd <z0=int6464#13,<y4=int6464#2
+# asm 2: paddd <z0=%xmm12,<y4=%xmm1
+paddd %xmm12,%xmm1
+
+# qhasm: 		r4 = y4
+# asm 1: movdqa <y4=int6464#2,>r4=int6464#3
+# asm 2: movdqa <y4=%xmm1,>r4=%xmm2
+movdqa %xmm1,%xmm2
+
+# qhasm: uint32323232	y4 <<= 7
+# asm 1: pslld $7,<y4=int6464#2
+# asm 2: pslld $7,<y4=%xmm1
+pslld $7,%xmm1
+
+# qhasm: 		z4 ^= y4
+# asm 1: pxor  <y4=int6464#2,<z4=int6464#15
+# asm 2: pxor  <y4=%xmm1,<z4=%xmm14
+pxor  %xmm1,%xmm14
+
+# qhasm: uint32323232	r4 >>= 25
+# asm 1: psrld $25,<r4=int6464#3
+# asm 2: psrld $25,<r4=%xmm2
+psrld $25,%xmm2
+
+# qhasm: 		z4 ^= r4
+# asm 1: pxor  <r4=int6464#3,<z4=int6464#15
+# asm 2: pxor  <r4=%xmm2,<z4=%xmm14
+pxor  %xmm2,%xmm14
+
+# qhasm: 				y9 = z1
+# asm 1: movdqa <z1=int6464#8,>y9=int6464#2
+# asm 2: movdqa <z1=%xmm7,>y9=%xmm1
+movdqa %xmm7,%xmm1
+
+# qhasm: uint32323232			y9 += z5
+# asm 1: paddd <z5=int6464#1,<y9=int6464#2
+# asm 2: paddd <z5=%xmm0,<y9=%xmm1
+paddd %xmm0,%xmm1
+
+# qhasm: 				r9 = y9
+# asm 1: movdqa <y9=int6464#2,>r9=int6464#3
+# asm 2: movdqa <y9=%xmm1,>r9=%xmm2
+movdqa %xmm1,%xmm2
+
+# qhasm: uint32323232			y9 <<= 7
+# asm 1: pslld $7,<y9=int6464#2
+# asm 2: pslld $7,<y9=%xmm1
+pslld $7,%xmm1
+
+# qhasm: 				z9 ^= y9
+# asm 1: pxor  <y9=int6464#2,<z9=int6464#12
+# asm 2: pxor  <y9=%xmm1,<z9=%xmm11
+pxor  %xmm1,%xmm11
+
+# qhasm: uint32323232			r9 >>= 25
+# asm 1: psrld $25,<r9=int6464#3
+# asm 2: psrld $25,<r9=%xmm2
+psrld $25,%xmm2
+
+# qhasm: 				z9 ^= r9
+# asm 1: pxor  <r9=int6464#3,<z9=int6464#12
+# asm 2: pxor  <r9=%xmm2,<z9=%xmm11
+pxor  %xmm2,%xmm11
+
+# qhasm: 		y8 = z0
+# asm 1: movdqa <z0=int6464#13,>y8=int6464#2
+# asm 2: movdqa <z0=%xmm12,>y8=%xmm1
+movdqa %xmm12,%xmm1
+
+# qhasm: uint32323232	y8 += z4
+# asm 1: paddd <z4=int6464#15,<y8=int6464#2
+# asm 2: paddd <z4=%xmm14,<y8=%xmm1
+paddd %xmm14,%xmm1
+
+# qhasm: 		r8 = y8
+# asm 1: movdqa <y8=int6464#2,>r8=int6464#3
+# asm 2: movdqa <y8=%xmm1,>r8=%xmm2
+movdqa %xmm1,%xmm2
+
+# qhasm: uint32323232	y8 <<= 9
+# asm 1: pslld $9,<y8=int6464#2
+# asm 2: pslld $9,<y8=%xmm1
+pslld $9,%xmm1
+
+# qhasm: 		z8 ^= y8
+# asm 1: pxor  <y8=int6464#2,<z8=int6464#16
+# asm 2: pxor  <y8=%xmm1,<z8=%xmm15
+pxor  %xmm1,%xmm15
+
+# qhasm: uint32323232	r8 >>= 23
+# asm 1: psrld $23,<r8=int6464#3
+# asm 2: psrld $23,<r8=%xmm2
+psrld $23,%xmm2
+
+# qhasm: 		z8 ^= r8
+# asm 1: pxor  <r8=int6464#3,<z8=int6464#16
+# asm 2: pxor  <r8=%xmm2,<z8=%xmm15
+pxor  %xmm2,%xmm15
+
+# qhasm: 				y13 = z5
+# asm 1: movdqa <z5=int6464#1,>y13=int6464#2
+# asm 2: movdqa <z5=%xmm0,>y13=%xmm1
+movdqa %xmm0,%xmm1
+
+# qhasm: uint32323232			y13 += z9
+# asm 1: paddd <z9=int6464#12,<y13=int6464#2
+# asm 2: paddd <z9=%xmm11,<y13=%xmm1
+paddd %xmm11,%xmm1
+
+# qhasm: 				r13 = y13
+# asm 1: movdqa <y13=int6464#2,>r13=int6464#3
+# asm 2: movdqa <y13=%xmm1,>r13=%xmm2
+movdqa %xmm1,%xmm2
+
+# qhasm: uint32323232			y13 <<= 9
+# asm 1: pslld $9,<y13=int6464#2
+# asm 2: pslld $9,<y13=%xmm1
+pslld $9,%xmm1
+
+# qhasm: 				z13 ^= y13
+# asm 1: pxor  <y13=int6464#2,<z13=int6464#10
+# asm 2: pxor  <y13=%xmm1,<z13=%xmm9
+pxor  %xmm1,%xmm9
+
+# qhasm: uint32323232			r13 >>= 23
+# asm 1: psrld $23,<r13=int6464#3
+# asm 2: psrld $23,<r13=%xmm2
+psrld $23,%xmm2
+
+# qhasm: 				z13 ^= r13
+# asm 1: pxor  <r13=int6464#3,<z13=int6464#10
+# asm 2: pxor  <r13=%xmm2,<z13=%xmm9
+pxor  %xmm2,%xmm9
+
+# qhasm: 		y12 = z4
+# asm 1: movdqa <z4=int6464#15,>y12=int6464#2
+# asm 2: movdqa <z4=%xmm14,>y12=%xmm1
+movdqa %xmm14,%xmm1
+
+# qhasm: uint32323232	y12 += z8
+# asm 1: paddd <z8=int6464#16,<y12=int6464#2
+# asm 2: paddd <z8=%xmm15,<y12=%xmm1
+paddd %xmm15,%xmm1
+
+# qhasm: 		r12 = y12
+# asm 1: movdqa <y12=int6464#2,>r12=int6464#3
+# asm 2: movdqa <y12=%xmm1,>r12=%xmm2
+movdqa %xmm1,%xmm2
+
+# qhasm: uint32323232	y12 <<= 13
+# asm 1: pslld $13,<y12=int6464#2
+# asm 2: pslld $13,<y12=%xmm1
+pslld $13,%xmm1
+
+# qhasm: 		z12 ^= y12
+# asm 1: pxor  <y12=int6464#2,<z12=int6464#14
+# asm 2: pxor  <y12=%xmm1,<z12=%xmm13
+pxor  %xmm1,%xmm13
+
+# qhasm: uint32323232	r12 >>= 19
+# asm 1: psrld $19,<r12=int6464#3
+# asm 2: psrld $19,<r12=%xmm2
+psrld $19,%xmm2
+
+# qhasm: 		z12 ^= r12
+# asm 1: pxor  <r12=int6464#3,<z12=int6464#14
+# asm 2: pxor  <r12=%xmm2,<z12=%xmm13
+pxor  %xmm2,%xmm13
+
+# qhasm: 				y1 = z9
+# asm 1: movdqa <z9=int6464#12,>y1=int6464#2
+# asm 2: movdqa <z9=%xmm11,>y1=%xmm1
+movdqa %xmm11,%xmm1
+
+# qhasm: uint32323232			y1 += z13
+# asm 1: paddd <z13=int6464#10,<y1=int6464#2
+# asm 2: paddd <z13=%xmm9,<y1=%xmm1
+paddd %xmm9,%xmm1
+
+# qhasm: 				r1 = y1
+# asm 1: movdqa <y1=int6464#2,>r1=int6464#3
+# asm 2: movdqa <y1=%xmm1,>r1=%xmm2
+movdqa %xmm1,%xmm2
+
+# qhasm: uint32323232			y1 <<= 13
+# asm 1: pslld $13,<y1=int6464#2
+# asm 2: pslld $13,<y1=%xmm1
+pslld $13,%xmm1
+
+# qhasm: 				z1 ^= y1
+# asm 1: pxor  <y1=int6464#2,<z1=int6464#8
+# asm 2: pxor  <y1=%xmm1,<z1=%xmm7
+pxor  %xmm1,%xmm7
+
+# qhasm: uint32323232			r1 >>= 19
+# asm 1: psrld $19,<r1=int6464#3
+# asm 2: psrld $19,<r1=%xmm2
+psrld $19,%xmm2
+
+# qhasm: 				z1 ^= r1
+# asm 1: pxor  <r1=int6464#3,<z1=int6464#8
+# asm 2: pxor  <r1=%xmm2,<z1=%xmm7
+pxor  %xmm2,%xmm7
+
+# qhasm: 		y0 = z8
+# asm 1: movdqa <z8=int6464#16,>y0=int6464#2
+# asm 2: movdqa <z8=%xmm15,>y0=%xmm1
+movdqa %xmm15,%xmm1
+
+# qhasm: uint32323232	y0 += z12
+# asm 1: paddd <z12=int6464#14,<y0=int6464#2
+# asm 2: paddd <z12=%xmm13,<y0=%xmm1
+paddd %xmm13,%xmm1
+
+# qhasm: 		r0 = y0
+# asm 1: movdqa <y0=int6464#2,>r0=int6464#3
+# asm 2: movdqa <y0=%xmm1,>r0=%xmm2
+movdqa %xmm1,%xmm2
+
+# qhasm: uint32323232	y0 <<= 18
+# asm 1: pslld $18,<y0=int6464#2
+# asm 2: pslld $18,<y0=%xmm1
+pslld $18,%xmm1
+
+# qhasm: 		z0 ^= y0
+# asm 1: pxor  <y0=int6464#2,<z0=int6464#13
+# asm 2: pxor  <y0=%xmm1,<z0=%xmm12
+pxor  %xmm1,%xmm12
+
+# qhasm: uint32323232	r0 >>= 14
+# asm 1: psrld $14,<r0=int6464#3
+# asm 2: psrld $14,<r0=%xmm2
+psrld $14,%xmm2
+
+# qhasm: 		z0 ^= r0
+# asm 1: pxor  <r0=int6464#3,<z0=int6464#13
+# asm 2: pxor  <r0=%xmm2,<z0=%xmm12
+pxor  %xmm2,%xmm12
+
+# qhasm: 						z10 = z10_stack
+# asm 1: movdqa <z10_stack=stack128#21,>z10=int6464#2
+# asm 2: movdqa <z10_stack=320(%rsp),>z10=%xmm1
+movdqa 320(%rsp),%xmm1
+
+# qhasm: 		z0_stack = z0
+# asm 1: movdqa <z0=int6464#13,>z0_stack=stack128#21
+# asm 2: movdqa <z0=%xmm12,>z0_stack=320(%rsp)
+movdqa %xmm12,320(%rsp)
+
+# qhasm: 				y5 = z13
+# asm 1: movdqa <z13=int6464#10,>y5=int6464#3
+# asm 2: movdqa <z13=%xmm9,>y5=%xmm2
+movdqa %xmm9,%xmm2
+
+# qhasm: uint32323232			y5 += z1
+# asm 1: paddd <z1=int6464#8,<y5=int6464#3
+# asm 2: paddd <z1=%xmm7,<y5=%xmm2
+paddd %xmm7,%xmm2
+
+# qhasm: 				r5 = y5
+# asm 1: movdqa <y5=int6464#3,>r5=int6464#13
+# asm 2: movdqa <y5=%xmm2,>r5=%xmm12
+movdqa %xmm2,%xmm12
+
+# qhasm: uint32323232			y5 <<= 18
+# asm 1: pslld $18,<y5=int6464#3
+# asm 2: pslld $18,<y5=%xmm2
+pslld $18,%xmm2
+
+# qhasm: 				z5 ^= y5
+# asm 1: pxor  <y5=int6464#3,<z5=int6464#1
+# asm 2: pxor  <y5=%xmm2,<z5=%xmm0
+pxor  %xmm2,%xmm0
+
+# qhasm: uint32323232			r5 >>= 14
+# asm 1: psrld $14,<r5=int6464#13
+# asm 2: psrld $14,<r5=%xmm12
+psrld $14,%xmm12
+
+# qhasm: 				z5 ^= r5
+# asm 1: pxor  <r5=int6464#13,<z5=int6464#1
+# asm 2: pxor  <r5=%xmm12,<z5=%xmm0
+pxor  %xmm12,%xmm0
+
+# qhasm: 						y14 = z6
+# asm 1: movdqa <z6=int6464#6,>y14=int6464#3
+# asm 2: movdqa <z6=%xmm5,>y14=%xmm2
+movdqa %xmm5,%xmm2
+
+# qhasm: uint32323232					y14 += z10
+# asm 1: paddd <z10=int6464#2,<y14=int6464#3
+# asm 2: paddd <z10=%xmm1,<y14=%xmm2
+paddd %xmm1,%xmm2
+
+# qhasm: 						r14 = y14
+# asm 1: movdqa <y14=int6464#3,>r14=int6464#13
+# asm 2: movdqa <y14=%xmm2,>r14=%xmm12
+movdqa %xmm2,%xmm12
+
+# qhasm: uint32323232					y14 <<= 7
+# asm 1: pslld $7,<y14=int6464#3
+# asm 2: pslld $7,<y14=%xmm2
+pslld $7,%xmm2
+
+# qhasm: 						z14 ^= y14
+# asm 1: pxor  <y14=int6464#3,<z14=int6464#4
+# asm 2: pxor  <y14=%xmm2,<z14=%xmm3
+pxor  %xmm2,%xmm3
+
+# qhasm: uint32323232					r14 >>= 25
+# asm 1: psrld $25,<r14=int6464#13
+# asm 2: psrld $25,<r14=%xmm12
+psrld $25,%xmm12
+
+# qhasm: 						z14 ^= r14
+# asm 1: pxor  <r14=int6464#13,<z14=int6464#4
+# asm 2: pxor  <r14=%xmm12,<z14=%xmm3
+pxor  %xmm12,%xmm3
+
+# qhasm: 								z15 = z15_stack
+# asm 1: movdqa <z15_stack=stack128#22,>z15=int6464#3
+# asm 2: movdqa <z15_stack=336(%rsp),>z15=%xmm2
+movdqa 336(%rsp),%xmm2
+
+# qhasm: 				z5_stack = z5
+# asm 1: movdqa <z5=int6464#1,>z5_stack=stack128#22
+# asm 2: movdqa <z5=%xmm0,>z5_stack=336(%rsp)
+movdqa %xmm0,336(%rsp)
+
+# qhasm: 								y3 = z11
+# asm 1: movdqa <z11=int6464#7,>y3=int6464#1
+# asm 2: movdqa <z11=%xmm6,>y3=%xmm0
+movdqa %xmm6,%xmm0
+
+# qhasm: uint32323232							y3 += z15
+# asm 1: paddd <z15=int6464#3,<y3=int6464#1
+# asm 2: paddd <z15=%xmm2,<y3=%xmm0
+paddd %xmm2,%xmm0
+
+# qhasm: 								r3 = y3
+# asm 1: movdqa <y3=int6464#1,>r3=int6464#13
+# asm 2: movdqa <y3=%xmm0,>r3=%xmm12
+movdqa %xmm0,%xmm12
+
+# qhasm: uint32323232							y3 <<= 7
+# asm 1: pslld $7,<y3=int6464#1
+# asm 2: pslld $7,<y3=%xmm0
+pslld $7,%xmm0
+
+# qhasm: 								z3 ^= y3
+# asm 1: pxor  <y3=int6464#1,<z3=int6464#5
+# asm 2: pxor  <y3=%xmm0,<z3=%xmm4
+pxor  %xmm0,%xmm4
+
+# qhasm: uint32323232							r3 >>= 25
+# asm 1: psrld $25,<r3=int6464#13
+# asm 2: psrld $25,<r3=%xmm12
+psrld $25,%xmm12
+
+# qhasm: 								z3 ^= r3
+# asm 1: pxor  <r3=int6464#13,<z3=int6464#5
+# asm 2: pxor  <r3=%xmm12,<z3=%xmm4
+pxor  %xmm12,%xmm4
+
+# qhasm: 						y2 = z10
+# asm 1: movdqa <z10=int6464#2,>y2=int6464#1
+# asm 2: movdqa <z10=%xmm1,>y2=%xmm0
+movdqa %xmm1,%xmm0
+
+# qhasm: uint32323232					y2 += z14
+# asm 1: paddd <z14=int6464#4,<y2=int6464#1
+# asm 2: paddd <z14=%xmm3,<y2=%xmm0
+paddd %xmm3,%xmm0
+
+# qhasm: 						r2 = y2
+# asm 1: movdqa <y2=int6464#1,>r2=int6464#13
+# asm 2: movdqa <y2=%xmm0,>r2=%xmm12
+movdqa %xmm0,%xmm12
+
+# qhasm: uint32323232					y2 <<= 9
+# asm 1: pslld $9,<y2=int6464#1
+# asm 2: pslld $9,<y2=%xmm0
+pslld $9,%xmm0
+
+# qhasm: 						z2 ^= y2
+# asm 1: pxor  <y2=int6464#1,<z2=int6464#11
+# asm 2: pxor  <y2=%xmm0,<z2=%xmm10
+pxor  %xmm0,%xmm10
+
+# qhasm: uint32323232					r2 >>= 23
+# asm 1: psrld $23,<r2=int6464#13
+# asm 2: psrld $23,<r2=%xmm12
+psrld $23,%xmm12
+
+# qhasm: 						z2 ^= r2
+# asm 1: pxor  <r2=int6464#13,<z2=int6464#11
+# asm 2: pxor  <r2=%xmm12,<z2=%xmm10
+pxor  %xmm12,%xmm10
+
+# qhasm: 								y7 = z15
+# asm 1: movdqa <z15=int6464#3,>y7=int6464#1
+# asm 2: movdqa <z15=%xmm2,>y7=%xmm0
+movdqa %xmm2,%xmm0
+
+# qhasm: uint32323232							y7 += z3
+# asm 1: paddd <z3=int6464#5,<y7=int6464#1
+# asm 2: paddd <z3=%xmm4,<y7=%xmm0
+paddd %xmm4,%xmm0
+
+# qhasm: 								r7 = y7
+# asm 1: movdqa <y7=int6464#1,>r7=int6464#13
+# asm 2: movdqa <y7=%xmm0,>r7=%xmm12
+movdqa %xmm0,%xmm12
+
+# qhasm: uint32323232							y7 <<= 9
+# asm 1: pslld $9,<y7=int6464#1
+# asm 2: pslld $9,<y7=%xmm0
+pslld $9,%xmm0
+
+# qhasm: 								z7 ^= y7
+# asm 1: pxor  <y7=int6464#1,<z7=int6464#9
+# asm 2: pxor  <y7=%xmm0,<z7=%xmm8
+pxor  %xmm0,%xmm8
+
+# qhasm: uint32323232							r7 >>= 23
+# asm 1: psrld $23,<r7=int6464#13
+# asm 2: psrld $23,<r7=%xmm12
+psrld $23,%xmm12
+
+# qhasm: 								z7 ^= r7
+# asm 1: pxor  <r7=int6464#13,<z7=int6464#9
+# asm 2: pxor  <r7=%xmm12,<z7=%xmm8
+pxor  %xmm12,%xmm8
+
+# qhasm: 						y6 = z14
+# asm 1: movdqa <z14=int6464#4,>y6=int6464#1
+# asm 2: movdqa <z14=%xmm3,>y6=%xmm0
+movdqa %xmm3,%xmm0
+
+# qhasm: uint32323232					y6 += z2
+# asm 1: paddd <z2=int6464#11,<y6=int6464#1
+# asm 2: paddd <z2=%xmm10,<y6=%xmm0
+paddd %xmm10,%xmm0
+
+# qhasm: 						r6 = y6
+# asm 1: movdqa <y6=int6464#1,>r6=int6464#13
+# asm 2: movdqa <y6=%xmm0,>r6=%xmm12
+movdqa %xmm0,%xmm12
+
+# qhasm: uint32323232					y6 <<= 13
+# asm 1: pslld $13,<y6=int6464#1
+# asm 2: pslld $13,<y6=%xmm0
+pslld $13,%xmm0
+
+# qhasm: 						z6 ^= y6
+# asm 1: pxor  <y6=int6464#1,<z6=int6464#6
+# asm 2: pxor  <y6=%xmm0,<z6=%xmm5
+pxor  %xmm0,%xmm5
+
+# qhasm: uint32323232					r6 >>= 19
+# asm 1: psrld $19,<r6=int6464#13
+# asm 2: psrld $19,<r6=%xmm12
+psrld $19,%xmm12
+
+# qhasm: 						z6 ^= r6
+# asm 1: pxor  <r6=int6464#13,<z6=int6464#6
+# asm 2: pxor  <r6=%xmm12,<z6=%xmm5
+pxor  %xmm12,%xmm5
+
+# qhasm: 								y11 = z3
+# asm 1: movdqa <z3=int6464#5,>y11=int6464#1
+# asm 2: movdqa <z3=%xmm4,>y11=%xmm0
+movdqa %xmm4,%xmm0
+
+# qhasm: uint32323232							y11 += z7
+# asm 1: paddd <z7=int6464#9,<y11=int6464#1
+# asm 2: paddd <z7=%xmm8,<y11=%xmm0
+paddd %xmm8,%xmm0
+
+# qhasm: 								r11 = y11
+# asm 1: movdqa <y11=int6464#1,>r11=int6464#13
+# asm 2: movdqa <y11=%xmm0,>r11=%xmm12
+movdqa %xmm0,%xmm12
+
+# qhasm: uint32323232							y11 <<= 13
+# asm 1: pslld $13,<y11=int6464#1
+# asm 2: pslld $13,<y11=%xmm0
+pslld $13,%xmm0
+
+# qhasm: 								z11 ^= y11
+# asm 1: pxor  <y11=int6464#1,<z11=int6464#7
+# asm 2: pxor  <y11=%xmm0,<z11=%xmm6
+pxor  %xmm0,%xmm6
+
+# qhasm: uint32323232							r11 >>= 19
+# asm 1: psrld $19,<r11=int6464#13
+# asm 2: psrld $19,<r11=%xmm12
+psrld $19,%xmm12
+
+# qhasm: 								z11 ^= r11
+# asm 1: pxor  <r11=int6464#13,<z11=int6464#7
+# asm 2: pxor  <r11=%xmm12,<z11=%xmm6
+pxor  %xmm12,%xmm6
+
+# qhasm: 						y10 = z2
+# asm 1: movdqa <z2=int6464#11,>y10=int6464#1
+# asm 2: movdqa <z2=%xmm10,>y10=%xmm0
+movdqa %xmm10,%xmm0
+
+# qhasm: uint32323232					y10 += z6
+# asm 1: paddd <z6=int6464#6,<y10=int6464#1
+# asm 2: paddd <z6=%xmm5,<y10=%xmm0
+paddd %xmm5,%xmm0
+
+# qhasm: 						r10 = y10
+# asm 1: movdqa <y10=int6464#1,>r10=int6464#13
+# asm 2: movdqa <y10=%xmm0,>r10=%xmm12
+movdqa %xmm0,%xmm12
+
+# qhasm: uint32323232					y10 <<= 18
+# asm 1: pslld $18,<y10=int6464#1
+# asm 2: pslld $18,<y10=%xmm0
+pslld $18,%xmm0
+
+# qhasm: 						z10 ^= y10
+# asm 1: pxor  <y10=int6464#1,<z10=int6464#2
+# asm 2: pxor  <y10=%xmm0,<z10=%xmm1
+pxor  %xmm0,%xmm1
+
+# qhasm: uint32323232					r10 >>= 14
+# asm 1: psrld $14,<r10=int6464#13
+# asm 2: psrld $14,<r10=%xmm12
+psrld $14,%xmm12
+
+# qhasm: 						z10 ^= r10
+# asm 1: pxor  <r10=int6464#13,<z10=int6464#2
+# asm 2: pxor  <r10=%xmm12,<z10=%xmm1
+pxor  %xmm12,%xmm1
+
+# qhasm: 		z0 = z0_stack
+# asm 1: movdqa <z0_stack=stack128#21,>z0=int6464#1
+# asm 2: movdqa <z0_stack=320(%rsp),>z0=%xmm0
+movdqa 320(%rsp),%xmm0
+
+# qhasm: 						z10_stack = z10
+# asm 1: movdqa <z10=int6464#2,>z10_stack=stack128#21
+# asm 2: movdqa <z10=%xmm1,>z10_stack=320(%rsp)
+movdqa %xmm1,320(%rsp)
+
+# qhasm: 		y1 = z3
+# asm 1: movdqa <z3=int6464#5,>y1=int6464#2
+# asm 2: movdqa <z3=%xmm4,>y1=%xmm1
+movdqa %xmm4,%xmm1
+
+# qhasm: uint32323232	y1 += z0
+# asm 1: paddd <z0=int6464#1,<y1=int6464#2
+# asm 2: paddd <z0=%xmm0,<y1=%xmm1
+paddd %xmm0,%xmm1
+
+# qhasm: 		r1 = y1
+# asm 1: movdqa <y1=int6464#2,>r1=int6464#13
+# asm 2: movdqa <y1=%xmm1,>r1=%xmm12
+movdqa %xmm1,%xmm12
+
+# qhasm: uint32323232	y1 <<= 7
+# asm 1: pslld $7,<y1=int6464#2
+# asm 2: pslld $7,<y1=%xmm1
+pslld $7,%xmm1
+
+# qhasm: 		z1 ^= y1
+# asm 1: pxor  <y1=int6464#2,<z1=int6464#8
+# asm 2: pxor  <y1=%xmm1,<z1=%xmm7
+pxor  %xmm1,%xmm7
+
+# qhasm: uint32323232	r1 >>= 25
+# asm 1: psrld $25,<r1=int6464#13
+# asm 2: psrld $25,<r1=%xmm12
+psrld $25,%xmm12
+
+# qhasm: 		z1 ^= r1
+# asm 1: pxor  <r1=int6464#13,<z1=int6464#8
+# asm 2: pxor  <r1=%xmm12,<z1=%xmm7
+pxor  %xmm12,%xmm7
+
+# qhasm: 								y15 = z7
+# asm 1: movdqa <z7=int6464#9,>y15=int6464#2
+# asm 2: movdqa <z7=%xmm8,>y15=%xmm1
+movdqa %xmm8,%xmm1
+
+# qhasm: uint32323232							y15 += z11
+# asm 1: paddd <z11=int6464#7,<y15=int6464#2
+# asm 2: paddd <z11=%xmm6,<y15=%xmm1
+paddd %xmm6,%xmm1
+
+# qhasm: 								r15 = y15
+# asm 1: movdqa <y15=int6464#2,>r15=int6464#13
+# asm 2: movdqa <y15=%xmm1,>r15=%xmm12
+movdqa %xmm1,%xmm12
+
+# qhasm: uint32323232							y15 <<= 18
+# asm 1: pslld $18,<y15=int6464#2
+# asm 2: pslld $18,<y15=%xmm1
+pslld $18,%xmm1
+
+# qhasm: 								z15 ^= y15
+# asm 1: pxor  <y15=int6464#2,<z15=int6464#3
+# asm 2: pxor  <y15=%xmm1,<z15=%xmm2
+pxor  %xmm1,%xmm2
+
+# qhasm: uint32323232							r15 >>= 14
+# asm 1: psrld $14,<r15=int6464#13
+# asm 2: psrld $14,<r15=%xmm12
+psrld $14,%xmm12
+
+# qhasm: 								z15 ^= r15
+# asm 1: pxor  <r15=int6464#13,<z15=int6464#3
+# asm 2: pxor  <r15=%xmm12,<z15=%xmm2
+pxor  %xmm12,%xmm2
+
+# qhasm: 				z5 = z5_stack
+# asm 1: movdqa <z5_stack=stack128#22,>z5=int6464#13
+# asm 2: movdqa <z5_stack=336(%rsp),>z5=%xmm12
+movdqa 336(%rsp),%xmm12
+
+# qhasm: 								z15_stack = z15
+# asm 1: movdqa <z15=int6464#3,>z15_stack=stack128#22
+# asm 2: movdqa <z15=%xmm2,>z15_stack=336(%rsp)
+movdqa %xmm2,336(%rsp)
+
+# qhasm: 				y6 = z4
+# asm 1: movdqa <z4=int6464#15,>y6=int6464#2
+# asm 2: movdqa <z4=%xmm14,>y6=%xmm1
+movdqa %xmm14,%xmm1
+
+# qhasm: uint32323232			y6 += z5
+# asm 1: paddd <z5=int6464#13,<y6=int6464#2
+# asm 2: paddd <z5=%xmm12,<y6=%xmm1
+paddd %xmm12,%xmm1
+
+# qhasm: 				r6 = y6
+# asm 1: movdqa <y6=int6464#2,>r6=int6464#3
+# asm 2: movdqa <y6=%xmm1,>r6=%xmm2
+movdqa %xmm1,%xmm2
+
+# qhasm: uint32323232			y6 <<= 7
+# asm 1: pslld $7,<y6=int6464#2
+# asm 2: pslld $7,<y6=%xmm1
+pslld $7,%xmm1
+
+# qhasm: 				z6 ^= y6
+# asm 1: pxor  <y6=int6464#2,<z6=int6464#6
+# asm 2: pxor  <y6=%xmm1,<z6=%xmm5
+pxor  %xmm1,%xmm5
+
+# qhasm: uint32323232			r6 >>= 25
+# asm 1: psrld $25,<r6=int6464#3
+# asm 2: psrld $25,<r6=%xmm2
+psrld $25,%xmm2
+
+# qhasm: 				z6 ^= r6
+# asm 1: pxor  <r6=int6464#3,<z6=int6464#6
+# asm 2: pxor  <r6=%xmm2,<z6=%xmm5
+pxor  %xmm2,%xmm5
+
+# qhasm: 		y2 = z0
+# asm 1: movdqa <z0=int6464#1,>y2=int6464#2
+# asm 2: movdqa <z0=%xmm0,>y2=%xmm1
+movdqa %xmm0,%xmm1
+
+# qhasm: uint32323232	y2 += z1
+# asm 1: paddd <z1=int6464#8,<y2=int6464#2
+# asm 2: paddd <z1=%xmm7,<y2=%xmm1
+paddd %xmm7,%xmm1
+
+# qhasm: 		r2 = y2
+# asm 1: movdqa <y2=int6464#2,>r2=int6464#3
+# asm 2: movdqa <y2=%xmm1,>r2=%xmm2
+movdqa %xmm1,%xmm2
+
+# qhasm: uint32323232	y2 <<= 9
+# asm 1: pslld $9,<y2=int6464#2
+# asm 2: pslld $9,<y2=%xmm1
+pslld $9,%xmm1
+
+# qhasm: 		z2 ^= y2
+# asm 1: pxor  <y2=int6464#2,<z2=int6464#11
+# asm 2: pxor  <y2=%xmm1,<z2=%xmm10
+pxor  %xmm1,%xmm10
+
+# qhasm: uint32323232	r2 >>= 23
+# asm 1: psrld $23,<r2=int6464#3
+# asm 2: psrld $23,<r2=%xmm2
+psrld $23,%xmm2
+
+# qhasm: 		z2 ^= r2
+# asm 1: pxor  <r2=int6464#3,<z2=int6464#11
+# asm 2: pxor  <r2=%xmm2,<z2=%xmm10
+pxor  %xmm2,%xmm10
+
+# qhasm: 				y7 = z5
+# asm 1: movdqa <z5=int6464#13,>y7=int6464#2
+# asm 2: movdqa <z5=%xmm12,>y7=%xmm1
+movdqa %xmm12,%xmm1
+
+# qhasm: uint32323232			y7 += z6
+# asm 1: paddd <z6=int6464#6,<y7=int6464#2
+# asm 2: paddd <z6=%xmm5,<y7=%xmm1
+paddd %xmm5,%xmm1
+
+# qhasm: 				r7 = y7
+# asm 1: movdqa <y7=int6464#2,>r7=int6464#3
+# asm 2: movdqa <y7=%xmm1,>r7=%xmm2
+movdqa %xmm1,%xmm2
+
+# qhasm: uint32323232			y7 <<= 9
+# asm 1: pslld $9,<y7=int6464#2
+# asm 2: pslld $9,<y7=%xmm1
+pslld $9,%xmm1
+
+# qhasm: 				z7 ^= y7
+# asm 1: pxor  <y7=int6464#2,<z7=int6464#9
+# asm 2: pxor  <y7=%xmm1,<z7=%xmm8
+pxor  %xmm1,%xmm8
+
+# qhasm: uint32323232			r7 >>= 23
+# asm 1: psrld $23,<r7=int6464#3
+# asm 2: psrld $23,<r7=%xmm2
+psrld $23,%xmm2
+
+# qhasm: 				z7 ^= r7
+# asm 1: pxor  <r7=int6464#3,<z7=int6464#9
+# asm 2: pxor  <r7=%xmm2,<z7=%xmm8
+pxor  %xmm2,%xmm8
+
+# qhasm: 		y3 = z1
+# asm 1: movdqa <z1=int6464#8,>y3=int6464#2
+# asm 2: movdqa <z1=%xmm7,>y3=%xmm1
+movdqa %xmm7,%xmm1
+
+# qhasm: uint32323232	y3 += z2
+# asm 1: paddd <z2=int6464#11,<y3=int6464#2
+# asm 2: paddd <z2=%xmm10,<y3=%xmm1
+paddd %xmm10,%xmm1
+
+# qhasm: 		r3 = y3
+# asm 1: movdqa <y3=int6464#2,>r3=int6464#3
+# asm 2: movdqa <y3=%xmm1,>r3=%xmm2
+movdqa %xmm1,%xmm2
+
+# qhasm: uint32323232	y3 <<= 13
+# asm 1: pslld $13,<y3=int6464#2
+# asm 2: pslld $13,<y3=%xmm1
+pslld $13,%xmm1
+
+# qhasm: 		z3 ^= y3
+# asm 1: pxor  <y3=int6464#2,<z3=int6464#5
+# asm 2: pxor  <y3=%xmm1,<z3=%xmm4
+pxor  %xmm1,%xmm4
+
+# qhasm: uint32323232	r3 >>= 19
+# asm 1: psrld $19,<r3=int6464#3
+# asm 2: psrld $19,<r3=%xmm2
+psrld $19,%xmm2
+
+# qhasm: 		z3 ^= r3
+# asm 1: pxor  <r3=int6464#3,<z3=int6464#5
+# asm 2: pxor  <r3=%xmm2,<z3=%xmm4
+pxor  %xmm2,%xmm4
+
+# qhasm: 				y4 = z6
+# asm 1: movdqa <z6=int6464#6,>y4=int6464#2
+# asm 2: movdqa <z6=%xmm5,>y4=%xmm1
+movdqa %xmm5,%xmm1
+
+# qhasm: uint32323232			y4 += z7
+# asm 1: paddd <z7=int6464#9,<y4=int6464#2
+# asm 2: paddd <z7=%xmm8,<y4=%xmm1
+paddd %xmm8,%xmm1
+
+# qhasm: 				r4 = y4
+# asm 1: movdqa <y4=int6464#2,>r4=int6464#3
+# asm 2: movdqa <y4=%xmm1,>r4=%xmm2
+movdqa %xmm1,%xmm2
+
+# qhasm: uint32323232			y4 <<= 13
+# asm 1: pslld $13,<y4=int6464#2
+# asm 2: pslld $13,<y4=%xmm1
+pslld $13,%xmm1
+
+# qhasm: 				z4 ^= y4
+# asm 1: pxor  <y4=int6464#2,<z4=int6464#15
+# asm 2: pxor  <y4=%xmm1,<z4=%xmm14
+pxor  %xmm1,%xmm14
+
+# qhasm: uint32323232			r4 >>= 19
+# asm 1: psrld $19,<r4=int6464#3
+# asm 2: psrld $19,<r4=%xmm2
+psrld $19,%xmm2
+
+# qhasm: 				z4 ^= r4
+# asm 1: pxor  <r4=int6464#3,<z4=int6464#15
+# asm 2: pxor  <r4=%xmm2,<z4=%xmm14
+pxor  %xmm2,%xmm14
+
+# qhasm: 		y0 = z2
+# asm 1: movdqa <z2=int6464#11,>y0=int6464#2
+# asm 2: movdqa <z2=%xmm10,>y0=%xmm1
+movdqa %xmm10,%xmm1
+
+# qhasm: uint32323232	y0 += z3
+# asm 1: paddd <z3=int6464#5,<y0=int6464#2
+# asm 2: paddd <z3=%xmm4,<y0=%xmm1
+paddd %xmm4,%xmm1
+
+# qhasm: 		r0 = y0
+# asm 1: movdqa <y0=int6464#2,>r0=int6464#3
+# asm 2: movdqa <y0=%xmm1,>r0=%xmm2
+movdqa %xmm1,%xmm2
+
+# qhasm: uint32323232	y0 <<= 18
+# asm 1: pslld $18,<y0=int6464#2
+# asm 2: pslld $18,<y0=%xmm1
+pslld $18,%xmm1
+
+# qhasm: 		z0 ^= y0
+# asm 1: pxor  <y0=int6464#2,<z0=int6464#1
+# asm 2: pxor  <y0=%xmm1,<z0=%xmm0
+pxor  %xmm1,%xmm0
+
+# qhasm: uint32323232	r0 >>= 14
+# asm 1: psrld $14,<r0=int6464#3
+# asm 2: psrld $14,<r0=%xmm2
+psrld $14,%xmm2
+
+# qhasm: 		z0 ^= r0
+# asm 1: pxor  <r0=int6464#3,<z0=int6464#1
+# asm 2: pxor  <r0=%xmm2,<z0=%xmm0
+pxor  %xmm2,%xmm0
+
+# qhasm: 						z10 = z10_stack
+# asm 1: movdqa <z10_stack=stack128#21,>z10=int6464#2
+# asm 2: movdqa <z10_stack=320(%rsp),>z10=%xmm1
+movdqa 320(%rsp),%xmm1
+
+# qhasm: 		z0_stack = z0
+# asm 1: movdqa <z0=int6464#1,>z0_stack=stack128#21
+# asm 2: movdqa <z0=%xmm0,>z0_stack=320(%rsp)
+movdqa %xmm0,320(%rsp)
+
+# qhasm: 				y5 = z7
+# asm 1: movdqa <z7=int6464#9,>y5=int6464#1
+# asm 2: movdqa <z7=%xmm8,>y5=%xmm0
+movdqa %xmm8,%xmm0
+
+# qhasm: uint32323232			y5 += z4
+# asm 1: paddd <z4=int6464#15,<y5=int6464#1
+# asm 2: paddd <z4=%xmm14,<y5=%xmm0
+paddd %xmm14,%xmm0
+
+# qhasm: 				r5 = y5
+# asm 1: movdqa <y5=int6464#1,>r5=int6464#3
+# asm 2: movdqa <y5=%xmm0,>r5=%xmm2
+movdqa %xmm0,%xmm2
+
+# qhasm: uint32323232			y5 <<= 18
+# asm 1: pslld $18,<y5=int6464#1
+# asm 2: pslld $18,<y5=%xmm0
+pslld $18,%xmm0
+
+# qhasm: 				z5 ^= y5
+# asm 1: pxor  <y5=int6464#1,<z5=int6464#13
+# asm 2: pxor  <y5=%xmm0,<z5=%xmm12
+pxor  %xmm0,%xmm12
+
+# qhasm: uint32323232			r5 >>= 14
+# asm 1: psrld $14,<r5=int6464#3
+# asm 2: psrld $14,<r5=%xmm2
+psrld $14,%xmm2
+
+# qhasm: 				z5 ^= r5
+# asm 1: pxor  <r5=int6464#3,<z5=int6464#13
+# asm 2: pxor  <r5=%xmm2,<z5=%xmm12
+pxor  %xmm2,%xmm12
+
+# qhasm: 						y11 = z9
+# asm 1: movdqa <z9=int6464#12,>y11=int6464#1
+# asm 2: movdqa <z9=%xmm11,>y11=%xmm0
+movdqa %xmm11,%xmm0
+
+# qhasm: uint32323232					y11 += z10
+# asm 1: paddd <z10=int6464#2,<y11=int6464#1
+# asm 2: paddd <z10=%xmm1,<y11=%xmm0
+paddd %xmm1,%xmm0
+
+# qhasm: 						r11 = y11
+# asm 1: movdqa <y11=int6464#1,>r11=int6464#3
+# asm 2: movdqa <y11=%xmm0,>r11=%xmm2
+movdqa %xmm0,%xmm2
+
+# qhasm: uint32323232					y11 <<= 7
+# asm 1: pslld $7,<y11=int6464#1
+# asm 2: pslld $7,<y11=%xmm0
+pslld $7,%xmm0
+
+# qhasm: 						z11 ^= y11
+# asm 1: pxor  <y11=int6464#1,<z11=int6464#7
+# asm 2: pxor  <y11=%xmm0,<z11=%xmm6
+pxor  %xmm0,%xmm6
+
+# qhasm: uint32323232					r11 >>= 25
+# asm 1: psrld $25,<r11=int6464#3
+# asm 2: psrld $25,<r11=%xmm2
+psrld $25,%xmm2
+
+# qhasm: 						z11 ^= r11
+# asm 1: pxor  <r11=int6464#3,<z11=int6464#7
+# asm 2: pxor  <r11=%xmm2,<z11=%xmm6
+pxor  %xmm2,%xmm6
+
+# qhasm: 								z15 = z15_stack
+# asm 1: movdqa <z15_stack=stack128#22,>z15=int6464#3
+# asm 2: movdqa <z15_stack=336(%rsp),>z15=%xmm2
+movdqa 336(%rsp),%xmm2
+
+# qhasm: 				z5_stack = z5
+# asm 1: movdqa <z5=int6464#13,>z5_stack=stack128#22
+# asm 2: movdqa <z5=%xmm12,>z5_stack=336(%rsp)
+movdqa %xmm12,336(%rsp)
+
+# qhasm: 								y12 = z14
+# asm 1: movdqa <z14=int6464#4,>y12=int6464#1
+# asm 2: movdqa <z14=%xmm3,>y12=%xmm0
+movdqa %xmm3,%xmm0
+
+# qhasm: uint32323232							y12 += z15
+# asm 1: paddd <z15=int6464#3,<y12=int6464#1
+# asm 2: paddd <z15=%xmm2,<y12=%xmm0
+paddd %xmm2,%xmm0
+
+# qhasm: 								r12 = y12
+# asm 1: movdqa <y12=int6464#1,>r12=int6464#13
+# asm 2: movdqa <y12=%xmm0,>r12=%xmm12
+movdqa %xmm0,%xmm12
+
+# qhasm: uint32323232							y12 <<= 7
+# asm 1: pslld $7,<y12=int6464#1
+# asm 2: pslld $7,<y12=%xmm0
+pslld $7,%xmm0
+
+# qhasm: 								z12 ^= y12
+# asm 1: pxor  <y12=int6464#1,<z12=int6464#14
+# asm 2: pxor  <y12=%xmm0,<z12=%xmm13
+pxor  %xmm0,%xmm13
+
+# qhasm: uint32323232							r12 >>= 25
+# asm 1: psrld $25,<r12=int6464#13
+# asm 2: psrld $25,<r12=%xmm12
+psrld $25,%xmm12
+
+# qhasm: 								z12 ^= r12
+# asm 1: pxor  <r12=int6464#13,<z12=int6464#14
+# asm 2: pxor  <r12=%xmm12,<z12=%xmm13
+pxor  %xmm12,%xmm13
+
+# qhasm: 						y8 = z10
+# asm 1: movdqa <z10=int6464#2,>y8=int6464#1
+# asm 2: movdqa <z10=%xmm1,>y8=%xmm0
+movdqa %xmm1,%xmm0
+
+# qhasm: uint32323232					y8 += z11
+# asm 1: paddd <z11=int6464#7,<y8=int6464#1
+# asm 2: paddd <z11=%xmm6,<y8=%xmm0
+paddd %xmm6,%xmm0
+
+# qhasm: 						r8 = y8
+# asm 1: movdqa <y8=int6464#1,>r8=int6464#13
+# asm 2: movdqa <y8=%xmm0,>r8=%xmm12
+movdqa %xmm0,%xmm12
+
+# qhasm: uint32323232					y8 <<= 9
+# asm 1: pslld $9,<y8=int6464#1
+# asm 2: pslld $9,<y8=%xmm0
+pslld $9,%xmm0
+
+# qhasm: 						z8 ^= y8
+# asm 1: pxor  <y8=int6464#1,<z8=int6464#16
+# asm 2: pxor  <y8=%xmm0,<z8=%xmm15
+pxor  %xmm0,%xmm15
+
+# qhasm: uint32323232					r8 >>= 23
+# asm 1: psrld $23,<r8=int6464#13
+# asm 2: psrld $23,<r8=%xmm12
+psrld $23,%xmm12
+
+# qhasm: 						z8 ^= r8
+# asm 1: pxor  <r8=int6464#13,<z8=int6464#16
+# asm 2: pxor  <r8=%xmm12,<z8=%xmm15
+pxor  %xmm12,%xmm15
+
+# qhasm: 								y13 = z15
+# asm 1: movdqa <z15=int6464#3,>y13=int6464#1
+# asm 2: movdqa <z15=%xmm2,>y13=%xmm0
+movdqa %xmm2,%xmm0
+
+# qhasm: uint32323232							y13 += z12
+# asm 1: paddd <z12=int6464#14,<y13=int6464#1
+# asm 2: paddd <z12=%xmm13,<y13=%xmm0
+paddd %xmm13,%xmm0
+
+# qhasm: 								r13 = y13
+# asm 1: movdqa <y13=int6464#1,>r13=int6464#13
+# asm 2: movdqa <y13=%xmm0,>r13=%xmm12
+movdqa %xmm0,%xmm12
+
+# qhasm: uint32323232							y13 <<= 9
+# asm 1: pslld $9,<y13=int6464#1
+# asm 2: pslld $9,<y13=%xmm0
+pslld $9,%xmm0
+
+# qhasm: 								z13 ^= y13
+# asm 1: pxor  <y13=int6464#1,<z13=int6464#10
+# asm 2: pxor  <y13=%xmm0,<z13=%xmm9
+pxor  %xmm0,%xmm9
+
+# qhasm: uint32323232							r13 >>= 23
+# asm 1: psrld $23,<r13=int6464#13
+# asm 2: psrld $23,<r13=%xmm12
+psrld $23,%xmm12
+
+# qhasm: 								z13 ^= r13
+# asm 1: pxor  <r13=int6464#13,<z13=int6464#10
+# asm 2: pxor  <r13=%xmm12,<z13=%xmm9
+pxor  %xmm12,%xmm9
+
+# qhasm: 						y9 = z11
+# asm 1: movdqa <z11=int6464#7,>y9=int6464#1
+# asm 2: movdqa <z11=%xmm6,>y9=%xmm0
+movdqa %xmm6,%xmm0
+
+# qhasm: uint32323232					y9 += z8
+# asm 1: paddd <z8=int6464#16,<y9=int6464#1
+# asm 2: paddd <z8=%xmm15,<y9=%xmm0
+paddd %xmm15,%xmm0
+
+# qhasm: 						r9 = y9
+# asm 1: movdqa <y9=int6464#1,>r9=int6464#13
+# asm 2: movdqa <y9=%xmm0,>r9=%xmm12
+movdqa %xmm0,%xmm12
+
+# qhasm: uint32323232					y9 <<= 13
+# asm 1: pslld $13,<y9=int6464#1
+# asm 2: pslld $13,<y9=%xmm0
+pslld $13,%xmm0
+
+# qhasm: 						z9 ^= y9
+# asm 1: pxor  <y9=int6464#1,<z9=int6464#12
+# asm 2: pxor  <y9=%xmm0,<z9=%xmm11
+pxor  %xmm0,%xmm11
+
+# qhasm: uint32323232					r9 >>= 19
+# asm 1: psrld $19,<r9=int6464#13
+# asm 2: psrld $19,<r9=%xmm12
+psrld $19,%xmm12
+
+# qhasm: 						z9 ^= r9
+# asm 1: pxor  <r9=int6464#13,<z9=int6464#12
+# asm 2: pxor  <r9=%xmm12,<z9=%xmm11
+pxor  %xmm12,%xmm11
+
+# qhasm: 								y14 = z12
+# asm 1: movdqa <z12=int6464#14,>y14=int6464#1
+# asm 2: movdqa <z12=%xmm13,>y14=%xmm0
+movdqa %xmm13,%xmm0
+
+# qhasm: uint32323232							y14 += z13
+# asm 1: paddd <z13=int6464#10,<y14=int6464#1
+# asm 2: paddd <z13=%xmm9,<y14=%xmm0
+paddd %xmm9,%xmm0
+
+# qhasm: 								r14 = y14
+# asm 1: movdqa <y14=int6464#1,>r14=int6464#13
+# asm 2: movdqa <y14=%xmm0,>r14=%xmm12
+movdqa %xmm0,%xmm12
+
+# qhasm: uint32323232							y14 <<= 13
+# asm 1: pslld $13,<y14=int6464#1
+# asm 2: pslld $13,<y14=%xmm0
+pslld $13,%xmm0
+
+# qhasm: 								z14 ^= y14
+# asm 1: pxor  <y14=int6464#1,<z14=int6464#4
+# asm 2: pxor  <y14=%xmm0,<z14=%xmm3
+pxor  %xmm0,%xmm3
+
+# qhasm: uint32323232							r14 >>= 19
+# asm 1: psrld $19,<r14=int6464#13
+# asm 2: psrld $19,<r14=%xmm12
+psrld $19,%xmm12
+
+# qhasm: 								z14 ^= r14
+# asm 1: pxor  <r14=int6464#13,<z14=int6464#4
+# asm 2: pxor  <r14=%xmm12,<z14=%xmm3
+pxor  %xmm12,%xmm3
+
+# qhasm: 						y10 = z8
+# asm 1: movdqa <z8=int6464#16,>y10=int6464#1
+# asm 2: movdqa <z8=%xmm15,>y10=%xmm0
+movdqa %xmm15,%xmm0
+
+# qhasm: uint32323232					y10 += z9
+# asm 1: paddd <z9=int6464#12,<y10=int6464#1
+# asm 2: paddd <z9=%xmm11,<y10=%xmm0
+paddd %xmm11,%xmm0
+
+# qhasm: 						r10 = y10
+# asm 1: movdqa <y10=int6464#1,>r10=int6464#13
+# asm 2: movdqa <y10=%xmm0,>r10=%xmm12
+movdqa %xmm0,%xmm12
+
+# qhasm: uint32323232					y10 <<= 18
+# asm 1: pslld $18,<y10=int6464#1
+# asm 2: pslld $18,<y10=%xmm0
+pslld $18,%xmm0
+
+# qhasm: 						z10 ^= y10
+# asm 1: pxor  <y10=int6464#1,<z10=int6464#2
+# asm 2: pxor  <y10=%xmm0,<z10=%xmm1
+pxor  %xmm0,%xmm1
+
+# qhasm: uint32323232					r10 >>= 14
+# asm 1: psrld $14,<r10=int6464#13
+# asm 2: psrld $14,<r10=%xmm12
+psrld $14,%xmm12
+
+# qhasm: 						z10 ^= r10
+# asm 1: pxor  <r10=int6464#13,<z10=int6464#2
+# asm 2: pxor  <r10=%xmm12,<z10=%xmm1
+pxor  %xmm12,%xmm1
+
+# qhasm: 								y15 = z13
+# asm 1: movdqa <z13=int6464#10,>y15=int6464#1
+# asm 2: movdqa <z13=%xmm9,>y15=%xmm0
+movdqa %xmm9,%xmm0
+
+# qhasm: uint32323232							y15 += z14
+# asm 1: paddd <z14=int6464#4,<y15=int6464#1
+# asm 2: paddd <z14=%xmm3,<y15=%xmm0
+paddd %xmm3,%xmm0
+
+# qhasm: 								r15 = y15
+# asm 1: movdqa <y15=int6464#1,>r15=int6464#13
+# asm 2: movdqa <y15=%xmm0,>r15=%xmm12
+movdqa %xmm0,%xmm12
+
+# qhasm: uint32323232							y15 <<= 18
+# asm 1: pslld $18,<y15=int6464#1
+# asm 2: pslld $18,<y15=%xmm0
+pslld $18,%xmm0
+
+# qhasm: 								z15 ^= y15
+# asm 1: pxor  <y15=int6464#1,<z15=int6464#3
+# asm 2: pxor  <y15=%xmm0,<z15=%xmm2
+pxor  %xmm0,%xmm2
+
+# qhasm: uint32323232							r15 >>= 14
+# asm 1: psrld $14,<r15=int6464#13
+# asm 2: psrld $14,<r15=%xmm12
+psrld $14,%xmm12
+
+# qhasm: 								z15 ^= r15
+# asm 1: pxor  <r15=int6464#13,<z15=int6464#3
+# asm 2: pxor  <r15=%xmm12,<z15=%xmm2
+pxor  %xmm12,%xmm2
+
+# qhasm: 		z0 = z0_stack
+# asm 1: movdqa <z0_stack=stack128#21,>z0=int6464#13
+# asm 2: movdqa <z0_stack=320(%rsp),>z0=%xmm12
+movdqa 320(%rsp),%xmm12
+
+# qhasm: 				z5 = z5_stack
+# asm 1: movdqa <z5_stack=stack128#22,>z5=int6464#1
+# asm 2: movdqa <z5_stack=336(%rsp),>z5=%xmm0
+movdqa 336(%rsp),%xmm0
+
+# qhasm:                   unsigned>? i -= 2
+# asm 1: sub  $2,<i=int64#3
+# asm 2: sub  $2,<i=%rdx
+sub  $2,%rdx
+# comment:fp stack unchanged by jump
+
+# qhasm: goto mainloop1 if unsigned>
+ja ._mainloop1
+
+# qhasm:   uint32323232 z0 += orig0
+# asm 1: paddd <orig0=stack128#8,<z0=int6464#13
+# asm 2: paddd <orig0=112(%rsp),<z0=%xmm12
+paddd 112(%rsp),%xmm12
+
+# qhasm:   uint32323232 z1 += orig1
+# asm 1: paddd <orig1=stack128#12,<z1=int6464#8
+# asm 2: paddd <orig1=176(%rsp),<z1=%xmm7
+paddd 176(%rsp),%xmm7
+
+# qhasm:   uint32323232 z2 += orig2
+# asm 1: paddd <orig2=stack128#15,<z2=int6464#11
+# asm 2: paddd <orig2=224(%rsp),<z2=%xmm10
+paddd 224(%rsp),%xmm10
+
+# qhasm:   uint32323232 z3 += orig3
+# asm 1: paddd <orig3=stack128#18,<z3=int6464#5
+# asm 2: paddd <orig3=272(%rsp),<z3=%xmm4
+paddd 272(%rsp),%xmm4
+
+# qhasm:   in0 = z0
+# asm 1: movd   <z0=int6464#13,>in0=int64#3
+# asm 2: movd   <z0=%xmm12,>in0=%rdx
+movd   %xmm12,%rdx
+
+# qhasm:   in1 = z1
+# asm 1: movd   <z1=int6464#8,>in1=int64#4
+# asm 2: movd   <z1=%xmm7,>in1=%rcx
+movd   %xmm7,%rcx
+
+# qhasm:   in2 = z2
+# asm 1: movd   <z2=int6464#11,>in2=int64#5
+# asm 2: movd   <z2=%xmm10,>in2=%r8
+movd   %xmm10,%r8
+
+# qhasm:   in3 = z3
+# asm 1: movd   <z3=int6464#5,>in3=int64#6
+# asm 2: movd   <z3=%xmm4,>in3=%r9
+movd   %xmm4,%r9
+
+# qhasm:   z0 <<<= 96
+# asm 1: pshufd $0x39,<z0=int6464#13,<z0=int6464#13
+# asm 2: pshufd $0x39,<z0=%xmm12,<z0=%xmm12
+pshufd $0x39,%xmm12,%xmm12
+
+# qhasm:   z1 <<<= 96
+# asm 1: pshufd $0x39,<z1=int6464#8,<z1=int6464#8
+# asm 2: pshufd $0x39,<z1=%xmm7,<z1=%xmm7
+pshufd $0x39,%xmm7,%xmm7
+
+# qhasm:   z2 <<<= 96
+# asm 1: pshufd $0x39,<z2=int6464#11,<z2=int6464#11
+# asm 2: pshufd $0x39,<z2=%xmm10,<z2=%xmm10
+pshufd $0x39,%xmm10,%xmm10
+
+# qhasm:   z3 <<<= 96
+# asm 1: pshufd $0x39,<z3=int6464#5,<z3=int6464#5
+# asm 2: pshufd $0x39,<z3=%xmm4,<z3=%xmm4
+pshufd $0x39,%xmm4,%xmm4
+
+# qhasm:   (uint32) in0 ^= *(uint32 *) (m + 0)
+# asm 1: xorl 0(<m=int64#2),<in0=int64#3d
+# asm 2: xorl 0(<m=%rsi),<in0=%edx
+xorl 0(%rsi),%edx
+
+# qhasm:   (uint32) in1 ^= *(uint32 *) (m + 4)
+# asm 1: xorl 4(<m=int64#2),<in1=int64#4d
+# asm 2: xorl 4(<m=%rsi),<in1=%ecx
+xorl 4(%rsi),%ecx
+
+# qhasm:   (uint32) in2 ^= *(uint32 *) (m + 8)
+# asm 1: xorl 8(<m=int64#2),<in2=int64#5d
+# asm 2: xorl 8(<m=%rsi),<in2=%r8d
+xorl 8(%rsi),%r8d
+
+# qhasm:   (uint32) in3 ^= *(uint32 *) (m + 12)
+# asm 1: xorl 12(<m=int64#2),<in3=int64#6d
+# asm 2: xorl 12(<m=%rsi),<in3=%r9d
+xorl 12(%rsi),%r9d
+
+# qhasm:   *(uint32 *) (out + 0) = in0
+# asm 1: movl   <in0=int64#3d,0(<out=int64#1)
+# asm 2: movl   <in0=%edx,0(<out=%rdi)
+movl   %edx,0(%rdi)
+
+# qhasm:   *(uint32 *) (out + 4) = in1
+# asm 1: movl   <in1=int64#4d,4(<out=int64#1)
+# asm 2: movl   <in1=%ecx,4(<out=%rdi)
+movl   %ecx,4(%rdi)
+
+# qhasm:   *(uint32 *) (out + 8) = in2
+# asm 1: movl   <in2=int64#5d,8(<out=int64#1)
+# asm 2: movl   <in2=%r8d,8(<out=%rdi)
+movl   %r8d,8(%rdi)
+
+# qhasm:   *(uint32 *) (out + 12) = in3
+# asm 1: movl   <in3=int64#6d,12(<out=int64#1)
+# asm 2: movl   <in3=%r9d,12(<out=%rdi)
+movl   %r9d,12(%rdi)
+
+# qhasm:   in0 = z0
+# asm 1: movd   <z0=int6464#13,>in0=int64#3
+# asm 2: movd   <z0=%xmm12,>in0=%rdx
+movd   %xmm12,%rdx
+
+# qhasm:   in1 = z1
+# asm 1: movd   <z1=int6464#8,>in1=int64#4
+# asm 2: movd   <z1=%xmm7,>in1=%rcx
+movd   %xmm7,%rcx
+
+# qhasm:   in2 = z2
+# asm 1: movd   <z2=int6464#11,>in2=int64#5
+# asm 2: movd   <z2=%xmm10,>in2=%r8
+movd   %xmm10,%r8
+
+# qhasm:   in3 = z3
+# asm 1: movd   <z3=int6464#5,>in3=int64#6
+# asm 2: movd   <z3=%xmm4,>in3=%r9
+movd   %xmm4,%r9
+
+# qhasm:   z0 <<<= 96
+# asm 1: pshufd $0x39,<z0=int6464#13,<z0=int6464#13
+# asm 2: pshufd $0x39,<z0=%xmm12,<z0=%xmm12
+pshufd $0x39,%xmm12,%xmm12
+
+# qhasm:   z1 <<<= 96
+# asm 1: pshufd $0x39,<z1=int6464#8,<z1=int6464#8
+# asm 2: pshufd $0x39,<z1=%xmm7,<z1=%xmm7
+pshufd $0x39,%xmm7,%xmm7
+
+# qhasm:   z2 <<<= 96
+# asm 1: pshufd $0x39,<z2=int6464#11,<z2=int6464#11
+# asm 2: pshufd $0x39,<z2=%xmm10,<z2=%xmm10
+pshufd $0x39,%xmm10,%xmm10
+
+# qhasm:   z3 <<<= 96
+# asm 1: pshufd $0x39,<z3=int6464#5,<z3=int6464#5
+# asm 2: pshufd $0x39,<z3=%xmm4,<z3=%xmm4
+pshufd $0x39,%xmm4,%xmm4
+
+# qhasm:   (uint32) in0 ^= *(uint32 *) (m + 64)
+# asm 1: xorl 64(<m=int64#2),<in0=int64#3d
+# asm 2: xorl 64(<m=%rsi),<in0=%edx
+xorl 64(%rsi),%edx
+
+# qhasm:   (uint32) in1 ^= *(uint32 *) (m + 68)
+# asm 1: xorl 68(<m=int64#2),<in1=int64#4d
+# asm 2: xorl 68(<m=%rsi),<in1=%ecx
+xorl 68(%rsi),%ecx
+
+# qhasm:   (uint32) in2 ^= *(uint32 *) (m + 72)
+# asm 1: xorl 72(<m=int64#2),<in2=int64#5d
+# asm 2: xorl 72(<m=%rsi),<in2=%r8d
+xorl 72(%rsi),%r8d
+
+# qhasm:   (uint32) in3 ^= *(uint32 *) (m + 76)
+# asm 1: xorl 76(<m=int64#2),<in3=int64#6d
+# asm 2: xorl 76(<m=%rsi),<in3=%r9d
+xorl 76(%rsi),%r9d
+
+# qhasm:   *(uint32 *) (out + 64) = in0
+# asm 1: movl   <in0=int64#3d,64(<out=int64#1)
+# asm 2: movl   <in0=%edx,64(<out=%rdi)
+movl   %edx,64(%rdi)
+
+# qhasm:   *(uint32 *) (out + 68) = in1
+# asm 1: movl   <in1=int64#4d,68(<out=int64#1)
+# asm 2: movl   <in1=%ecx,68(<out=%rdi)
+movl   %ecx,68(%rdi)
+
+# qhasm:   *(uint32 *) (out + 72) = in2
+# asm 1: movl   <in2=int64#5d,72(<out=int64#1)
+# asm 2: movl   <in2=%r8d,72(<out=%rdi)
+movl   %r8d,72(%rdi)
+
+# qhasm:   *(uint32 *) (out + 76) = in3
+# asm 1: movl   <in3=int64#6d,76(<out=int64#1)
+# asm 2: movl   <in3=%r9d,76(<out=%rdi)
+movl   %r9d,76(%rdi)
+
+# qhasm:   in0 = z0
+# asm 1: movd   <z0=int6464#13,>in0=int64#3
+# asm 2: movd   <z0=%xmm12,>in0=%rdx
+movd   %xmm12,%rdx
+
+# qhasm:   in1 = z1
+# asm 1: movd   <z1=int6464#8,>in1=int64#4
+# asm 2: movd   <z1=%xmm7,>in1=%rcx
+movd   %xmm7,%rcx
+
+# qhasm:   in2 = z2
+# asm 1: movd   <z2=int6464#11,>in2=int64#5
+# asm 2: movd   <z2=%xmm10,>in2=%r8
+movd   %xmm10,%r8
+
+# qhasm:   in3 = z3
+# asm 1: movd   <z3=int6464#5,>in3=int64#6
+# asm 2: movd   <z3=%xmm4,>in3=%r9
+movd   %xmm4,%r9
+
+# qhasm:   z0 <<<= 96
+# asm 1: pshufd $0x39,<z0=int6464#13,<z0=int6464#13
+# asm 2: pshufd $0x39,<z0=%xmm12,<z0=%xmm12
+pshufd $0x39,%xmm12,%xmm12
+
+# qhasm:   z1 <<<= 96
+# asm 1: pshufd $0x39,<z1=int6464#8,<z1=int6464#8
+# asm 2: pshufd $0x39,<z1=%xmm7,<z1=%xmm7
+pshufd $0x39,%xmm7,%xmm7
+
+# qhasm:   z2 <<<= 96
+# asm 1: pshufd $0x39,<z2=int6464#11,<z2=int6464#11
+# asm 2: pshufd $0x39,<z2=%xmm10,<z2=%xmm10
+pshufd $0x39,%xmm10,%xmm10
+
+# qhasm:   z3 <<<= 96
+# asm 1: pshufd $0x39,<z3=int6464#5,<z3=int6464#5
+# asm 2: pshufd $0x39,<z3=%xmm4,<z3=%xmm4
+pshufd $0x39,%xmm4,%xmm4
+
+# qhasm:   (uint32) in0 ^= *(uint32 *) (m + 128)
+# asm 1: xorl 128(<m=int64#2),<in0=int64#3d
+# asm 2: xorl 128(<m=%rsi),<in0=%edx
+xorl 128(%rsi),%edx
+
+# qhasm:   (uint32) in1 ^= *(uint32 *) (m + 132)
+# asm 1: xorl 132(<m=int64#2),<in1=int64#4d
+# asm 2: xorl 132(<m=%rsi),<in1=%ecx
+xorl 132(%rsi),%ecx
+
+# qhasm:   (uint32) in2 ^= *(uint32 *) (m + 136)
+# asm 1: xorl 136(<m=int64#2),<in2=int64#5d
+# asm 2: xorl 136(<m=%rsi),<in2=%r8d
+xorl 136(%rsi),%r8d
+
+# qhasm:   (uint32) in3 ^= *(uint32 *) (m + 140)
+# asm 1: xorl 140(<m=int64#2),<in3=int64#6d
+# asm 2: xorl 140(<m=%rsi),<in3=%r9d
+xorl 140(%rsi),%r9d
+
+# qhasm:   *(uint32 *) (out + 128) = in0
+# asm 1: movl   <in0=int64#3d,128(<out=int64#1)
+# asm 2: movl   <in0=%edx,128(<out=%rdi)
+movl   %edx,128(%rdi)
+
+# qhasm:   *(uint32 *) (out + 132) = in1
+# asm 1: movl   <in1=int64#4d,132(<out=int64#1)
+# asm 2: movl   <in1=%ecx,132(<out=%rdi)
+movl   %ecx,132(%rdi)
+
+# qhasm:   *(uint32 *) (out + 136) = in2
+# asm 1: movl   <in2=int64#5d,136(<out=int64#1)
+# asm 2: movl   <in2=%r8d,136(<out=%rdi)
+movl   %r8d,136(%rdi)
+
+# qhasm:   *(uint32 *) (out + 140) = in3
+# asm 1: movl   <in3=int64#6d,140(<out=int64#1)
+# asm 2: movl   <in3=%r9d,140(<out=%rdi)
+movl   %r9d,140(%rdi)
+
+# qhasm:   in0 = z0
+# asm 1: movd   <z0=int6464#13,>in0=int64#3
+# asm 2: movd   <z0=%xmm12,>in0=%rdx
+movd   %xmm12,%rdx
+
+# qhasm:   in1 = z1
+# asm 1: movd   <z1=int6464#8,>in1=int64#4
+# asm 2: movd   <z1=%xmm7,>in1=%rcx
+movd   %xmm7,%rcx
+
+# qhasm:   in2 = z2
+# asm 1: movd   <z2=int6464#11,>in2=int64#5
+# asm 2: movd   <z2=%xmm10,>in2=%r8
+movd   %xmm10,%r8
+
+# qhasm:   in3 = z3
+# asm 1: movd   <z3=int6464#5,>in3=int64#6
+# asm 2: movd   <z3=%xmm4,>in3=%r9
+movd   %xmm4,%r9
+
+# qhasm:   (uint32) in0 ^= *(uint32 *) (m + 192)
+# asm 1: xorl 192(<m=int64#2),<in0=int64#3d
+# asm 2: xorl 192(<m=%rsi),<in0=%edx
+xorl 192(%rsi),%edx
+
+# qhasm:   (uint32) in1 ^= *(uint32 *) (m + 196)
+# asm 1: xorl 196(<m=int64#2),<in1=int64#4d
+# asm 2: xorl 196(<m=%rsi),<in1=%ecx
+xorl 196(%rsi),%ecx
+
+# qhasm:   (uint32) in2 ^= *(uint32 *) (m + 200)
+# asm 1: xorl 200(<m=int64#2),<in2=int64#5d
+# asm 2: xorl 200(<m=%rsi),<in2=%r8d
+xorl 200(%rsi),%r8d
+
+# qhasm:   (uint32) in3 ^= *(uint32 *) (m + 204)
+# asm 1: xorl 204(<m=int64#2),<in3=int64#6d
+# asm 2: xorl 204(<m=%rsi),<in3=%r9d
+xorl 204(%rsi),%r9d
+
+# qhasm:   *(uint32 *) (out + 192) = in0
+# asm 1: movl   <in0=int64#3d,192(<out=int64#1)
+# asm 2: movl   <in0=%edx,192(<out=%rdi)
+movl   %edx,192(%rdi)
+
+# qhasm:   *(uint32 *) (out + 196) = in1
+# asm 1: movl   <in1=int64#4d,196(<out=int64#1)
+# asm 2: movl   <in1=%ecx,196(<out=%rdi)
+movl   %ecx,196(%rdi)
+
+# qhasm:   *(uint32 *) (out + 200) = in2
+# asm 1: movl   <in2=int64#5d,200(<out=int64#1)
+# asm 2: movl   <in2=%r8d,200(<out=%rdi)
+movl   %r8d,200(%rdi)
+
+# qhasm:   *(uint32 *) (out + 204) = in3
+# asm 1: movl   <in3=int64#6d,204(<out=int64#1)
+# asm 2: movl   <in3=%r9d,204(<out=%rdi)
+movl   %r9d,204(%rdi)
+
+# qhasm:   uint32323232 z4 += orig4
+# asm 1: paddd <orig4=stack128#16,<z4=int6464#15
+# asm 2: paddd <orig4=240(%rsp),<z4=%xmm14
+paddd 240(%rsp),%xmm14
+
+# qhasm:   uint32323232 z5 += orig5
+# asm 1: paddd <orig5=stack128#5,<z5=int6464#1
+# asm 2: paddd <orig5=64(%rsp),<z5=%xmm0
+paddd 64(%rsp),%xmm0
+
+# qhasm:   uint32323232 z6 += orig6
+# asm 1: paddd <orig6=stack128#9,<z6=int6464#6
+# asm 2: paddd <orig6=128(%rsp),<z6=%xmm5
+paddd 128(%rsp),%xmm5
+
+# qhasm:   uint32323232 z7 += orig7
+# asm 1: paddd <orig7=stack128#13,<z7=int6464#9
+# asm 2: paddd <orig7=192(%rsp),<z7=%xmm8
+paddd 192(%rsp),%xmm8
+
+# qhasm:   in4 = z4
+# asm 1: movd   <z4=int6464#15,>in4=int64#3
+# asm 2: movd   <z4=%xmm14,>in4=%rdx
+movd   %xmm14,%rdx
+
+# qhasm:   in5 = z5
+# asm 1: movd   <z5=int6464#1,>in5=int64#4
+# asm 2: movd   <z5=%xmm0,>in5=%rcx
+movd   %xmm0,%rcx
+
+# qhasm:   in6 = z6
+# asm 1: movd   <z6=int6464#6,>in6=int64#5
+# asm 2: movd   <z6=%xmm5,>in6=%r8
+movd   %xmm5,%r8
+
+# qhasm:   in7 = z7
+# asm 1: movd   <z7=int6464#9,>in7=int64#6
+# asm 2: movd   <z7=%xmm8,>in7=%r9
+movd   %xmm8,%r9
+
+# qhasm:   z4 <<<= 96
+# asm 1: pshufd $0x39,<z4=int6464#15,<z4=int6464#15
+# asm 2: pshufd $0x39,<z4=%xmm14,<z4=%xmm14
+pshufd $0x39,%xmm14,%xmm14
+
+# qhasm:   z5 <<<= 96
+# asm 1: pshufd $0x39,<z5=int6464#1,<z5=int6464#1
+# asm 2: pshufd $0x39,<z5=%xmm0,<z5=%xmm0
+pshufd $0x39,%xmm0,%xmm0
+
+# qhasm:   z6 <<<= 96
+# asm 1: pshufd $0x39,<z6=int6464#6,<z6=int6464#6
+# asm 2: pshufd $0x39,<z6=%xmm5,<z6=%xmm5
+pshufd $0x39,%xmm5,%xmm5
+
+# qhasm:   z7 <<<= 96
+# asm 1: pshufd $0x39,<z7=int6464#9,<z7=int6464#9
+# asm 2: pshufd $0x39,<z7=%xmm8,<z7=%xmm8
+pshufd $0x39,%xmm8,%xmm8
+
+# qhasm:   (uint32) in4 ^= *(uint32 *) (m + 16)
+# asm 1: xorl 16(<m=int64#2),<in4=int64#3d
+# asm 2: xorl 16(<m=%rsi),<in4=%edx
+xorl 16(%rsi),%edx
+
+# qhasm:   (uint32) in5 ^= *(uint32 *) (m + 20)
+# asm 1: xorl 20(<m=int64#2),<in5=int64#4d
+# asm 2: xorl 20(<m=%rsi),<in5=%ecx
+xorl 20(%rsi),%ecx
+
+# qhasm:   (uint32) in6 ^= *(uint32 *) (m + 24)
+# asm 1: xorl 24(<m=int64#2),<in6=int64#5d
+# asm 2: xorl 24(<m=%rsi),<in6=%r8d
+xorl 24(%rsi),%r8d
+
+# qhasm:   (uint32) in7 ^= *(uint32 *) (m + 28)
+# asm 1: xorl 28(<m=int64#2),<in7=int64#6d
+# asm 2: xorl 28(<m=%rsi),<in7=%r9d
+xorl 28(%rsi),%r9d
+
+# qhasm:   *(uint32 *) (out + 16) = in4
+# asm 1: movl   <in4=int64#3d,16(<out=int64#1)
+# asm 2: movl   <in4=%edx,16(<out=%rdi)
+movl   %edx,16(%rdi)
+
+# qhasm:   *(uint32 *) (out + 20) = in5
+# asm 1: movl   <in5=int64#4d,20(<out=int64#1)
+# asm 2: movl   <in5=%ecx,20(<out=%rdi)
+movl   %ecx,20(%rdi)
+
+# qhasm:   *(uint32 *) (out + 24) = in6
+# asm 1: movl   <in6=int64#5d,24(<out=int64#1)
+# asm 2: movl   <in6=%r8d,24(<out=%rdi)
+movl   %r8d,24(%rdi)
+
+# qhasm:   *(uint32 *) (out + 28) = in7
+# asm 1: movl   <in7=int64#6d,28(<out=int64#1)
+# asm 2: movl   <in7=%r9d,28(<out=%rdi)
+movl   %r9d,28(%rdi)
+
+# qhasm:   in4 = z4
+# asm 1: movd   <z4=int6464#15,>in4=int64#3
+# asm 2: movd   <z4=%xmm14,>in4=%rdx
+movd   %xmm14,%rdx
+
+# qhasm:   in5 = z5
+# asm 1: movd   <z5=int6464#1,>in5=int64#4
+# asm 2: movd   <z5=%xmm0,>in5=%rcx
+movd   %xmm0,%rcx
+
+# qhasm:   in6 = z6
+# asm 1: movd   <z6=int6464#6,>in6=int64#5
+# asm 2: movd   <z6=%xmm5,>in6=%r8
+movd   %xmm5,%r8
+
+# qhasm:   in7 = z7
+# asm 1: movd   <z7=int6464#9,>in7=int64#6
+# asm 2: movd   <z7=%xmm8,>in7=%r9
+movd   %xmm8,%r9
+
+# qhasm:   z4 <<<= 96
+# asm 1: pshufd $0x39,<z4=int6464#15,<z4=int6464#15
+# asm 2: pshufd $0x39,<z4=%xmm14,<z4=%xmm14
+pshufd $0x39,%xmm14,%xmm14
+
+# qhasm:   z5 <<<= 96
+# asm 1: pshufd $0x39,<z5=int6464#1,<z5=int6464#1
+# asm 2: pshufd $0x39,<z5=%xmm0,<z5=%xmm0
+pshufd $0x39,%xmm0,%xmm0
+
+# qhasm:   z6 <<<= 96
+# asm 1: pshufd $0x39,<z6=int6464#6,<z6=int6464#6
+# asm 2: pshufd $0x39,<z6=%xmm5,<z6=%xmm5
+pshufd $0x39,%xmm5,%xmm5
+
+# qhasm:   z7 <<<= 96
+# asm 1: pshufd $0x39,<z7=int6464#9,<z7=int6464#9
+# asm 2: pshufd $0x39,<z7=%xmm8,<z7=%xmm8
+pshufd $0x39,%xmm8,%xmm8
+
+# qhasm:   (uint32) in4 ^= *(uint32 *) (m + 80)
+# asm 1: xorl 80(<m=int64#2),<in4=int64#3d
+# asm 2: xorl 80(<m=%rsi),<in4=%edx
+xorl 80(%rsi),%edx
+
+# qhasm:   (uint32) in5 ^= *(uint32 *) (m + 84)
+# asm 1: xorl 84(<m=int64#2),<in5=int64#4d
+# asm 2: xorl 84(<m=%rsi),<in5=%ecx
+xorl 84(%rsi),%ecx
+
+# qhasm:   (uint32) in6 ^= *(uint32 *) (m + 88)
+# asm 1: xorl 88(<m=int64#2),<in6=int64#5d
+# asm 2: xorl 88(<m=%rsi),<in6=%r8d
+xorl 88(%rsi),%r8d
+
+# qhasm:   (uint32) in7 ^= *(uint32 *) (m + 92)
+# asm 1: xorl 92(<m=int64#2),<in7=int64#6d
+# asm 2: xorl 92(<m=%rsi),<in7=%r9d
+xorl 92(%rsi),%r9d
+
+# qhasm:   *(uint32 *) (out + 80) = in4
+# asm 1: movl   <in4=int64#3d,80(<out=int64#1)
+# asm 2: movl   <in4=%edx,80(<out=%rdi)
+movl   %edx,80(%rdi)
+
+# qhasm:   *(uint32 *) (out + 84) = in5
+# asm 1: movl   <in5=int64#4d,84(<out=int64#1)
+# asm 2: movl   <in5=%ecx,84(<out=%rdi)
+movl   %ecx,84(%rdi)
+
+# qhasm:   *(uint32 *) (out + 88) = in6
+# asm 1: movl   <in6=int64#5d,88(<out=int64#1)
+# asm 2: movl   <in6=%r8d,88(<out=%rdi)
+movl   %r8d,88(%rdi)
+
+# qhasm:   *(uint32 *) (out + 92) = in7
+# asm 1: movl   <in7=int64#6d,92(<out=int64#1)
+# asm 2: movl   <in7=%r9d,92(<out=%rdi)
+movl   %r9d,92(%rdi)
+
+# qhasm:   in4 = z4
+# asm 1: movd   <z4=int6464#15,>in4=int64#3
+# asm 2: movd   <z4=%xmm14,>in4=%rdx
+movd   %xmm14,%rdx
+
+# qhasm:   in5 = z5
+# asm 1: movd   <z5=int6464#1,>in5=int64#4
+# asm 2: movd   <z5=%xmm0,>in5=%rcx
+movd   %xmm0,%rcx
+
+# qhasm:   in6 = z6
+# asm 1: movd   <z6=int6464#6,>in6=int64#5
+# asm 2: movd   <z6=%xmm5,>in6=%r8
+movd   %xmm5,%r8
+
+# qhasm:   in7 = z7
+# asm 1: movd   <z7=int6464#9,>in7=int64#6
+# asm 2: movd   <z7=%xmm8,>in7=%r9
+movd   %xmm8,%r9
+
+# qhasm:   z4 <<<= 96
+# asm 1: pshufd $0x39,<z4=int6464#15,<z4=int6464#15
+# asm 2: pshufd $0x39,<z4=%xmm14,<z4=%xmm14
+pshufd $0x39,%xmm14,%xmm14
+
+# qhasm:   z5 <<<= 96
+# asm 1: pshufd $0x39,<z5=int6464#1,<z5=int6464#1
+# asm 2: pshufd $0x39,<z5=%xmm0,<z5=%xmm0
+pshufd $0x39,%xmm0,%xmm0
+
+# qhasm:   z6 <<<= 96
+# asm 1: pshufd $0x39,<z6=int6464#6,<z6=int6464#6
+# asm 2: pshufd $0x39,<z6=%xmm5,<z6=%xmm5
+pshufd $0x39,%xmm5,%xmm5
+
+# qhasm:   z7 <<<= 96
+# asm 1: pshufd $0x39,<z7=int6464#9,<z7=int6464#9
+# asm 2: pshufd $0x39,<z7=%xmm8,<z7=%xmm8
+pshufd $0x39,%xmm8,%xmm8
+
+# qhasm:   (uint32) in4 ^= *(uint32 *) (m + 144)
+# asm 1: xorl 144(<m=int64#2),<in4=int64#3d
+# asm 2: xorl 144(<m=%rsi),<in4=%edx
+xorl 144(%rsi),%edx
+
+# qhasm:   (uint32) in5 ^= *(uint32 *) (m + 148)
+# asm 1: xorl 148(<m=int64#2),<in5=int64#4d
+# asm 2: xorl 148(<m=%rsi),<in5=%ecx
+xorl 148(%rsi),%ecx
+
+# qhasm:   (uint32) in6 ^= *(uint32 *) (m + 152)
+# asm 1: xorl 152(<m=int64#2),<in6=int64#5d
+# asm 2: xorl 152(<m=%rsi),<in6=%r8d
+xorl 152(%rsi),%r8d
+
+# qhasm:   (uint32) in7 ^= *(uint32 *) (m + 156)
+# asm 1: xorl 156(<m=int64#2),<in7=int64#6d
+# asm 2: xorl 156(<m=%rsi),<in7=%r9d
+xorl 156(%rsi),%r9d
+
+# qhasm:   *(uint32 *) (out + 144) = in4
+# asm 1: movl   <in4=int64#3d,144(<out=int64#1)
+# asm 2: movl   <in4=%edx,144(<out=%rdi)
+movl   %edx,144(%rdi)
+
+# qhasm:   *(uint32 *) (out + 148) = in5
+# asm 1: movl   <in5=int64#4d,148(<out=int64#1)
+# asm 2: movl   <in5=%ecx,148(<out=%rdi)
+movl   %ecx,148(%rdi)
+
+# qhasm:   *(uint32 *) (out + 152) = in6
+# asm 1: movl   <in6=int64#5d,152(<out=int64#1)
+# asm 2: movl   <in6=%r8d,152(<out=%rdi)
+movl   %r8d,152(%rdi)
+
+# qhasm:   *(uint32 *) (out + 156) = in7
+# asm 1: movl   <in7=int64#6d,156(<out=int64#1)
+# asm 2: movl   <in7=%r9d,156(<out=%rdi)
+movl   %r9d,156(%rdi)
+
+# qhasm:   in4 = z4
+# asm 1: movd   <z4=int6464#15,>in4=int64#3
+# asm 2: movd   <z4=%xmm14,>in4=%rdx
+movd   %xmm14,%rdx
+
+# qhasm:   in5 = z5
+# asm 1: movd   <z5=int6464#1,>in5=int64#4
+# asm 2: movd   <z5=%xmm0,>in5=%rcx
+movd   %xmm0,%rcx
+
+# qhasm:   in6 = z6
+# asm 1: movd   <z6=int6464#6,>in6=int64#5
+# asm 2: movd   <z6=%xmm5,>in6=%r8
+movd   %xmm5,%r8
+
+# qhasm:   in7 = z7
+# asm 1: movd   <z7=int6464#9,>in7=int64#6
+# asm 2: movd   <z7=%xmm8,>in7=%r9
+movd   %xmm8,%r9
+
+# qhasm:   (uint32) in4 ^= *(uint32 *) (m + 208)
+# asm 1: xorl 208(<m=int64#2),<in4=int64#3d
+# asm 2: xorl 208(<m=%rsi),<in4=%edx
+xorl 208(%rsi),%edx
+
+# qhasm:   (uint32) in5 ^= *(uint32 *) (m + 212)
+# asm 1: xorl 212(<m=int64#2),<in5=int64#4d
+# asm 2: xorl 212(<m=%rsi),<in5=%ecx
+xorl 212(%rsi),%ecx
+
+# qhasm:   (uint32) in6 ^= *(uint32 *) (m + 216)
+# asm 1: xorl 216(<m=int64#2),<in6=int64#5d
+# asm 2: xorl 216(<m=%rsi),<in6=%r8d
+xorl 216(%rsi),%r8d
+
+# qhasm:   (uint32) in7 ^= *(uint32 *) (m + 220)
+# asm 1: xorl 220(<m=int64#2),<in7=int64#6d
+# asm 2: xorl 220(<m=%rsi),<in7=%r9d
+xorl 220(%rsi),%r9d
+
+# qhasm:   *(uint32 *) (out + 208) = in4
+# asm 1: movl   <in4=int64#3d,208(<out=int64#1)
+# asm 2: movl   <in4=%edx,208(<out=%rdi)
+movl   %edx,208(%rdi)
+
+# qhasm:   *(uint32 *) (out + 212) = in5
+# asm 1: movl   <in5=int64#4d,212(<out=int64#1)
+# asm 2: movl   <in5=%ecx,212(<out=%rdi)
+movl   %ecx,212(%rdi)
+
+# qhasm:   *(uint32 *) (out + 216) = in6
+# asm 1: movl   <in6=int64#5d,216(<out=int64#1)
+# asm 2: movl   <in6=%r8d,216(<out=%rdi)
+movl   %r8d,216(%rdi)
+
+# qhasm:   *(uint32 *) (out + 220) = in7
+# asm 1: movl   <in7=int64#6d,220(<out=int64#1)
+# asm 2: movl   <in7=%r9d,220(<out=%rdi)
+movl   %r9d,220(%rdi)
+
+# qhasm:   uint32323232 z8 += orig8
+# asm 1: paddd <orig8=stack128#19,<z8=int6464#16
+# asm 2: paddd <orig8=288(%rsp),<z8=%xmm15
+paddd 288(%rsp),%xmm15
+
+# qhasm:   uint32323232 z9 += orig9
+# asm 1: paddd <orig9=stack128#20,<z9=int6464#12
+# asm 2: paddd <orig9=304(%rsp),<z9=%xmm11
+paddd 304(%rsp),%xmm11
+
+# qhasm:   uint32323232 z10 += orig10
+# asm 1: paddd <orig10=stack128#6,<z10=int6464#2
+# asm 2: paddd <orig10=80(%rsp),<z10=%xmm1
+paddd 80(%rsp),%xmm1
+
+# qhasm:   uint32323232 z11 += orig11
+# asm 1: paddd <orig11=stack128#10,<z11=int6464#7
+# asm 2: paddd <orig11=144(%rsp),<z11=%xmm6
+paddd 144(%rsp),%xmm6
+
+# qhasm:   in8 = z8
+# asm 1: movd   <z8=int6464#16,>in8=int64#3
+# asm 2: movd   <z8=%xmm15,>in8=%rdx
+movd   %xmm15,%rdx
+
+# qhasm:   in9 = z9
+# asm 1: movd   <z9=int6464#12,>in9=int64#4
+# asm 2: movd   <z9=%xmm11,>in9=%rcx
+movd   %xmm11,%rcx
+
+# qhasm:   in10 = z10
+# asm 1: movd   <z10=int6464#2,>in10=int64#5
+# asm 2: movd   <z10=%xmm1,>in10=%r8
+movd   %xmm1,%r8
+
+# qhasm:   in11 = z11
+# asm 1: movd   <z11=int6464#7,>in11=int64#6
+# asm 2: movd   <z11=%xmm6,>in11=%r9
+movd   %xmm6,%r9
+
+# qhasm:   z8 <<<= 96
+# asm 1: pshufd $0x39,<z8=int6464#16,<z8=int6464#16
+# asm 2: pshufd $0x39,<z8=%xmm15,<z8=%xmm15
+pshufd $0x39,%xmm15,%xmm15
+
+# qhasm:   z9 <<<= 96
+# asm 1: pshufd $0x39,<z9=int6464#12,<z9=int6464#12
+# asm 2: pshufd $0x39,<z9=%xmm11,<z9=%xmm11
+pshufd $0x39,%xmm11,%xmm11
+
+# qhasm:   z10 <<<= 96
+# asm 1: pshufd $0x39,<z10=int6464#2,<z10=int6464#2
+# asm 2: pshufd $0x39,<z10=%xmm1,<z10=%xmm1
+pshufd $0x39,%xmm1,%xmm1
+
+# qhasm:   z11 <<<= 96
+# asm 1: pshufd $0x39,<z11=int6464#7,<z11=int6464#7
+# asm 2: pshufd $0x39,<z11=%xmm6,<z11=%xmm6
+pshufd $0x39,%xmm6,%xmm6
+
+# qhasm:   (uint32) in8 ^= *(uint32 *) (m + 32)
+# asm 1: xorl 32(<m=int64#2),<in8=int64#3d
+# asm 2: xorl 32(<m=%rsi),<in8=%edx
+xorl 32(%rsi),%edx
+
+# qhasm:   (uint32) in9 ^= *(uint32 *) (m + 36)
+# asm 1: xorl 36(<m=int64#2),<in9=int64#4d
+# asm 2: xorl 36(<m=%rsi),<in9=%ecx
+xorl 36(%rsi),%ecx
+
+# qhasm:   (uint32) in10 ^= *(uint32 *) (m + 40)
+# asm 1: xorl 40(<m=int64#2),<in10=int64#5d
+# asm 2: xorl 40(<m=%rsi),<in10=%r8d
+xorl 40(%rsi),%r8d
+
+# qhasm:   (uint32) in11 ^= *(uint32 *) (m + 44)
+# asm 1: xorl 44(<m=int64#2),<in11=int64#6d
+# asm 2: xorl 44(<m=%rsi),<in11=%r9d
+xorl 44(%rsi),%r9d
+
+# qhasm:   *(uint32 *) (out + 32) = in8
+# asm 1: movl   <in8=int64#3d,32(<out=int64#1)
+# asm 2: movl   <in8=%edx,32(<out=%rdi)
+movl   %edx,32(%rdi)
+
+# qhasm:   *(uint32 *) (out + 36) = in9
+# asm 1: movl   <in9=int64#4d,36(<out=int64#1)
+# asm 2: movl   <in9=%ecx,36(<out=%rdi)
+movl   %ecx,36(%rdi)
+
+# qhasm:   *(uint32 *) (out + 40) = in10
+# asm 1: movl   <in10=int64#5d,40(<out=int64#1)
+# asm 2: movl   <in10=%r8d,40(<out=%rdi)
+movl   %r8d,40(%rdi)
+
+# qhasm:   *(uint32 *) (out + 44) = in11
+# asm 1: movl   <in11=int64#6d,44(<out=int64#1)
+# asm 2: movl   <in11=%r9d,44(<out=%rdi)
+movl   %r9d,44(%rdi)
+
+# qhasm:   in8 = z8
+# asm 1: movd   <z8=int6464#16,>in8=int64#3
+# asm 2: movd   <z8=%xmm15,>in8=%rdx
+movd   %xmm15,%rdx
+
+# qhasm:   in9 = z9
+# asm 1: movd   <z9=int6464#12,>in9=int64#4
+# asm 2: movd   <z9=%xmm11,>in9=%rcx
+movd   %xmm11,%rcx
+
+# qhasm:   in10 = z10
+# asm 1: movd   <z10=int6464#2,>in10=int64#5
+# asm 2: movd   <z10=%xmm1,>in10=%r8
+movd   %xmm1,%r8
+
+# qhasm:   in11 = z11
+# asm 1: movd   <z11=int6464#7,>in11=int64#6
+# asm 2: movd   <z11=%xmm6,>in11=%r9
+movd   %xmm6,%r9
+
+# qhasm:   z8 <<<= 96
+# asm 1: pshufd $0x39,<z8=int6464#16,<z8=int6464#16
+# asm 2: pshufd $0x39,<z8=%xmm15,<z8=%xmm15
+pshufd $0x39,%xmm15,%xmm15
+
+# qhasm:   z9 <<<= 96
+# asm 1: pshufd $0x39,<z9=int6464#12,<z9=int6464#12
+# asm 2: pshufd $0x39,<z9=%xmm11,<z9=%xmm11
+pshufd $0x39,%xmm11,%xmm11
+
+# qhasm:   z10 <<<= 96
+# asm 1: pshufd $0x39,<z10=int6464#2,<z10=int6464#2
+# asm 2: pshufd $0x39,<z10=%xmm1,<z10=%xmm1
+pshufd $0x39,%xmm1,%xmm1
+
+# qhasm:   z11 <<<= 96
+# asm 1: pshufd $0x39,<z11=int6464#7,<z11=int6464#7
+# asm 2: pshufd $0x39,<z11=%xmm6,<z11=%xmm6
+pshufd $0x39,%xmm6,%xmm6
+
+# qhasm:   (uint32) in8 ^= *(uint32 *) (m + 96)
+# asm 1: xorl 96(<m=int64#2),<in8=int64#3d
+# asm 2: xorl 96(<m=%rsi),<in8=%edx
+xorl 96(%rsi),%edx
+
+# qhasm:   (uint32) in9 ^= *(uint32 *) (m + 100)
+# asm 1: xorl 100(<m=int64#2),<in9=int64#4d
+# asm 2: xorl 100(<m=%rsi),<in9=%ecx
+xorl 100(%rsi),%ecx
+
+# qhasm:   (uint32) in10 ^= *(uint32 *) (m + 104)
+# asm 1: xorl 104(<m=int64#2),<in10=int64#5d
+# asm 2: xorl 104(<m=%rsi),<in10=%r8d
+xorl 104(%rsi),%r8d
+
+# qhasm:   (uint32) in11 ^= *(uint32 *) (m + 108)
+# asm 1: xorl 108(<m=int64#2),<in11=int64#6d
+# asm 2: xorl 108(<m=%rsi),<in11=%r9d
+xorl 108(%rsi),%r9d
+
+# qhasm:   *(uint32 *) (out + 96) = in8
+# asm 1: movl   <in8=int64#3d,96(<out=int64#1)
+# asm 2: movl   <in8=%edx,96(<out=%rdi)
+movl   %edx,96(%rdi)
+
+# qhasm:   *(uint32 *) (out + 100) = in9
+# asm 1: movl   <in9=int64#4d,100(<out=int64#1)
+# asm 2: movl   <in9=%ecx,100(<out=%rdi)
+movl   %ecx,100(%rdi)
+
+# qhasm:   *(uint32 *) (out + 104) = in10
+# asm 1: movl   <in10=int64#5d,104(<out=int64#1)
+# asm 2: movl   <in10=%r8d,104(<out=%rdi)
+movl   %r8d,104(%rdi)
+
+# qhasm:   *(uint32 *) (out + 108) = in11
+# asm 1: movl   <in11=int64#6d,108(<out=int64#1)
+# asm 2: movl   <in11=%r9d,108(<out=%rdi)
+movl   %r9d,108(%rdi)
+
+# qhasm:   in8 = z8
+# asm 1: movd   <z8=int6464#16,>in8=int64#3
+# asm 2: movd   <z8=%xmm15,>in8=%rdx
+movd   %xmm15,%rdx
+
+# qhasm:   in9 = z9
+# asm 1: movd   <z9=int6464#12,>in9=int64#4
+# asm 2: movd   <z9=%xmm11,>in9=%rcx
+movd   %xmm11,%rcx
+
+# qhasm:   in10 = z10
+# asm 1: movd   <z10=int6464#2,>in10=int64#5
+# asm 2: movd   <z10=%xmm1,>in10=%r8
+movd   %xmm1,%r8
+
+# qhasm:   in11 = z11
+# asm 1: movd   <z11=int6464#7,>in11=int64#6
+# asm 2: movd   <z11=%xmm6,>in11=%r9
+movd   %xmm6,%r9
+
+# qhasm:   z8 <<<= 96
+# asm 1: pshufd $0x39,<z8=int6464#16,<z8=int6464#16
+# asm 2: pshufd $0x39,<z8=%xmm15,<z8=%xmm15
+pshufd $0x39,%xmm15,%xmm15
+
+# qhasm:   z9 <<<= 96
+# asm 1: pshufd $0x39,<z9=int6464#12,<z9=int6464#12
+# asm 2: pshufd $0x39,<z9=%xmm11,<z9=%xmm11
+pshufd $0x39,%xmm11,%xmm11
+
+# qhasm:   z10 <<<= 96
+# asm 1: pshufd $0x39,<z10=int6464#2,<z10=int6464#2
+# asm 2: pshufd $0x39,<z10=%xmm1,<z10=%xmm1
+pshufd $0x39,%xmm1,%xmm1
+
+# qhasm:   z11 <<<= 96
+# asm 1: pshufd $0x39,<z11=int6464#7,<z11=int6464#7
+# asm 2: pshufd $0x39,<z11=%xmm6,<z11=%xmm6
+pshufd $0x39,%xmm6,%xmm6
+
+# qhasm:   (uint32) in8 ^= *(uint32 *) (m + 160)
+# asm 1: xorl 160(<m=int64#2),<in8=int64#3d
+# asm 2: xorl 160(<m=%rsi),<in8=%edx
+xorl 160(%rsi),%edx
+
+# qhasm:   (uint32) in9 ^= *(uint32 *) (m + 164)
+# asm 1: xorl 164(<m=int64#2),<in9=int64#4d
+# asm 2: xorl 164(<m=%rsi),<in9=%ecx
+xorl 164(%rsi),%ecx
+
+# qhasm:   (uint32) in10 ^= *(uint32 *) (m + 168)
+# asm 1: xorl 168(<m=int64#2),<in10=int64#5d
+# asm 2: xorl 168(<m=%rsi),<in10=%r8d
+xorl 168(%rsi),%r8d
+
+# qhasm:   (uint32) in11 ^= *(uint32 *) (m + 172)
+# asm 1: xorl 172(<m=int64#2),<in11=int64#6d
+# asm 2: xorl 172(<m=%rsi),<in11=%r9d
+xorl 172(%rsi),%r9d
+
+# qhasm:   *(uint32 *) (out + 160) = in8
+# asm 1: movl   <in8=int64#3d,160(<out=int64#1)
+# asm 2: movl   <in8=%edx,160(<out=%rdi)
+movl   %edx,160(%rdi)
+
+# qhasm:   *(uint32 *) (out + 164) = in9
+# asm 1: movl   <in9=int64#4d,164(<out=int64#1)
+# asm 2: movl   <in9=%ecx,164(<out=%rdi)
+movl   %ecx,164(%rdi)
+
+# qhasm:   *(uint32 *) (out + 168) = in10
+# asm 1: movl   <in10=int64#5d,168(<out=int64#1)
+# asm 2: movl   <in10=%r8d,168(<out=%rdi)
+movl   %r8d,168(%rdi)
+
+# qhasm:   *(uint32 *) (out + 172) = in11
+# asm 1: movl   <in11=int64#6d,172(<out=int64#1)
+# asm 2: movl   <in11=%r9d,172(<out=%rdi)
+movl   %r9d,172(%rdi)
+
+# qhasm:   in8 = z8
+# asm 1: movd   <z8=int6464#16,>in8=int64#3
+# asm 2: movd   <z8=%xmm15,>in8=%rdx
+movd   %xmm15,%rdx
+
+# qhasm:   in9 = z9
+# asm 1: movd   <z9=int6464#12,>in9=int64#4
+# asm 2: movd   <z9=%xmm11,>in9=%rcx
+movd   %xmm11,%rcx
+
+# qhasm:   in10 = z10
+# asm 1: movd   <z10=int6464#2,>in10=int64#5
+# asm 2: movd   <z10=%xmm1,>in10=%r8
+movd   %xmm1,%r8
+
+# qhasm:   in11 = z11
+# asm 1: movd   <z11=int6464#7,>in11=int64#6
+# asm 2: movd   <z11=%xmm6,>in11=%r9
+movd   %xmm6,%r9
+
+# qhasm:   (uint32) in8 ^= *(uint32 *) (m + 224)
+# asm 1: xorl 224(<m=int64#2),<in8=int64#3d
+# asm 2: xorl 224(<m=%rsi),<in8=%edx
+xorl 224(%rsi),%edx
+
+# qhasm:   (uint32) in9 ^= *(uint32 *) (m + 228)
+# asm 1: xorl 228(<m=int64#2),<in9=int64#4d
+# asm 2: xorl 228(<m=%rsi),<in9=%ecx
+xorl 228(%rsi),%ecx
+
+# qhasm:   (uint32) in10 ^= *(uint32 *) (m + 232)
+# asm 1: xorl 232(<m=int64#2),<in10=int64#5d
+# asm 2: xorl 232(<m=%rsi),<in10=%r8d
+xorl 232(%rsi),%r8d
+
+# qhasm:   (uint32) in11 ^= *(uint32 *) (m + 236)
+# asm 1: xorl 236(<m=int64#2),<in11=int64#6d
+# asm 2: xorl 236(<m=%rsi),<in11=%r9d
+xorl 236(%rsi),%r9d
+
+# qhasm:   *(uint32 *) (out + 224) = in8
+# asm 1: movl   <in8=int64#3d,224(<out=int64#1)
+# asm 2: movl   <in8=%edx,224(<out=%rdi)
+movl   %edx,224(%rdi)
+
+# qhasm:   *(uint32 *) (out + 228) = in9
+# asm 1: movl   <in9=int64#4d,228(<out=int64#1)
+# asm 2: movl   <in9=%ecx,228(<out=%rdi)
+movl   %ecx,228(%rdi)
+
+# qhasm:   *(uint32 *) (out + 232) = in10
+# asm 1: movl   <in10=int64#5d,232(<out=int64#1)
+# asm 2: movl   <in10=%r8d,232(<out=%rdi)
+movl   %r8d,232(%rdi)
+
+# qhasm:   *(uint32 *) (out + 236) = in11
+# asm 1: movl   <in11=int64#6d,236(<out=int64#1)
+# asm 2: movl   <in11=%r9d,236(<out=%rdi)
+movl   %r9d,236(%rdi)
+
+# qhasm:   uint32323232 z12 += orig12
+# asm 1: paddd <orig12=stack128#11,<z12=int6464#14
+# asm 2: paddd <orig12=160(%rsp),<z12=%xmm13
+paddd 160(%rsp),%xmm13
+
+# qhasm:   uint32323232 z13 += orig13
+# asm 1: paddd <orig13=stack128#14,<z13=int6464#10
+# asm 2: paddd <orig13=208(%rsp),<z13=%xmm9
+paddd 208(%rsp),%xmm9
+
+# qhasm:   uint32323232 z14 += orig14
+# asm 1: paddd <orig14=stack128#17,<z14=int6464#4
+# asm 2: paddd <orig14=256(%rsp),<z14=%xmm3
+paddd 256(%rsp),%xmm3
+
+# qhasm:   uint32323232 z15 += orig15
+# asm 1: paddd <orig15=stack128#7,<z15=int6464#3
+# asm 2: paddd <orig15=96(%rsp),<z15=%xmm2
+paddd 96(%rsp),%xmm2
+
+# qhasm:   in12 = z12
+# asm 1: movd   <z12=int6464#14,>in12=int64#3
+# asm 2: movd   <z12=%xmm13,>in12=%rdx
+movd   %xmm13,%rdx
+
+# qhasm:   in13 = z13
+# asm 1: movd   <z13=int6464#10,>in13=int64#4
+# asm 2: movd   <z13=%xmm9,>in13=%rcx
+movd   %xmm9,%rcx
+
+# qhasm:   in14 = z14
+# asm 1: movd   <z14=int6464#4,>in14=int64#5
+# asm 2: movd   <z14=%xmm3,>in14=%r8
+movd   %xmm3,%r8
+
+# qhasm:   in15 = z15
+# asm 1: movd   <z15=int6464#3,>in15=int64#6
+# asm 2: movd   <z15=%xmm2,>in15=%r9
+movd   %xmm2,%r9
+
+# qhasm:   z12 <<<= 96
+# asm 1: pshufd $0x39,<z12=int6464#14,<z12=int6464#14
+# asm 2: pshufd $0x39,<z12=%xmm13,<z12=%xmm13
+pshufd $0x39,%xmm13,%xmm13
+
+# qhasm:   z13 <<<= 96
+# asm 1: pshufd $0x39,<z13=int6464#10,<z13=int6464#10
+# asm 2: pshufd $0x39,<z13=%xmm9,<z13=%xmm9
+pshufd $0x39,%xmm9,%xmm9
+
+# qhasm:   z14 <<<= 96
+# asm 1: pshufd $0x39,<z14=int6464#4,<z14=int6464#4
+# asm 2: pshufd $0x39,<z14=%xmm3,<z14=%xmm3
+pshufd $0x39,%xmm3,%xmm3
+
+# qhasm:   z15 <<<= 96
+# asm 1: pshufd $0x39,<z15=int6464#3,<z15=int6464#3
+# asm 2: pshufd $0x39,<z15=%xmm2,<z15=%xmm2
+pshufd $0x39,%xmm2,%xmm2
+
+# qhasm:   (uint32) in12 ^= *(uint32 *) (m + 48)
+# asm 1: xorl 48(<m=int64#2),<in12=int64#3d
+# asm 2: xorl 48(<m=%rsi),<in12=%edx
+xorl 48(%rsi),%edx
+
+# qhasm:   (uint32) in13 ^= *(uint32 *) (m + 52)
+# asm 1: xorl 52(<m=int64#2),<in13=int64#4d
+# asm 2: xorl 52(<m=%rsi),<in13=%ecx
+xorl 52(%rsi),%ecx
+
+# qhasm:   (uint32) in14 ^= *(uint32 *) (m + 56)
+# asm 1: xorl 56(<m=int64#2),<in14=int64#5d
+# asm 2: xorl 56(<m=%rsi),<in14=%r8d
+xorl 56(%rsi),%r8d
+
+# qhasm:   (uint32) in15 ^= *(uint32 *) (m + 60)
+# asm 1: xorl 60(<m=int64#2),<in15=int64#6d
+# asm 2: xorl 60(<m=%rsi),<in15=%r9d
+xorl 60(%rsi),%r9d
+
+# qhasm:   *(uint32 *) (out + 48) = in12
+# asm 1: movl   <in12=int64#3d,48(<out=int64#1)
+# asm 2: movl   <in12=%edx,48(<out=%rdi)
+movl   %edx,48(%rdi)
+
+# qhasm:   *(uint32 *) (out + 52) = in13
+# asm 1: movl   <in13=int64#4d,52(<out=int64#1)
+# asm 2: movl   <in13=%ecx,52(<out=%rdi)
+movl   %ecx,52(%rdi)
+
+# qhasm:   *(uint32 *) (out + 56) = in14
+# asm 1: movl   <in14=int64#5d,56(<out=int64#1)
+# asm 2: movl   <in14=%r8d,56(<out=%rdi)
+movl   %r8d,56(%rdi)
+
+# qhasm:   *(uint32 *) (out + 60) = in15
+# asm 1: movl   <in15=int64#6d,60(<out=int64#1)
+# asm 2: movl   <in15=%r9d,60(<out=%rdi)
+movl   %r9d,60(%rdi)
+
+# qhasm:   in12 = z12
+# asm 1: movd   <z12=int6464#14,>in12=int64#3
+# asm 2: movd   <z12=%xmm13,>in12=%rdx
+movd   %xmm13,%rdx
+
+# qhasm:   in13 = z13
+# asm 1: movd   <z13=int6464#10,>in13=int64#4
+# asm 2: movd   <z13=%xmm9,>in13=%rcx
+movd   %xmm9,%rcx
+
+# qhasm:   in14 = z14
+# asm 1: movd   <z14=int6464#4,>in14=int64#5
+# asm 2: movd   <z14=%xmm3,>in14=%r8
+movd   %xmm3,%r8
+
+# qhasm:   in15 = z15
+# asm 1: movd   <z15=int6464#3,>in15=int64#6
+# asm 2: movd   <z15=%xmm2,>in15=%r9
+movd   %xmm2,%r9
+
+# qhasm:   z12 <<<= 96
+# asm 1: pshufd $0x39,<z12=int6464#14,<z12=int6464#14
+# asm 2: pshufd $0x39,<z12=%xmm13,<z12=%xmm13
+pshufd $0x39,%xmm13,%xmm13
+
+# qhasm:   z13 <<<= 96
+# asm 1: pshufd $0x39,<z13=int6464#10,<z13=int6464#10
+# asm 2: pshufd $0x39,<z13=%xmm9,<z13=%xmm9
+pshufd $0x39,%xmm9,%xmm9
+
+# qhasm:   z14 <<<= 96
+# asm 1: pshufd $0x39,<z14=int6464#4,<z14=int6464#4
+# asm 2: pshufd $0x39,<z14=%xmm3,<z14=%xmm3
+pshufd $0x39,%xmm3,%xmm3
+
+# qhasm:   z15 <<<= 96
+# asm 1: pshufd $0x39,<z15=int6464#3,<z15=int6464#3
+# asm 2: pshufd $0x39,<z15=%xmm2,<z15=%xmm2
+pshufd $0x39,%xmm2,%xmm2
+
+# qhasm:   (uint32) in12 ^= *(uint32 *) (m + 112)
+# asm 1: xorl 112(<m=int64#2),<in12=int64#3d
+# asm 2: xorl 112(<m=%rsi),<in12=%edx
+xorl 112(%rsi),%edx
+
+# qhasm:   (uint32) in13 ^= *(uint32 *) (m + 116)
+# asm 1: xorl 116(<m=int64#2),<in13=int64#4d
+# asm 2: xorl 116(<m=%rsi),<in13=%ecx
+xorl 116(%rsi),%ecx
+
+# qhasm:   (uint32) in14 ^= *(uint32 *) (m + 120)
+# asm 1: xorl 120(<m=int64#2),<in14=int64#5d
+# asm 2: xorl 120(<m=%rsi),<in14=%r8d
+xorl 120(%rsi),%r8d
+
+# qhasm:   (uint32) in15 ^= *(uint32 *) (m + 124)
+# asm 1: xorl 124(<m=int64#2),<in15=int64#6d
+# asm 2: xorl 124(<m=%rsi),<in15=%r9d
+xorl 124(%rsi),%r9d
+
+# qhasm:   *(uint32 *) (out + 112) = in12
+# asm 1: movl   <in12=int64#3d,112(<out=int64#1)
+# asm 2: movl   <in12=%edx,112(<out=%rdi)
+movl   %edx,112(%rdi)
+
+# qhasm:   *(uint32 *) (out + 116) = in13
+# asm 1: movl   <in13=int64#4d,116(<out=int64#1)
+# asm 2: movl   <in13=%ecx,116(<out=%rdi)
+movl   %ecx,116(%rdi)
+
+# qhasm:   *(uint32 *) (out + 120) = in14
+# asm 1: movl   <in14=int64#5d,120(<out=int64#1)
+# asm 2: movl   <in14=%r8d,120(<out=%rdi)
+movl   %r8d,120(%rdi)
+
+# qhasm:   *(uint32 *) (out + 124) = in15
+# asm 1: movl   <in15=int64#6d,124(<out=int64#1)
+# asm 2: movl   <in15=%r9d,124(<out=%rdi)
+movl   %r9d,124(%rdi)
+
+# qhasm:   in12 = z12
+# asm 1: movd   <z12=int6464#14,>in12=int64#3
+# asm 2: movd   <z12=%xmm13,>in12=%rdx
+movd   %xmm13,%rdx
+
+# qhasm:   in13 = z13
+# asm 1: movd   <z13=int6464#10,>in13=int64#4
+# asm 2: movd   <z13=%xmm9,>in13=%rcx
+movd   %xmm9,%rcx
+
+# qhasm:   in14 = z14
+# asm 1: movd   <z14=int6464#4,>in14=int64#5
+# asm 2: movd   <z14=%xmm3,>in14=%r8
+movd   %xmm3,%r8
+
+# qhasm:   in15 = z15
+# asm 1: movd   <z15=int6464#3,>in15=int64#6
+# asm 2: movd   <z15=%xmm2,>in15=%r9
+movd   %xmm2,%r9
+
+# qhasm:   z12 <<<= 96
+# asm 1: pshufd $0x39,<z12=int6464#14,<z12=int6464#14
+# asm 2: pshufd $0x39,<z12=%xmm13,<z12=%xmm13
+pshufd $0x39,%xmm13,%xmm13
+
+# qhasm:   z13 <<<= 96
+# asm 1: pshufd $0x39,<z13=int6464#10,<z13=int6464#10
+# asm 2: pshufd $0x39,<z13=%xmm9,<z13=%xmm9
+pshufd $0x39,%xmm9,%xmm9
+
+# qhasm:   z14 <<<= 96
+# asm 1: pshufd $0x39,<z14=int6464#4,<z14=int6464#4
+# asm 2: pshufd $0x39,<z14=%xmm3,<z14=%xmm3
+pshufd $0x39,%xmm3,%xmm3
+
+# qhasm:   z15 <<<= 96
+# asm 1: pshufd $0x39,<z15=int6464#3,<z15=int6464#3
+# asm 2: pshufd $0x39,<z15=%xmm2,<z15=%xmm2
+pshufd $0x39,%xmm2,%xmm2
+
+# qhasm:   (uint32) in12 ^= *(uint32 *) (m + 176)
+# asm 1: xorl 176(<m=int64#2),<in12=int64#3d
+# asm 2: xorl 176(<m=%rsi),<in12=%edx
+xorl 176(%rsi),%edx
+
+# qhasm:   (uint32) in13 ^= *(uint32 *) (m + 180)
+# asm 1: xorl 180(<m=int64#2),<in13=int64#4d
+# asm 2: xorl 180(<m=%rsi),<in13=%ecx
+xorl 180(%rsi),%ecx
+
+# qhasm:   (uint32) in14 ^= *(uint32 *) (m + 184)
+# asm 1: xorl 184(<m=int64#2),<in14=int64#5d
+# asm 2: xorl 184(<m=%rsi),<in14=%r8d
+xorl 184(%rsi),%r8d
+
+# qhasm:   (uint32) in15 ^= *(uint32 *) (m + 188)
+# asm 1: xorl 188(<m=int64#2),<in15=int64#6d
+# asm 2: xorl 188(<m=%rsi),<in15=%r9d
+xorl 188(%rsi),%r9d
+
+# qhasm:   *(uint32 *) (out + 176) = in12
+# asm 1: movl   <in12=int64#3d,176(<out=int64#1)
+# asm 2: movl   <in12=%edx,176(<out=%rdi)
+movl   %edx,176(%rdi)
+
+# qhasm:   *(uint32 *) (out + 180) = in13
+# asm 1: movl   <in13=int64#4d,180(<out=int64#1)
+# asm 2: movl   <in13=%ecx,180(<out=%rdi)
+movl   %ecx,180(%rdi)
+
+# qhasm:   *(uint32 *) (out + 184) = in14
+# asm 1: movl   <in14=int64#5d,184(<out=int64#1)
+# asm 2: movl   <in14=%r8d,184(<out=%rdi)
+movl   %r8d,184(%rdi)
+
+# qhasm:   *(uint32 *) (out + 188) = in15
+# asm 1: movl   <in15=int64#6d,188(<out=int64#1)
+# asm 2: movl   <in15=%r9d,188(<out=%rdi)
+movl   %r9d,188(%rdi)
+
+# qhasm:   in12 = z12
+# asm 1: movd   <z12=int6464#14,>in12=int64#3
+# asm 2: movd   <z12=%xmm13,>in12=%rdx
+movd   %xmm13,%rdx
+
+# qhasm:   in13 = z13
+# asm 1: movd   <z13=int6464#10,>in13=int64#4
+# asm 2: movd   <z13=%xmm9,>in13=%rcx
+movd   %xmm9,%rcx
+
+# qhasm:   in14 = z14
+# asm 1: movd   <z14=int6464#4,>in14=int64#5
+# asm 2: movd   <z14=%xmm3,>in14=%r8
+movd   %xmm3,%r8
+
+# qhasm:   in15 = z15
+# asm 1: movd   <z15=int6464#3,>in15=int64#6
+# asm 2: movd   <z15=%xmm2,>in15=%r9
+movd   %xmm2,%r9
+
+# qhasm:   (uint32) in12 ^= *(uint32 *) (m + 240)
+# asm 1: xorl 240(<m=int64#2),<in12=int64#3d
+# asm 2: xorl 240(<m=%rsi),<in12=%edx
+xorl 240(%rsi),%edx
+
+# qhasm:   (uint32) in13 ^= *(uint32 *) (m + 244)
+# asm 1: xorl 244(<m=int64#2),<in13=int64#4d
+# asm 2: xorl 244(<m=%rsi),<in13=%ecx
+xorl 244(%rsi),%ecx
+
+# qhasm:   (uint32) in14 ^= *(uint32 *) (m + 248)
+# asm 1: xorl 248(<m=int64#2),<in14=int64#5d
+# asm 2: xorl 248(<m=%rsi),<in14=%r8d
+xorl 248(%rsi),%r8d
+
+# qhasm:   (uint32) in15 ^= *(uint32 *) (m + 252)
+# asm 1: xorl 252(<m=int64#2),<in15=int64#6d
+# asm 2: xorl 252(<m=%rsi),<in15=%r9d
+xorl 252(%rsi),%r9d
+
+# qhasm:   *(uint32 *) (out + 240) = in12
+# asm 1: movl   <in12=int64#3d,240(<out=int64#1)
+# asm 2: movl   <in12=%edx,240(<out=%rdi)
+movl   %edx,240(%rdi)
+
+# qhasm:   *(uint32 *) (out + 244) = in13
+# asm 1: movl   <in13=int64#4d,244(<out=int64#1)
+# asm 2: movl   <in13=%ecx,244(<out=%rdi)
+movl   %ecx,244(%rdi)
+
+# qhasm:   *(uint32 *) (out + 248) = in14
+# asm 1: movl   <in14=int64#5d,248(<out=int64#1)
+# asm 2: movl   <in14=%r8d,248(<out=%rdi)
+movl   %r8d,248(%rdi)
+
+# qhasm:   *(uint32 *) (out + 252) = in15
+# asm 1: movl   <in15=int64#6d,252(<out=int64#1)
+# asm 2: movl   <in15=%r9d,252(<out=%rdi)
+movl   %r9d,252(%rdi)
+
+# qhasm:   bytes = bytes_backup
+# asm 1: movq <bytes_backup=stack64#8,>bytes=int64#6
+# asm 2: movq <bytes_backup=408(%rsp),>bytes=%r9
+movq 408(%rsp),%r9
+
+# qhasm:   bytes -= 256
+# asm 1: sub  $256,<bytes=int64#6
+# asm 2: sub  $256,<bytes=%r9
+sub  $256,%r9
+
+# qhasm:   m += 256
+# asm 1: add  $256,<m=int64#2
+# asm 2: add  $256,<m=%rsi
+add  $256,%rsi
+
+# qhasm:   out += 256
+# asm 1: add  $256,<out=int64#1
+# asm 2: add  $256,<out=%rdi
+add  $256,%rdi
+
+# qhasm:                            unsigned<? bytes - 256
+# asm 1: cmp  $256,<bytes=int64#6
+# asm 2: cmp  $256,<bytes=%r9
+cmp  $256,%r9
+# comment:fp stack unchanged by jump
+
+# qhasm:   goto bytesatleast256 if !unsigned<
+jae ._bytesatleast256
+
+# qhasm:                 unsigned>? bytes - 0
+# asm 1: cmp  $0,<bytes=int64#6
+# asm 2: cmp  $0,<bytes=%r9
+cmp  $0,%r9
+# comment:fp stack unchanged by jump
+
+# qhasm:   goto done if !unsigned>
+jbe ._done
+# comment:fp stack unchanged by fallthrough
+
+# qhasm: bytesbetween1and255:
+._bytesbetween1and255:
+
+# qhasm:                   unsigned<? bytes - 64
+# asm 1: cmp  $64,<bytes=int64#6
+# asm 2: cmp  $64,<bytes=%r9
+cmp  $64,%r9
+# comment:fp stack unchanged by jump
+
+# qhasm:   goto nocopy if !unsigned<
+jae ._nocopy
+
+# qhasm:     ctarget = out
+# asm 1: mov  <out=int64#1,>ctarget=int64#3
+# asm 2: mov  <out=%rdi,>ctarget=%rdx
+mov  %rdi,%rdx
+
+# qhasm:     out = &tmp
+# asm 1: leaq <tmp=stack512#1,>out=int64#1
+# asm 2: leaq <tmp=416(%rsp),>out=%rdi
+leaq 416(%rsp),%rdi
+
+# qhasm:     i = bytes
+# asm 1: mov  <bytes=int64#6,>i=int64#4
+# asm 2: mov  <bytes=%r9,>i=%rcx
+mov  %r9,%rcx
+
+# qhasm:     while (i) { *out++ = *m++; --i }
+rep movsb
+
+# qhasm:     out = &tmp
+# asm 1: leaq <tmp=stack512#1,>out=int64#1
+# asm 2: leaq <tmp=416(%rsp),>out=%rdi
+leaq 416(%rsp),%rdi
+
+# qhasm:     m = &tmp
+# asm 1: leaq <tmp=stack512#1,>m=int64#2
+# asm 2: leaq <tmp=416(%rsp),>m=%rsi
+leaq 416(%rsp),%rsi
+# comment:fp stack unchanged by fallthrough
+
+# qhasm:   nocopy:
+._nocopy:
+
+# qhasm:   bytes_backup = bytes
+# asm 1: movq <bytes=int64#6,>bytes_backup=stack64#8
+# asm 2: movq <bytes=%r9,>bytes_backup=408(%rsp)
+movq %r9,408(%rsp)
+
+# qhasm: diag0 = x0
+# asm 1: movdqa <x0=stack128#4,>diag0=int6464#1
+# asm 2: movdqa <x0=48(%rsp),>diag0=%xmm0
+movdqa 48(%rsp),%xmm0
+
+# qhasm: diag1 = x1
+# asm 1: movdqa <x1=stack128#1,>diag1=int6464#2
+# asm 2: movdqa <x1=0(%rsp),>diag1=%xmm1
+movdqa 0(%rsp),%xmm1
+
+# qhasm: diag2 = x2
+# asm 1: movdqa <x2=stack128#2,>diag2=int6464#3
+# asm 2: movdqa <x2=16(%rsp),>diag2=%xmm2
+movdqa 16(%rsp),%xmm2
+
+# qhasm: diag3 = x3
+# asm 1: movdqa <x3=stack128#3,>diag3=int6464#4
+# asm 2: movdqa <x3=32(%rsp),>diag3=%xmm3
+movdqa 32(%rsp),%xmm3
+
+# qhasm:                     a0 = diag1
+# asm 1: movdqa <diag1=int6464#2,>a0=int6464#5
+# asm 2: movdqa <diag1=%xmm1,>a0=%xmm4
+movdqa %xmm1,%xmm4
+
+# qhasm: i = 20
+# asm 1: mov  $20,>i=int64#4
+# asm 2: mov  $20,>i=%rcx
+mov  $20,%rcx
+
+# qhasm: mainloop2:
+._mainloop2:
+
+# qhasm: uint32323232        a0 += diag0
+# asm 1: paddd <diag0=int6464#1,<a0=int6464#5
+# asm 2: paddd <diag0=%xmm0,<a0=%xmm4
+paddd %xmm0,%xmm4
+
+# qhasm:                                 a1 = diag0
+# asm 1: movdqa <diag0=int6464#1,>a1=int6464#6
+# asm 2: movdqa <diag0=%xmm0,>a1=%xmm5
+movdqa %xmm0,%xmm5
+
+# qhasm:                     b0 = a0
+# asm 1: movdqa <a0=int6464#5,>b0=int6464#7
+# asm 2: movdqa <a0=%xmm4,>b0=%xmm6
+movdqa %xmm4,%xmm6
+
+# qhasm: uint32323232        a0 <<= 7
+# asm 1: pslld $7,<a0=int6464#5
+# asm 2: pslld $7,<a0=%xmm4
+pslld $7,%xmm4
+
+# qhasm: uint32323232        b0 >>= 25
+# asm 1: psrld $25,<b0=int6464#7
+# asm 2: psrld $25,<b0=%xmm6
+psrld $25,%xmm6
+
+# qhasm:                 diag3 ^= a0
+# asm 1: pxor  <a0=int6464#5,<diag3=int6464#4
+# asm 2: pxor  <a0=%xmm4,<diag3=%xmm3
+pxor  %xmm4,%xmm3
+
+# qhasm:                 diag3 ^= b0
+# asm 1: pxor  <b0=int6464#7,<diag3=int6464#4
+# asm 2: pxor  <b0=%xmm6,<diag3=%xmm3
+pxor  %xmm6,%xmm3
+
+# qhasm: uint32323232                        a1 += diag3
+# asm 1: paddd <diag3=int6464#4,<a1=int6464#6
+# asm 2: paddd <diag3=%xmm3,<a1=%xmm5
+paddd %xmm3,%xmm5
+
+# qhasm:                                                 a2 = diag3
+# asm 1: movdqa <diag3=int6464#4,>a2=int6464#5
+# asm 2: movdqa <diag3=%xmm3,>a2=%xmm4
+movdqa %xmm3,%xmm4
+
+# qhasm:                                     b1 = a1
+# asm 1: movdqa <a1=int6464#6,>b1=int6464#7
+# asm 2: movdqa <a1=%xmm5,>b1=%xmm6
+movdqa %xmm5,%xmm6
+
+# qhasm: uint32323232                        a1 <<= 9
+# asm 1: pslld $9,<a1=int6464#6
+# asm 2: pslld $9,<a1=%xmm5
+pslld $9,%xmm5
+
+# qhasm: uint32323232                        b1 >>= 23
+# asm 1: psrld $23,<b1=int6464#7
+# asm 2: psrld $23,<b1=%xmm6
+psrld $23,%xmm6
+
+# qhasm:                                 diag2 ^= a1
+# asm 1: pxor  <a1=int6464#6,<diag2=int6464#3
+# asm 2: pxor  <a1=%xmm5,<diag2=%xmm2
+pxor  %xmm5,%xmm2
+
+# qhasm:                 diag3 <<<= 32
+# asm 1: pshufd $0x93,<diag3=int6464#4,<diag3=int6464#4
+# asm 2: pshufd $0x93,<diag3=%xmm3,<diag3=%xmm3
+pshufd $0x93,%xmm3,%xmm3
+
+# qhasm:                                 diag2 ^= b1
+# asm 1: pxor  <b1=int6464#7,<diag2=int6464#3
+# asm 2: pxor  <b1=%xmm6,<diag2=%xmm2
+pxor  %xmm6,%xmm2
+
+# qhasm: uint32323232                                        a2 += diag2
+# asm 1: paddd <diag2=int6464#3,<a2=int6464#5
+# asm 2: paddd <diag2=%xmm2,<a2=%xmm4
+paddd %xmm2,%xmm4
+
+# qhasm:                                                                 a3 = diag2
+# asm 1: movdqa <diag2=int6464#3,>a3=int6464#6
+# asm 2: movdqa <diag2=%xmm2,>a3=%xmm5
+movdqa %xmm2,%xmm5
+
+# qhasm:                                                     b2 = a2
+# asm 1: movdqa <a2=int6464#5,>b2=int6464#7
+# asm 2: movdqa <a2=%xmm4,>b2=%xmm6
+movdqa %xmm4,%xmm6
+
+# qhasm: uint32323232                                        a2 <<= 13
+# asm 1: pslld $13,<a2=int6464#5
+# asm 2: pslld $13,<a2=%xmm4
+pslld $13,%xmm4
+
+# qhasm: uint32323232                                        b2 >>= 19
+# asm 1: psrld $19,<b2=int6464#7
+# asm 2: psrld $19,<b2=%xmm6
+psrld $19,%xmm6
+
+# qhasm:                                                 diag1 ^= a2
+# asm 1: pxor  <a2=int6464#5,<diag1=int6464#2
+# asm 2: pxor  <a2=%xmm4,<diag1=%xmm1
+pxor  %xmm4,%xmm1
+
+# qhasm:                                 diag2 <<<= 64
+# asm 1: pshufd $0x4e,<diag2=int6464#3,<diag2=int6464#3
+# asm 2: pshufd $0x4e,<diag2=%xmm2,<diag2=%xmm2
+pshufd $0x4e,%xmm2,%xmm2
+
+# qhasm:                                                 diag1 ^= b2
+# asm 1: pxor  <b2=int6464#7,<diag1=int6464#2
+# asm 2: pxor  <b2=%xmm6,<diag1=%xmm1
+pxor  %xmm6,%xmm1
+
+# qhasm: uint32323232                                                        a3 += diag1
+# asm 1: paddd <diag1=int6464#2,<a3=int6464#6
+# asm 2: paddd <diag1=%xmm1,<a3=%xmm5
+paddd %xmm1,%xmm5
+
+# qhasm:                 a4 = diag3
+# asm 1: movdqa <diag3=int6464#4,>a4=int6464#5
+# asm 2: movdqa <diag3=%xmm3,>a4=%xmm4
+movdqa %xmm3,%xmm4
+
+# qhasm:                                                                     b3 = a3
+# asm 1: movdqa <a3=int6464#6,>b3=int6464#7
+# asm 2: movdqa <a3=%xmm5,>b3=%xmm6
+movdqa %xmm5,%xmm6
+
+# qhasm: uint32323232                                                        a3 <<= 18
+# asm 1: pslld $18,<a3=int6464#6
+# asm 2: pslld $18,<a3=%xmm5
+pslld $18,%xmm5
+
+# qhasm: uint32323232                                                        b3 >>= 14
+# asm 1: psrld $14,<b3=int6464#7
+# asm 2: psrld $14,<b3=%xmm6
+psrld $14,%xmm6
+
+# qhasm:                                                                 diag0 ^= a3
+# asm 1: pxor  <a3=int6464#6,<diag0=int6464#1
+# asm 2: pxor  <a3=%xmm5,<diag0=%xmm0
+pxor  %xmm5,%xmm0
+
+# qhasm:                                                 diag1 <<<= 96
+# asm 1: pshufd $0x39,<diag1=int6464#2,<diag1=int6464#2
+# asm 2: pshufd $0x39,<diag1=%xmm1,<diag1=%xmm1
+pshufd $0x39,%xmm1,%xmm1
+
+# qhasm:                                                                 diag0 ^= b3
+# asm 1: pxor  <b3=int6464#7,<diag0=int6464#1
+# asm 2: pxor  <b3=%xmm6,<diag0=%xmm0
+pxor  %xmm6,%xmm0
+
+# qhasm: uint32323232        a4 += diag0
+# asm 1: paddd <diag0=int6464#1,<a4=int6464#5
+# asm 2: paddd <diag0=%xmm0,<a4=%xmm4
+paddd %xmm0,%xmm4
+
+# qhasm:                                 a5 = diag0
+# asm 1: movdqa <diag0=int6464#1,>a5=int6464#6
+# asm 2: movdqa <diag0=%xmm0,>a5=%xmm5
+movdqa %xmm0,%xmm5
+
+# qhasm:                     b4 = a4
+# asm 1: movdqa <a4=int6464#5,>b4=int6464#7
+# asm 2: movdqa <a4=%xmm4,>b4=%xmm6
+movdqa %xmm4,%xmm6
+
+# qhasm: uint32323232        a4 <<= 7
+# asm 1: pslld $7,<a4=int6464#5
+# asm 2: pslld $7,<a4=%xmm4
+pslld $7,%xmm4
+
+# qhasm: uint32323232        b4 >>= 25
+# asm 1: psrld $25,<b4=int6464#7
+# asm 2: psrld $25,<b4=%xmm6
+psrld $25,%xmm6
+
+# qhasm:                 diag1 ^= a4
+# asm 1: pxor  <a4=int6464#5,<diag1=int6464#2
+# asm 2: pxor  <a4=%xmm4,<diag1=%xmm1
+pxor  %xmm4,%xmm1
+
+# qhasm:                 diag1 ^= b4
+# asm 1: pxor  <b4=int6464#7,<diag1=int6464#2
+# asm 2: pxor  <b4=%xmm6,<diag1=%xmm1
+pxor  %xmm6,%xmm1
+
+# qhasm: uint32323232                        a5 += diag1
+# asm 1: paddd <diag1=int6464#2,<a5=int6464#6
+# asm 2: paddd <diag1=%xmm1,<a5=%xmm5
+paddd %xmm1,%xmm5
+
+# qhasm:                                                 a6 = diag1
+# asm 1: movdqa <diag1=int6464#2,>a6=int6464#5
+# asm 2: movdqa <diag1=%xmm1,>a6=%xmm4
+movdqa %xmm1,%xmm4
+
+# qhasm:                                     b5 = a5
+# asm 1: movdqa <a5=int6464#6,>b5=int6464#7
+# asm 2: movdqa <a5=%xmm5,>b5=%xmm6
+movdqa %xmm5,%xmm6
+
+# qhasm: uint32323232                        a5 <<= 9
+# asm 1: pslld $9,<a5=int6464#6
+# asm 2: pslld $9,<a5=%xmm5
+pslld $9,%xmm5
+
+# qhasm: uint32323232                        b5 >>= 23
+# asm 1: psrld $23,<b5=int6464#7
+# asm 2: psrld $23,<b5=%xmm6
+psrld $23,%xmm6
+
+# qhasm:                                 diag2 ^= a5
+# asm 1: pxor  <a5=int6464#6,<diag2=int6464#3
+# asm 2: pxor  <a5=%xmm5,<diag2=%xmm2
+pxor  %xmm5,%xmm2
+
+# qhasm:                 diag1 <<<= 32
+# asm 1: pshufd $0x93,<diag1=int6464#2,<diag1=int6464#2
+# asm 2: pshufd $0x93,<diag1=%xmm1,<diag1=%xmm1
+pshufd $0x93,%xmm1,%xmm1
+
+# qhasm:                                 diag2 ^= b5
+# asm 1: pxor  <b5=int6464#7,<diag2=int6464#3
+# asm 2: pxor  <b5=%xmm6,<diag2=%xmm2
+pxor  %xmm6,%xmm2
+
+# qhasm: uint32323232                                        a6 += diag2
+# asm 1: paddd <diag2=int6464#3,<a6=int6464#5
+# asm 2: paddd <diag2=%xmm2,<a6=%xmm4
+paddd %xmm2,%xmm4
+
+# qhasm:                                                                 a7 = diag2
+# asm 1: movdqa <diag2=int6464#3,>a7=int6464#6
+# asm 2: movdqa <diag2=%xmm2,>a7=%xmm5
+movdqa %xmm2,%xmm5
+
+# qhasm:                                                     b6 = a6
+# asm 1: movdqa <a6=int6464#5,>b6=int6464#7
+# asm 2: movdqa <a6=%xmm4,>b6=%xmm6
+movdqa %xmm4,%xmm6
+
+# qhasm: uint32323232                                        a6 <<= 13
+# asm 1: pslld $13,<a6=int6464#5
+# asm 2: pslld $13,<a6=%xmm4
+pslld $13,%xmm4
+
+# qhasm: uint32323232                                        b6 >>= 19
+# asm 1: psrld $19,<b6=int6464#7
+# asm 2: psrld $19,<b6=%xmm6
+psrld $19,%xmm6
+
+# qhasm:                                                 diag3 ^= a6
+# asm 1: pxor  <a6=int6464#5,<diag3=int6464#4
+# asm 2: pxor  <a6=%xmm4,<diag3=%xmm3
+pxor  %xmm4,%xmm3
+
+# qhasm:                                 diag2 <<<= 64
+# asm 1: pshufd $0x4e,<diag2=int6464#3,<diag2=int6464#3
+# asm 2: pshufd $0x4e,<diag2=%xmm2,<diag2=%xmm2
+pshufd $0x4e,%xmm2,%xmm2
+
+# qhasm:                                                 diag3 ^= b6
+# asm 1: pxor  <b6=int6464#7,<diag3=int6464#4
+# asm 2: pxor  <b6=%xmm6,<diag3=%xmm3
+pxor  %xmm6,%xmm3
+
+# qhasm: uint32323232                                                        a7 += diag3
+# asm 1: paddd <diag3=int6464#4,<a7=int6464#6
+# asm 2: paddd <diag3=%xmm3,<a7=%xmm5
+paddd %xmm3,%xmm5
+
+# qhasm:                 a0 = diag1
+# asm 1: movdqa <diag1=int6464#2,>a0=int6464#5
+# asm 2: movdqa <diag1=%xmm1,>a0=%xmm4
+movdqa %xmm1,%xmm4
+
+# qhasm:                                                                     b7 = a7
+# asm 1: movdqa <a7=int6464#6,>b7=int6464#7
+# asm 2: movdqa <a7=%xmm5,>b7=%xmm6
+movdqa %xmm5,%xmm6
+
+# qhasm: uint32323232                                                        a7 <<= 18
+# asm 1: pslld $18,<a7=int6464#6
+# asm 2: pslld $18,<a7=%xmm5
+pslld $18,%xmm5
+
+# qhasm: uint32323232                                                        b7 >>= 14
+# asm 1: psrld $14,<b7=int6464#7
+# asm 2: psrld $14,<b7=%xmm6
+psrld $14,%xmm6
+
+# qhasm:                                                                 diag0 ^= a7
+# asm 1: pxor  <a7=int6464#6,<diag0=int6464#1
+# asm 2: pxor  <a7=%xmm5,<diag0=%xmm0
+pxor  %xmm5,%xmm0
+
+# qhasm:                                                 diag3 <<<= 96
+# asm 1: pshufd $0x39,<diag3=int6464#4,<diag3=int6464#4
+# asm 2: pshufd $0x39,<diag3=%xmm3,<diag3=%xmm3
+pshufd $0x39,%xmm3,%xmm3
+
+# qhasm:                                                                 diag0 ^= b7
+# asm 1: pxor  <b7=int6464#7,<diag0=int6464#1
+# asm 2: pxor  <b7=%xmm6,<diag0=%xmm0
+pxor  %xmm6,%xmm0
+
+# qhasm: uint32323232        a0 += diag0
+# asm 1: paddd <diag0=int6464#1,<a0=int6464#5
+# asm 2: paddd <diag0=%xmm0,<a0=%xmm4
+paddd %xmm0,%xmm4
+
+# qhasm:                                 a1 = diag0
+# asm 1: movdqa <diag0=int6464#1,>a1=int6464#6
+# asm 2: movdqa <diag0=%xmm0,>a1=%xmm5
+movdqa %xmm0,%xmm5
+
+# qhasm:                     b0 = a0
+# asm 1: movdqa <a0=int6464#5,>b0=int6464#7
+# asm 2: movdqa <a0=%xmm4,>b0=%xmm6
+movdqa %xmm4,%xmm6
+
+# qhasm: uint32323232        a0 <<= 7
+# asm 1: pslld $7,<a0=int6464#5
+# asm 2: pslld $7,<a0=%xmm4
+pslld $7,%xmm4
+
+# qhasm: uint32323232        b0 >>= 25
+# asm 1: psrld $25,<b0=int6464#7
+# asm 2: psrld $25,<b0=%xmm6
+psrld $25,%xmm6
+
+# qhasm:                 diag3 ^= a0
+# asm 1: pxor  <a0=int6464#5,<diag3=int6464#4
+# asm 2: pxor  <a0=%xmm4,<diag3=%xmm3
+pxor  %xmm4,%xmm3
+
+# qhasm:                 diag3 ^= b0
+# asm 1: pxor  <b0=int6464#7,<diag3=int6464#4
+# asm 2: pxor  <b0=%xmm6,<diag3=%xmm3
+pxor  %xmm6,%xmm3
+
+# qhasm: uint32323232                        a1 += diag3
+# asm 1: paddd <diag3=int6464#4,<a1=int6464#6
+# asm 2: paddd <diag3=%xmm3,<a1=%xmm5
+paddd %xmm3,%xmm5
+
+# qhasm:                                                 a2 = diag3
+# asm 1: movdqa <diag3=int6464#4,>a2=int6464#5
+# asm 2: movdqa <diag3=%xmm3,>a2=%xmm4
+movdqa %xmm3,%xmm4
+
+# qhasm:                                     b1 = a1
+# asm 1: movdqa <a1=int6464#6,>b1=int6464#7
+# asm 2: movdqa <a1=%xmm5,>b1=%xmm6
+movdqa %xmm5,%xmm6
+
+# qhasm: uint32323232                        a1 <<= 9
+# asm 1: pslld $9,<a1=int6464#6
+# asm 2: pslld $9,<a1=%xmm5
+pslld $9,%xmm5
+
+# qhasm: uint32323232                        b1 >>= 23
+# asm 1: psrld $23,<b1=int6464#7
+# asm 2: psrld $23,<b1=%xmm6
+psrld $23,%xmm6
+
+# qhasm:                                 diag2 ^= a1
+# asm 1: pxor  <a1=int6464#6,<diag2=int6464#3
+# asm 2: pxor  <a1=%xmm5,<diag2=%xmm2
+pxor  %xmm5,%xmm2
+
+# qhasm:                 diag3 <<<= 32
+# asm 1: pshufd $0x93,<diag3=int6464#4,<diag3=int6464#4
+# asm 2: pshufd $0x93,<diag3=%xmm3,<diag3=%xmm3
+pshufd $0x93,%xmm3,%xmm3
+
+# qhasm:                                 diag2 ^= b1
+# asm 1: pxor  <b1=int6464#7,<diag2=int6464#3
+# asm 2: pxor  <b1=%xmm6,<diag2=%xmm2
+pxor  %xmm6,%xmm2
+
+# qhasm: uint32323232                                        a2 += diag2
+# asm 1: paddd <diag2=int6464#3,<a2=int6464#5
+# asm 2: paddd <diag2=%xmm2,<a2=%xmm4
+paddd %xmm2,%xmm4
+
+# qhasm:                                                                 a3 = diag2
+# asm 1: movdqa <diag2=int6464#3,>a3=int6464#6
+# asm 2: movdqa <diag2=%xmm2,>a3=%xmm5
+movdqa %xmm2,%xmm5
+
+# qhasm:                                                     b2 = a2
+# asm 1: movdqa <a2=int6464#5,>b2=int6464#7
+# asm 2: movdqa <a2=%xmm4,>b2=%xmm6
+movdqa %xmm4,%xmm6
+
+# qhasm: uint32323232                                        a2 <<= 13
+# asm 1: pslld $13,<a2=int6464#5
+# asm 2: pslld $13,<a2=%xmm4
+pslld $13,%xmm4
+
+# qhasm: uint32323232                                        b2 >>= 19
+# asm 1: psrld $19,<b2=int6464#7
+# asm 2: psrld $19,<b2=%xmm6
+psrld $19,%xmm6
+
+# qhasm:                                                 diag1 ^= a2
+# asm 1: pxor  <a2=int6464#5,<diag1=int6464#2
+# asm 2: pxor  <a2=%xmm4,<diag1=%xmm1
+pxor  %xmm4,%xmm1
+
+# qhasm:                                 diag2 <<<= 64
+# asm 1: pshufd $0x4e,<diag2=int6464#3,<diag2=int6464#3
+# asm 2: pshufd $0x4e,<diag2=%xmm2,<diag2=%xmm2
+pshufd $0x4e,%xmm2,%xmm2
+
+# qhasm:                                                 diag1 ^= b2
+# asm 1: pxor  <b2=int6464#7,<diag1=int6464#2
+# asm 2: pxor  <b2=%xmm6,<diag1=%xmm1
+pxor  %xmm6,%xmm1
+
+# qhasm: uint32323232                                                        a3 += diag1
+# asm 1: paddd <diag1=int6464#2,<a3=int6464#6
+# asm 2: paddd <diag1=%xmm1,<a3=%xmm5
+paddd %xmm1,%xmm5
+
+# qhasm:                 a4 = diag3
+# asm 1: movdqa <diag3=int6464#4,>a4=int6464#5
+# asm 2: movdqa <diag3=%xmm3,>a4=%xmm4
+movdqa %xmm3,%xmm4
+
+# qhasm:                                                                     b3 = a3
+# asm 1: movdqa <a3=int6464#6,>b3=int6464#7
+# asm 2: movdqa <a3=%xmm5,>b3=%xmm6
+movdqa %xmm5,%xmm6
+
+# qhasm: uint32323232                                                        a3 <<= 18
+# asm 1: pslld $18,<a3=int6464#6
+# asm 2: pslld $18,<a3=%xmm5
+pslld $18,%xmm5
+
+# qhasm: uint32323232                                                        b3 >>= 14
+# asm 1: psrld $14,<b3=int6464#7
+# asm 2: psrld $14,<b3=%xmm6
+psrld $14,%xmm6
+
+# qhasm:                                                                 diag0 ^= a3
+# asm 1: pxor  <a3=int6464#6,<diag0=int6464#1
+# asm 2: pxor  <a3=%xmm5,<diag0=%xmm0
+pxor  %xmm5,%xmm0
+
+# qhasm:                                                 diag1 <<<= 96
+# asm 1: pshufd $0x39,<diag1=int6464#2,<diag1=int6464#2
+# asm 2: pshufd $0x39,<diag1=%xmm1,<diag1=%xmm1
+pshufd $0x39,%xmm1,%xmm1
+
+# qhasm:                                                                 diag0 ^= b3
+# asm 1: pxor  <b3=int6464#7,<diag0=int6464#1
+# asm 2: pxor  <b3=%xmm6,<diag0=%xmm0
+pxor  %xmm6,%xmm0
+
+# qhasm: uint32323232        a4 += diag0
+# asm 1: paddd <diag0=int6464#1,<a4=int6464#5
+# asm 2: paddd <diag0=%xmm0,<a4=%xmm4
+paddd %xmm0,%xmm4
+
+# qhasm:                                 a5 = diag0
+# asm 1: movdqa <diag0=int6464#1,>a5=int6464#6
+# asm 2: movdqa <diag0=%xmm0,>a5=%xmm5
+movdqa %xmm0,%xmm5
+
+# qhasm:                     b4 = a4
+# asm 1: movdqa <a4=int6464#5,>b4=int6464#7
+# asm 2: movdqa <a4=%xmm4,>b4=%xmm6
+movdqa %xmm4,%xmm6
+
+# qhasm: uint32323232        a4 <<= 7
+# asm 1: pslld $7,<a4=int6464#5
+# asm 2: pslld $7,<a4=%xmm4
+pslld $7,%xmm4
+
+# qhasm: uint32323232        b4 >>= 25
+# asm 1: psrld $25,<b4=int6464#7
+# asm 2: psrld $25,<b4=%xmm6
+psrld $25,%xmm6
+
+# qhasm:                 diag1 ^= a4
+# asm 1: pxor  <a4=int6464#5,<diag1=int6464#2
+# asm 2: pxor  <a4=%xmm4,<diag1=%xmm1
+pxor  %xmm4,%xmm1
+
+# qhasm:                 diag1 ^= b4
+# asm 1: pxor  <b4=int6464#7,<diag1=int6464#2
+# asm 2: pxor  <b4=%xmm6,<diag1=%xmm1
+pxor  %xmm6,%xmm1
+
+# qhasm: uint32323232                        a5 += diag1
+# asm 1: paddd <diag1=int6464#2,<a5=int6464#6
+# asm 2: paddd <diag1=%xmm1,<a5=%xmm5
+paddd %xmm1,%xmm5
+
+# qhasm:                                                 a6 = diag1
+# asm 1: movdqa <diag1=int6464#2,>a6=int6464#5
+# asm 2: movdqa <diag1=%xmm1,>a6=%xmm4
+movdqa %xmm1,%xmm4
+
+# qhasm:                                     b5 = a5
+# asm 1: movdqa <a5=int6464#6,>b5=int6464#7
+# asm 2: movdqa <a5=%xmm5,>b5=%xmm6
+movdqa %xmm5,%xmm6
+
+# qhasm: uint32323232                        a5 <<= 9
+# asm 1: pslld $9,<a5=int6464#6
+# asm 2: pslld $9,<a5=%xmm5
+pslld $9,%xmm5
+
+# qhasm: uint32323232                        b5 >>= 23
+# asm 1: psrld $23,<b5=int6464#7
+# asm 2: psrld $23,<b5=%xmm6
+psrld $23,%xmm6
+
+# qhasm:                                 diag2 ^= a5
+# asm 1: pxor  <a5=int6464#6,<diag2=int6464#3
+# asm 2: pxor  <a5=%xmm5,<diag2=%xmm2
+pxor  %xmm5,%xmm2
+
+# qhasm:                 diag1 <<<= 32
+# asm 1: pshufd $0x93,<diag1=int6464#2,<diag1=int6464#2
+# asm 2: pshufd $0x93,<diag1=%xmm1,<diag1=%xmm1
+pshufd $0x93,%xmm1,%xmm1
+
+# qhasm:                                 diag2 ^= b5
+# asm 1: pxor  <b5=int6464#7,<diag2=int6464#3
+# asm 2: pxor  <b5=%xmm6,<diag2=%xmm2
+pxor  %xmm6,%xmm2
+
+# qhasm: uint32323232                                        a6 += diag2
+# asm 1: paddd <diag2=int6464#3,<a6=int6464#5
+# asm 2: paddd <diag2=%xmm2,<a6=%xmm4
+paddd %xmm2,%xmm4
+
+# qhasm:                                                                 a7 = diag2
+# asm 1: movdqa <diag2=int6464#3,>a7=int6464#6
+# asm 2: movdqa <diag2=%xmm2,>a7=%xmm5
+movdqa %xmm2,%xmm5
+
+# qhasm:                                                     b6 = a6
+# asm 1: movdqa <a6=int6464#5,>b6=int6464#7
+# asm 2: movdqa <a6=%xmm4,>b6=%xmm6
+movdqa %xmm4,%xmm6
+
+# qhasm: uint32323232                                        a6 <<= 13
+# asm 1: pslld $13,<a6=int6464#5
+# asm 2: pslld $13,<a6=%xmm4
+pslld $13,%xmm4
+
+# qhasm: uint32323232                                        b6 >>= 19
+# asm 1: psrld $19,<b6=int6464#7
+# asm 2: psrld $19,<b6=%xmm6
+psrld $19,%xmm6
+
+# qhasm:                                                 diag3 ^= a6
+# asm 1: pxor  <a6=int6464#5,<diag3=int6464#4
+# asm 2: pxor  <a6=%xmm4,<diag3=%xmm3
+pxor  %xmm4,%xmm3
+
+# qhasm:                                 diag2 <<<= 64
+# asm 1: pshufd $0x4e,<diag2=int6464#3,<diag2=int6464#3
+# asm 2: pshufd $0x4e,<diag2=%xmm2,<diag2=%xmm2
+pshufd $0x4e,%xmm2,%xmm2
+
+# qhasm:                                                 diag3 ^= b6
+# asm 1: pxor  <b6=int6464#7,<diag3=int6464#4
+# asm 2: pxor  <b6=%xmm6,<diag3=%xmm3
+pxor  %xmm6,%xmm3
+
+# qhasm:                  unsigned>? i -= 4
+# asm 1: sub  $4,<i=int64#4
+# asm 2: sub  $4,<i=%rcx
+sub  $4,%rcx
+
+# qhasm: uint32323232                                                        a7 += diag3
+# asm 1: paddd <diag3=int6464#4,<a7=int6464#6
+# asm 2: paddd <diag3=%xmm3,<a7=%xmm5
+paddd %xmm3,%xmm5
+
+# qhasm:                 a0 = diag1
+# asm 1: movdqa <diag1=int6464#2,>a0=int6464#5
+# asm 2: movdqa <diag1=%xmm1,>a0=%xmm4
+movdqa %xmm1,%xmm4
+
+# qhasm:                                                                     b7 = a7
+# asm 1: movdqa <a7=int6464#6,>b7=int6464#7
+# asm 2: movdqa <a7=%xmm5,>b7=%xmm6
+movdqa %xmm5,%xmm6
+
+# qhasm: uint32323232                                                        a7 <<= 18
+# asm 1: pslld $18,<a7=int6464#6
+# asm 2: pslld $18,<a7=%xmm5
+pslld $18,%xmm5
+
+# qhasm:                 b0 = 0
+# asm 1: pxor   >b0=int6464#8,>b0=int6464#8
+# asm 2: pxor   >b0=%xmm7,>b0=%xmm7
+pxor   %xmm7,%xmm7
+
+# qhasm: uint32323232                                                        b7 >>= 14
+# asm 1: psrld $14,<b7=int6464#7
+# asm 2: psrld $14,<b7=%xmm6
+psrld $14,%xmm6
+
+# qhasm:                                                                 diag0 ^= a7
+# asm 1: pxor  <a7=int6464#6,<diag0=int6464#1
+# asm 2: pxor  <a7=%xmm5,<diag0=%xmm0
+pxor  %xmm5,%xmm0
+
+# qhasm:                                                 diag3 <<<= 96
+# asm 1: pshufd $0x39,<diag3=int6464#4,<diag3=int6464#4
+# asm 2: pshufd $0x39,<diag3=%xmm3,<diag3=%xmm3
+pshufd $0x39,%xmm3,%xmm3
+
+# qhasm:                                                                 diag0 ^= b7
+# asm 1: pxor  <b7=int6464#7,<diag0=int6464#1
+# asm 2: pxor  <b7=%xmm6,<diag0=%xmm0
+pxor  %xmm6,%xmm0
+# comment:fp stack unchanged by jump
+
+# qhasm: goto mainloop2 if unsigned>
+ja ._mainloop2
+
+# qhasm: uint32323232 diag0 += x0
+# asm 1: paddd <x0=stack128#4,<diag0=int6464#1
+# asm 2: paddd <x0=48(%rsp),<diag0=%xmm0
+paddd 48(%rsp),%xmm0
+
+# qhasm: uint32323232 diag1 += x1
+# asm 1: paddd <x1=stack128#1,<diag1=int6464#2
+# asm 2: paddd <x1=0(%rsp),<diag1=%xmm1
+paddd 0(%rsp),%xmm1
+
+# qhasm: uint32323232 diag2 += x2
+# asm 1: paddd <x2=stack128#2,<diag2=int6464#3
+# asm 2: paddd <x2=16(%rsp),<diag2=%xmm2
+paddd 16(%rsp),%xmm2
+
+# qhasm: uint32323232 diag3 += x3
+# asm 1: paddd <x3=stack128#3,<diag3=int6464#4
+# asm 2: paddd <x3=32(%rsp),<diag3=%xmm3
+paddd 32(%rsp),%xmm3
+
+# qhasm: in0 = diag0
+# asm 1: movd   <diag0=int6464#1,>in0=int64#4
+# asm 2: movd   <diag0=%xmm0,>in0=%rcx
+movd   %xmm0,%rcx
+
+# qhasm: in12 = diag1
+# asm 1: movd   <diag1=int6464#2,>in12=int64#5
+# asm 2: movd   <diag1=%xmm1,>in12=%r8
+movd   %xmm1,%r8
+
+# qhasm: in8 = diag2
+# asm 1: movd   <diag2=int6464#3,>in8=int64#6
+# asm 2: movd   <diag2=%xmm2,>in8=%r9
+movd   %xmm2,%r9
+
+# qhasm: in4 = diag3
+# asm 1: movd   <diag3=int6464#4,>in4=int64#7
+# asm 2: movd   <diag3=%xmm3,>in4=%rax
+movd   %xmm3,%rax
+
+# qhasm: diag0 <<<= 96
+# asm 1: pshufd $0x39,<diag0=int6464#1,<diag0=int6464#1
+# asm 2: pshufd $0x39,<diag0=%xmm0,<diag0=%xmm0
+pshufd $0x39,%xmm0,%xmm0
+
+# qhasm: diag1 <<<= 96
+# asm 1: pshufd $0x39,<diag1=int6464#2,<diag1=int6464#2
+# asm 2: pshufd $0x39,<diag1=%xmm1,<diag1=%xmm1
+pshufd $0x39,%xmm1,%xmm1
+
+# qhasm: diag2 <<<= 96
+# asm 1: pshufd $0x39,<diag2=int6464#3,<diag2=int6464#3
+# asm 2: pshufd $0x39,<diag2=%xmm2,<diag2=%xmm2
+pshufd $0x39,%xmm2,%xmm2
+
+# qhasm: diag3 <<<= 96
+# asm 1: pshufd $0x39,<diag3=int6464#4,<diag3=int6464#4
+# asm 2: pshufd $0x39,<diag3=%xmm3,<diag3=%xmm3
+pshufd $0x39,%xmm3,%xmm3
+
+# qhasm: (uint32) in0 ^= *(uint32 *) (m + 0)
+# asm 1: xorl 0(<m=int64#2),<in0=int64#4d
+# asm 2: xorl 0(<m=%rsi),<in0=%ecx
+xorl 0(%rsi),%ecx
+
+# qhasm: (uint32) in12 ^= *(uint32 *) (m + 48)
+# asm 1: xorl 48(<m=int64#2),<in12=int64#5d
+# asm 2: xorl 48(<m=%rsi),<in12=%r8d
+xorl 48(%rsi),%r8d
+
+# qhasm: (uint32) in8 ^= *(uint32 *) (m + 32)
+# asm 1: xorl 32(<m=int64#2),<in8=int64#6d
+# asm 2: xorl 32(<m=%rsi),<in8=%r9d
+xorl 32(%rsi),%r9d
+
+# qhasm: (uint32) in4 ^= *(uint32 *) (m + 16)
+# asm 1: xorl 16(<m=int64#2),<in4=int64#7d
+# asm 2: xorl 16(<m=%rsi),<in4=%eax
+xorl 16(%rsi),%eax
+
+# qhasm: *(uint32 *) (out + 0) = in0
+# asm 1: movl   <in0=int64#4d,0(<out=int64#1)
+# asm 2: movl   <in0=%ecx,0(<out=%rdi)
+movl   %ecx,0(%rdi)
+
+# qhasm: *(uint32 *) (out + 48) = in12
+# asm 1: movl   <in12=int64#5d,48(<out=int64#1)
+# asm 2: movl   <in12=%r8d,48(<out=%rdi)
+movl   %r8d,48(%rdi)
+
+# qhasm: *(uint32 *) (out + 32) = in8
+# asm 1: movl   <in8=int64#6d,32(<out=int64#1)
+# asm 2: movl   <in8=%r9d,32(<out=%rdi)
+movl   %r9d,32(%rdi)
+
+# qhasm: *(uint32 *) (out + 16) = in4
+# asm 1: movl   <in4=int64#7d,16(<out=int64#1)
+# asm 2: movl   <in4=%eax,16(<out=%rdi)
+movl   %eax,16(%rdi)
+
+# qhasm: in5 = diag0
+# asm 1: movd   <diag0=int6464#1,>in5=int64#4
+# asm 2: movd   <diag0=%xmm0,>in5=%rcx
+movd   %xmm0,%rcx
+
+# qhasm: in1 = diag1
+# asm 1: movd   <diag1=int6464#2,>in1=int64#5
+# asm 2: movd   <diag1=%xmm1,>in1=%r8
+movd   %xmm1,%r8
+
+# qhasm: in13 = diag2
+# asm 1: movd   <diag2=int6464#3,>in13=int64#6
+# asm 2: movd   <diag2=%xmm2,>in13=%r9
+movd   %xmm2,%r9
+
+# qhasm: in9 = diag3
+# asm 1: movd   <diag3=int6464#4,>in9=int64#7
+# asm 2: movd   <diag3=%xmm3,>in9=%rax
+movd   %xmm3,%rax
+
+# qhasm: diag0 <<<= 96
+# asm 1: pshufd $0x39,<diag0=int6464#1,<diag0=int6464#1
+# asm 2: pshufd $0x39,<diag0=%xmm0,<diag0=%xmm0
+pshufd $0x39,%xmm0,%xmm0
+
+# qhasm: diag1 <<<= 96
+# asm 1: pshufd $0x39,<diag1=int6464#2,<diag1=int6464#2
+# asm 2: pshufd $0x39,<diag1=%xmm1,<diag1=%xmm1
+pshufd $0x39,%xmm1,%xmm1
+
+# qhasm: diag2 <<<= 96
+# asm 1: pshufd $0x39,<diag2=int6464#3,<diag2=int6464#3
+# asm 2: pshufd $0x39,<diag2=%xmm2,<diag2=%xmm2
+pshufd $0x39,%xmm2,%xmm2
+
+# qhasm: diag3 <<<= 96
+# asm 1: pshufd $0x39,<diag3=int6464#4,<diag3=int6464#4
+# asm 2: pshufd $0x39,<diag3=%xmm3,<diag3=%xmm3
+pshufd $0x39,%xmm3,%xmm3
+
+# qhasm: (uint32) in5 ^= *(uint32 *) (m + 20)
+# asm 1: xorl 20(<m=int64#2),<in5=int64#4d
+# asm 2: xorl 20(<m=%rsi),<in5=%ecx
+xorl 20(%rsi),%ecx
+
+# qhasm: (uint32) in1 ^= *(uint32 *) (m + 4)
+# asm 1: xorl 4(<m=int64#2),<in1=int64#5d
+# asm 2: xorl 4(<m=%rsi),<in1=%r8d
+xorl 4(%rsi),%r8d
+
+# qhasm: (uint32) in13 ^= *(uint32 *) (m + 52)
+# asm 1: xorl 52(<m=int64#2),<in13=int64#6d
+# asm 2: xorl 52(<m=%rsi),<in13=%r9d
+xorl 52(%rsi),%r9d
+
+# qhasm: (uint32) in9 ^= *(uint32 *) (m + 36)
+# asm 1: xorl 36(<m=int64#2),<in9=int64#7d
+# asm 2: xorl 36(<m=%rsi),<in9=%eax
+xorl 36(%rsi),%eax
+
+# qhasm: *(uint32 *) (out + 20) = in5
+# asm 1: movl   <in5=int64#4d,20(<out=int64#1)
+# asm 2: movl   <in5=%ecx,20(<out=%rdi)
+movl   %ecx,20(%rdi)
+
+# qhasm: *(uint32 *) (out + 4) = in1
+# asm 1: movl   <in1=int64#5d,4(<out=int64#1)
+# asm 2: movl   <in1=%r8d,4(<out=%rdi)
+movl   %r8d,4(%rdi)
+
+# qhasm: *(uint32 *) (out + 52) = in13
+# asm 1: movl   <in13=int64#6d,52(<out=int64#1)
+# asm 2: movl   <in13=%r9d,52(<out=%rdi)
+movl   %r9d,52(%rdi)
+
+# qhasm: *(uint32 *) (out + 36) = in9
+# asm 1: movl   <in9=int64#7d,36(<out=int64#1)
+# asm 2: movl   <in9=%eax,36(<out=%rdi)
+movl   %eax,36(%rdi)
+
+# qhasm: in10 = diag0
+# asm 1: movd   <diag0=int6464#1,>in10=int64#4
+# asm 2: movd   <diag0=%xmm0,>in10=%rcx
+movd   %xmm0,%rcx
+
+# qhasm: in6 = diag1
+# asm 1: movd   <diag1=int6464#2,>in6=int64#5
+# asm 2: movd   <diag1=%xmm1,>in6=%r8
+movd   %xmm1,%r8
+
+# qhasm: in2 = diag2
+# asm 1: movd   <diag2=int6464#3,>in2=int64#6
+# asm 2: movd   <diag2=%xmm2,>in2=%r9
+movd   %xmm2,%r9
+
+# qhasm: in14 = diag3
+# asm 1: movd   <diag3=int6464#4,>in14=int64#7
+# asm 2: movd   <diag3=%xmm3,>in14=%rax
+movd   %xmm3,%rax
+
+# qhasm: diag0 <<<= 96
+# asm 1: pshufd $0x39,<diag0=int6464#1,<diag0=int6464#1
+# asm 2: pshufd $0x39,<diag0=%xmm0,<diag0=%xmm0
+pshufd $0x39,%xmm0,%xmm0
+
+# qhasm: diag1 <<<= 96
+# asm 1: pshufd $0x39,<diag1=int6464#2,<diag1=int6464#2
+# asm 2: pshufd $0x39,<diag1=%xmm1,<diag1=%xmm1
+pshufd $0x39,%xmm1,%xmm1
+
+# qhasm: diag2 <<<= 96
+# asm 1: pshufd $0x39,<diag2=int6464#3,<diag2=int6464#3
+# asm 2: pshufd $0x39,<diag2=%xmm2,<diag2=%xmm2
+pshufd $0x39,%xmm2,%xmm2
+
+# qhasm: diag3 <<<= 96
+# asm 1: pshufd $0x39,<diag3=int6464#4,<diag3=int6464#4
+# asm 2: pshufd $0x39,<diag3=%xmm3,<diag3=%xmm3
+pshufd $0x39,%xmm3,%xmm3
+
+# qhasm: (uint32) in10 ^= *(uint32 *) (m + 40)
+# asm 1: xorl 40(<m=int64#2),<in10=int64#4d
+# asm 2: xorl 40(<m=%rsi),<in10=%ecx
+xorl 40(%rsi),%ecx
+
+# qhasm: (uint32) in6 ^= *(uint32 *) (m + 24)
+# asm 1: xorl 24(<m=int64#2),<in6=int64#5d
+# asm 2: xorl 24(<m=%rsi),<in6=%r8d
+xorl 24(%rsi),%r8d
+
+# qhasm: (uint32) in2 ^= *(uint32 *) (m + 8)
+# asm 1: xorl 8(<m=int64#2),<in2=int64#6d
+# asm 2: xorl 8(<m=%rsi),<in2=%r9d
+xorl 8(%rsi),%r9d
+
+# qhasm: (uint32) in14 ^= *(uint32 *) (m + 56)
+# asm 1: xorl 56(<m=int64#2),<in14=int64#7d
+# asm 2: xorl 56(<m=%rsi),<in14=%eax
+xorl 56(%rsi),%eax
+
+# qhasm: *(uint32 *) (out + 40) = in10
+# asm 1: movl   <in10=int64#4d,40(<out=int64#1)
+# asm 2: movl   <in10=%ecx,40(<out=%rdi)
+movl   %ecx,40(%rdi)
+
+# qhasm: *(uint32 *) (out + 24) = in6
+# asm 1: movl   <in6=int64#5d,24(<out=int64#1)
+# asm 2: movl   <in6=%r8d,24(<out=%rdi)
+movl   %r8d,24(%rdi)
+
+# qhasm: *(uint32 *) (out + 8) = in2
+# asm 1: movl   <in2=int64#6d,8(<out=int64#1)
+# asm 2: movl   <in2=%r9d,8(<out=%rdi)
+movl   %r9d,8(%rdi)
+
+# qhasm: *(uint32 *) (out + 56) = in14
+# asm 1: movl   <in14=int64#7d,56(<out=int64#1)
+# asm 2: movl   <in14=%eax,56(<out=%rdi)
+movl   %eax,56(%rdi)
+
+# qhasm: in15 = diag0
+# asm 1: movd   <diag0=int6464#1,>in15=int64#4
+# asm 2: movd   <diag0=%xmm0,>in15=%rcx
+movd   %xmm0,%rcx
+
+# qhasm: in11 = diag1
+# asm 1: movd   <diag1=int6464#2,>in11=int64#5
+# asm 2: movd   <diag1=%xmm1,>in11=%r8
+movd   %xmm1,%r8
+
+# qhasm: in7 = diag2
+# asm 1: movd   <diag2=int6464#3,>in7=int64#6
+# asm 2: movd   <diag2=%xmm2,>in7=%r9
+movd   %xmm2,%r9
+
+# qhasm: in3 = diag3
+# asm 1: movd   <diag3=int6464#4,>in3=int64#7
+# asm 2: movd   <diag3=%xmm3,>in3=%rax
+movd   %xmm3,%rax
+
+# qhasm: (uint32) in15 ^= *(uint32 *) (m + 60)
+# asm 1: xorl 60(<m=int64#2),<in15=int64#4d
+# asm 2: xorl 60(<m=%rsi),<in15=%ecx
+xorl 60(%rsi),%ecx
+
+# qhasm: (uint32) in11 ^= *(uint32 *) (m + 44)
+# asm 1: xorl 44(<m=int64#2),<in11=int64#5d
+# asm 2: xorl 44(<m=%rsi),<in11=%r8d
+xorl 44(%rsi),%r8d
+
+# qhasm: (uint32) in7 ^= *(uint32 *) (m + 28)
+# asm 1: xorl 28(<m=int64#2),<in7=int64#6d
+# asm 2: xorl 28(<m=%rsi),<in7=%r9d
+xorl 28(%rsi),%r9d
+
+# qhasm: (uint32) in3 ^= *(uint32 *) (m + 12)
+# asm 1: xorl 12(<m=int64#2),<in3=int64#7d
+# asm 2: xorl 12(<m=%rsi),<in3=%eax
+xorl 12(%rsi),%eax
+
+# qhasm: *(uint32 *) (out + 60) = in15
+# asm 1: movl   <in15=int64#4d,60(<out=int64#1)
+# asm 2: movl   <in15=%ecx,60(<out=%rdi)
+movl   %ecx,60(%rdi)
+
+# qhasm: *(uint32 *) (out + 44) = in11
+# asm 1: movl   <in11=int64#5d,44(<out=int64#1)
+# asm 2: movl   <in11=%r8d,44(<out=%rdi)
+movl   %r8d,44(%rdi)
+
+# qhasm: *(uint32 *) (out + 28) = in7
+# asm 1: movl   <in7=int64#6d,28(<out=int64#1)
+# asm 2: movl   <in7=%r9d,28(<out=%rdi)
+movl   %r9d,28(%rdi)
+
+# qhasm: *(uint32 *) (out + 12) = in3
+# asm 1: movl   <in3=int64#7d,12(<out=int64#1)
+# asm 2: movl   <in3=%eax,12(<out=%rdi)
+movl   %eax,12(%rdi)
+
+# qhasm:   bytes = bytes_backup
+# asm 1: movq <bytes_backup=stack64#8,>bytes=int64#6
+# asm 2: movq <bytes_backup=408(%rsp),>bytes=%r9
+movq 408(%rsp),%r9
+
+# qhasm:   in8 = ((uint32 *)&x2)[0]
+# asm 1: movl <x2=stack128#2,>in8=int64#4d
+# asm 2: movl <x2=16(%rsp),>in8=%ecx
+movl 16(%rsp),%ecx
+
+# qhasm:   in9 = ((uint32 *)&x3)[1]
+# asm 1: movl 4+<x3=stack128#3,>in9=int64#5d
+# asm 2: movl 4+<x3=32(%rsp),>in9=%r8d
+movl 4+32(%rsp),%r8d
+
+# qhasm:   in8 += 1
+# asm 1: add  $1,<in8=int64#4
+# asm 2: add  $1,<in8=%rcx
+add  $1,%rcx
+
+# qhasm:   in9 <<= 32
+# asm 1: shl  $32,<in9=int64#5
+# asm 2: shl  $32,<in9=%r8
+shl  $32,%r8
+
+# qhasm:   in8 += in9
+# asm 1: add  <in9=int64#5,<in8=int64#4
+# asm 2: add  <in9=%r8,<in8=%rcx
+add  %r8,%rcx
+
+# qhasm:   in9 = in8
+# asm 1: mov  <in8=int64#4,>in9=int64#5
+# asm 2: mov  <in8=%rcx,>in9=%r8
+mov  %rcx,%r8
+
+# qhasm:   (uint64) in9 >>= 32
+# asm 1: shr  $32,<in9=int64#5
+# asm 2: shr  $32,<in9=%r8
+shr  $32,%r8
+
+# qhasm:   ((uint32 *)&x2)[0] = in8
+# asm 1: movl <in8=int64#4d,>x2=stack128#2
+# asm 2: movl <in8=%ecx,>x2=16(%rsp)
+movl %ecx,16(%rsp)
+
+# qhasm:   ((uint32 *)&x3)[1] = in9
+# asm 1: movl <in9=int64#5d,4+<x3=stack128#3
+# asm 2: movl <in9=%r8d,4+<x3=32(%rsp)
+movl %r8d,4+32(%rsp)
+
+# qhasm:                          unsigned>? unsigned<? bytes - 64
+# asm 1: cmp  $64,<bytes=int64#6
+# asm 2: cmp  $64,<bytes=%r9
+cmp  $64,%r9
+# comment:fp stack unchanged by jump
+
+# qhasm:   goto bytesatleast65 if unsigned>
+ja ._bytesatleast65
+# comment:fp stack unchanged by jump
+
+# qhasm:     goto bytesatleast64 if !unsigned<
+jae ._bytesatleast64
+
+# qhasm:       m = out
+# asm 1: mov  <out=int64#1,>m=int64#2
+# asm 2: mov  <out=%rdi,>m=%rsi
+mov  %rdi,%rsi
+
+# qhasm:       out = ctarget
+# asm 1: mov  <ctarget=int64#3,>out=int64#1
+# asm 2: mov  <ctarget=%rdx,>out=%rdi
+mov  %rdx,%rdi
+
+# qhasm:       i = bytes
+# asm 1: mov  <bytes=int64#6,>i=int64#4
+# asm 2: mov  <bytes=%r9,>i=%rcx
+mov  %r9,%rcx
+
+# qhasm:       while (i) { *out++ = *m++; --i }
+rep movsb
+# comment:fp stack unchanged by fallthrough
+
+# qhasm:     bytesatleast64:
+._bytesatleast64:
+# comment:fp stack unchanged by fallthrough
+
+# qhasm:     done:
+._done:
+
+# qhasm:     r11_caller = r11_stack
+# asm 1: movq <r11_stack=stack64#1,>r11_caller=int64#9
+# asm 2: movq <r11_stack=352(%rsp),>r11_caller=%r11
+movq 352(%rsp),%r11
+
+# qhasm:     r12_caller = r12_stack
+# asm 1: movq <r12_stack=stack64#2,>r12_caller=int64#10
+# asm 2: movq <r12_stack=360(%rsp),>r12_caller=%r12
+movq 360(%rsp),%r12
+
+# qhasm:     r13_caller = r13_stack
+# asm 1: movq <r13_stack=stack64#3,>r13_caller=int64#11
+# asm 2: movq <r13_stack=368(%rsp),>r13_caller=%r13
+movq 368(%rsp),%r13
+
+# qhasm:     r14_caller = r14_stack
+# asm 1: movq <r14_stack=stack64#4,>r14_caller=int64#12
+# asm 2: movq <r14_stack=376(%rsp),>r14_caller=%r14
+movq 376(%rsp),%r14
+
+# qhasm:     r15_caller = r15_stack
+# asm 1: movq <r15_stack=stack64#5,>r15_caller=int64#13
+# asm 2: movq <r15_stack=384(%rsp),>r15_caller=%r15
+movq 384(%rsp),%r15
+
+# qhasm:     rbx_caller = rbx_stack
+# asm 1: movq <rbx_stack=stack64#6,>rbx_caller=int64#14
+# asm 2: movq <rbx_stack=392(%rsp),>rbx_caller=%rbx
+movq 392(%rsp),%rbx
+
+# qhasm:     rbp_caller = rbp_stack
+# asm 1: movq <rbp_stack=stack64#7,>rbp_caller=int64#15
+# asm 2: movq <rbp_stack=400(%rsp),>rbp_caller=%rbp
+movq 400(%rsp),%rbp
+
+# qhasm:     leave
+add %r11,%rsp
+xor %rax,%rax
+xor %rdx,%rdx
+ret
+
+# qhasm:   bytesatleast65:
+._bytesatleast65:
+
+# qhasm:   bytes -= 64
+# asm 1: sub  $64,<bytes=int64#6
+# asm 2: sub  $64,<bytes=%r9
+sub  $64,%r9
+
+# qhasm:   out += 64
+# asm 1: add  $64,<out=int64#1
+# asm 2: add  $64,<out=%rdi
+add  $64,%rdi
+
+# qhasm:   m += 64
+# asm 1: add  $64,<m=int64#2
+# asm 2: add  $64,<m=%rsi
+add  $64,%rsi
+# comment:fp stack unchanged by jump
+
+# qhasm: goto bytesbetween1and255
+jmp ._bytesbetween1and255
diff --git a/src/crypto/cipher/salsa20/xmm/salsa20_x86_xmm5.s b/src/crypto/cipher/salsa20/xmm/salsa20_x86_xmm5.s
new file mode 100644
index 0000000..9e32ea4
--- /dev/null
+++ b/src/crypto/cipher/salsa20/xmm/salsa20_x86_xmm5.s
@@ -0,0 +1,5078 @@
+
+# qhasm: int32 a
+
+# qhasm: stack32 arg1
+
+# qhasm: stack32 arg2
+
+# qhasm: stack32 arg3
+
+# qhasm: stack32 arg4
+
+# qhasm: stack32 arg5
+
+# qhasm: stack32 arg6
+
+# qhasm: input arg1
+
+# qhasm: input arg2
+
+# qhasm: input arg3
+
+# qhasm: input arg4
+
+# qhasm: input arg5
+
+# qhasm: input arg6
+
+# qhasm: int32 eax
+
+# qhasm: int32 ebx
+
+# qhasm: int32 esi
+
+# qhasm: int32 edi
+
+# qhasm: int32 ebp
+
+# qhasm: caller eax
+
+# qhasm: caller ebx
+
+# qhasm: caller esi
+
+# qhasm: caller edi
+
+# qhasm: caller ebp
+
+# qhasm: int32 k
+
+# qhasm: int32 kbits
+
+# qhasm: int32 iv
+
+# qhasm: int32 i
+
+# qhasm: stack128 x0
+
+# qhasm: stack128 x1
+
+# qhasm: stack128 x2
+
+# qhasm: stack128 x3
+
+# qhasm: int32 m
+
+# qhasm: stack32 out_stack
+
+# qhasm: int32 out
+
+# qhasm: stack32 bytes_stack
+
+# qhasm: int32 bytes
+
+# qhasm: stack32 eax_stack
+
+# qhasm: stack32 ebx_stack
+
+# qhasm: stack32 esi_stack
+
+# qhasm: stack32 edi_stack
+
+# qhasm: stack32 ebp_stack
+
+# qhasm: int6464 diag0
+
+# qhasm: int6464 diag1
+
+# qhasm: int6464 diag2
+
+# qhasm: int6464 diag3
+
+# qhasm: int6464 a0
+
+# qhasm: int6464 a1
+
+# qhasm: int6464 a2
+
+# qhasm: int6464 a3
+
+# qhasm: int6464 a4
+
+# qhasm: int6464 a5
+
+# qhasm: int6464 a6
+
+# qhasm: int6464 a7
+
+# qhasm: int6464 b0
+
+# qhasm: int6464 b1
+
+# qhasm: int6464 b2
+
+# qhasm: int6464 b3
+
+# qhasm: int6464 b4
+
+# qhasm: int6464 b5
+
+# qhasm: int6464 b6
+
+# qhasm: int6464 b7
+
+# qhasm: int6464 z0
+
+# qhasm: int6464 z1
+
+# qhasm: int6464 z2
+
+# qhasm: int6464 z3
+
+# qhasm: int6464 z4
+
+# qhasm: int6464 z5
+
+# qhasm: int6464 z6
+
+# qhasm: int6464 z7
+
+# qhasm: int6464 z8
+
+# qhasm: int6464 z9
+
+# qhasm: int6464 z10
+
+# qhasm: int6464 z11
+
+# qhasm: int6464 z12
+
+# qhasm: int6464 z13
+
+# qhasm: int6464 z14
+
+# qhasm: int6464 z15
+
+# qhasm: stack128 z0_stack
+
+# qhasm: stack128 z1_stack
+
+# qhasm: stack128 z2_stack
+
+# qhasm: stack128 z3_stack
+
+# qhasm: stack128 z4_stack
+
+# qhasm: stack128 z5_stack
+
+# qhasm: stack128 z6_stack
+
+# qhasm: stack128 z7_stack
+
+# qhasm: stack128 z8_stack
+
+# qhasm: stack128 z9_stack
+
+# qhasm: stack128 z10_stack
+
+# qhasm: stack128 z11_stack
+
+# qhasm: stack128 z12_stack
+
+# qhasm: stack128 z13_stack
+
+# qhasm: stack128 z14_stack
+
+# qhasm: stack128 z15_stack
+
+# qhasm: stack128 orig0
+
+# qhasm: stack128 orig1
+
+# qhasm: stack128 orig2
+
+# qhasm: stack128 orig3
+
+# qhasm: stack128 orig4
+
+# qhasm: stack128 orig5
+
+# qhasm: stack128 orig6
+
+# qhasm: stack128 orig7
+
+# qhasm: stack128 orig8
+
+# qhasm: stack128 orig9
+
+# qhasm: stack128 orig10
+
+# qhasm: stack128 orig11
+
+# qhasm: stack128 orig12
+
+# qhasm: stack128 orig13
+
+# qhasm: stack128 orig14
+
+# qhasm: stack128 orig15
+
+# qhasm: int6464 p
+
+# qhasm: int6464 q
+
+# qhasm: int6464 r
+
+# qhasm: int6464 s
+
+# qhasm: int6464 t
+
+# qhasm: int6464 u
+
+# qhasm: int6464 v
+
+# qhasm: int6464 w
+
+# qhasm: int6464 mp
+
+# qhasm: int6464 mq
+
+# qhasm: int6464 mr
+
+# qhasm: int6464 ms
+
+# qhasm: int6464 mt
+
+# qhasm: int6464 mu
+
+# qhasm: int6464 mv
+
+# qhasm: int6464 mw
+
+# qhasm: int32 in0
+
+# qhasm: int32 in1
+
+# qhasm: int32 in2
+
+# qhasm: int32 in3
+
+# qhasm: int32 in4
+
+# qhasm: int32 in5
+
+# qhasm: int32 in6
+
+# qhasm: int32 in7
+
+# qhasm: int32 in8
+
+# qhasm: int32 in9
+
+# qhasm: int32 in10
+
+# qhasm: int32 in11
+
+# qhasm: int32 in12
+
+# qhasm: int32 in13
+
+# qhasm: int32 in14
+
+# qhasm: int32 in15
+
+# qhasm: stack512 tmp
+
+# qhasm: stack32 ctarget
+
+# qhasm: enter crypto_stream_salsa20_x86_xmm5
+.text
+.p2align 5
+.globl _crypto_stream_salsa20_x86_xmm5
+.globl crypto_stream_salsa20_x86_xmm5
+_crypto_stream_salsa20_x86_xmm5:
+crypto_stream_salsa20_x86_xmm5:
+mov %esp,%eax
+and $31,%eax
+add $704,%eax
+sub %eax,%esp
+
+# qhasm: eax_stack = eax
+# asm 1: movl <eax=int32#1,>eax_stack=stack32#1
+# asm 2: movl <eax=%eax,>eax_stack=0(%esp)
+movl %eax,0(%esp)
+
+# qhasm: ebx_stack = ebx
+# asm 1: movl <ebx=int32#4,>ebx_stack=stack32#2
+# asm 2: movl <ebx=%ebx,>ebx_stack=4(%esp)
+movl %ebx,4(%esp)
+
+# qhasm: esi_stack = esi
+# asm 1: movl <esi=int32#5,>esi_stack=stack32#3
+# asm 2: movl <esi=%esi,>esi_stack=8(%esp)
+movl %esi,8(%esp)
+
+# qhasm: edi_stack = edi
+# asm 1: movl <edi=int32#6,>edi_stack=stack32#4
+# asm 2: movl <edi=%edi,>edi_stack=12(%esp)
+movl %edi,12(%esp)
+
+# qhasm: ebp_stack = ebp
+# asm 1: movl <ebp=int32#7,>ebp_stack=stack32#5
+# asm 2: movl <ebp=%ebp,>ebp_stack=16(%esp)
+movl %ebp,16(%esp)
+
+# qhasm: bytes = arg2
+# asm 1: movl <arg2=stack32#-2,>bytes=int32#3
+# asm 2: movl <arg2=8(%esp,%eax),>bytes=%edx
+movl 8(%esp,%eax),%edx
+
+# qhasm: out = arg1
+# asm 1: movl <arg1=stack32#-1,>out=int32#6
+# asm 2: movl <arg1=4(%esp,%eax),>out=%edi
+movl 4(%esp,%eax),%edi
+
+# qhasm: m = out
+# asm 1: mov  <out=int32#6,>m=int32#5
+# asm 2: mov  <out=%edi,>m=%esi
+mov  %edi,%esi
+
+# qhasm: iv = arg4
+# asm 1: movl <arg4=stack32#-4,>iv=int32#4
+# asm 2: movl <arg4=16(%esp,%eax),>iv=%ebx
+movl 16(%esp,%eax),%ebx
+
+# qhasm: k = arg5
+# asm 1: movl <arg5=stack32#-5,>k=int32#7
+# asm 2: movl <arg5=20(%esp,%eax),>k=%ebp
+movl 20(%esp,%eax),%ebp
+
+# qhasm:               unsigned>? bytes - 0
+# asm 1: cmp  $0,<bytes=int32#3
+# asm 2: cmp  $0,<bytes=%edx
+cmp  $0,%edx
+# comment:fp stack unchanged by jump
+
+# qhasm: goto done if !unsigned>
+jbe ._done
+
+# qhasm: a = 0
+# asm 1: mov  $0,>a=int32#1
+# asm 2: mov  $0,>a=%eax
+mov  $0,%eax
+
+# qhasm: i = bytes
+# asm 1: mov  <bytes=int32#3,>i=int32#2
+# asm 2: mov  <bytes=%edx,>i=%ecx
+mov  %edx,%ecx
+
+# qhasm: while (i) { *out++ = a; --i }
+rep stosb
+
+# qhasm: out -= bytes
+# asm 1: subl <bytes=int32#3,<out=int32#6
+# asm 2: subl <bytes=%edx,<out=%edi
+subl %edx,%edi
+# comment:fp stack unchanged by jump
+
+# qhasm: goto start
+jmp ._start
+
+# qhasm: enter crypto_stream_salsa20_x86_xmm5_xor
+.text
+.p2align 5
+.globl _crypto_stream_salsa20_x86_xmm5_xor
+.globl crypto_stream_salsa20_x86_xmm5_xor
+_crypto_stream_salsa20_x86_xmm5_xor:
+crypto_stream_salsa20_x86_xmm5_xor:
+mov %esp,%eax
+and $31,%eax
+add $704,%eax
+sub %eax,%esp
+
+# qhasm: eax_stack = eax
+# asm 1: movl <eax=int32#1,>eax_stack=stack32#1
+# asm 2: movl <eax=%eax,>eax_stack=0(%esp)
+movl %eax,0(%esp)
+
+# qhasm: ebx_stack = ebx
+# asm 1: movl <ebx=int32#4,>ebx_stack=stack32#2
+# asm 2: movl <ebx=%ebx,>ebx_stack=4(%esp)
+movl %ebx,4(%esp)
+
+# qhasm: esi_stack = esi
+# asm 1: movl <esi=int32#5,>esi_stack=stack32#3
+# asm 2: movl <esi=%esi,>esi_stack=8(%esp)
+movl %esi,8(%esp)
+
+# qhasm: edi_stack = edi
+# asm 1: movl <edi=int32#6,>edi_stack=stack32#4
+# asm 2: movl <edi=%edi,>edi_stack=12(%esp)
+movl %edi,12(%esp)
+
+# qhasm: ebp_stack = ebp
+# asm 1: movl <ebp=int32#7,>ebp_stack=stack32#5
+# asm 2: movl <ebp=%ebp,>ebp_stack=16(%esp)
+movl %ebp,16(%esp)
+
+# qhasm: out = arg1
+# asm 1: movl <arg1=stack32#-1,>out=int32#6
+# asm 2: movl <arg1=4(%esp,%eax),>out=%edi
+movl 4(%esp,%eax),%edi
+
+# qhasm: m = arg2
+# asm 1: movl <arg2=stack32#-2,>m=int32#5
+# asm 2: movl <arg2=8(%esp,%eax),>m=%esi
+movl 8(%esp,%eax),%esi
+
+# qhasm: bytes = arg3
+# asm 1: movl <arg3=stack32#-3,>bytes=int32#3
+# asm 2: movl <arg3=12(%esp,%eax),>bytes=%edx
+movl 12(%esp,%eax),%edx
+
+# qhasm: iv = arg5
+# asm 1: movl <arg5=stack32#-5,>iv=int32#4
+# asm 2: movl <arg5=20(%esp,%eax),>iv=%ebx
+movl 20(%esp,%eax),%ebx
+
+# qhasm: k = arg6
+# asm 1: movl <arg6=stack32#-6,>k=int32#7
+# asm 2: movl <arg6=24(%esp,%eax),>k=%ebp
+movl 24(%esp,%eax),%ebp
+
+# qhasm:               unsigned>? bytes - 0
+# asm 1: cmp  $0,<bytes=int32#3
+# asm 2: cmp  $0,<bytes=%edx
+cmp  $0,%edx
+# comment:fp stack unchanged by jump
+
+# qhasm: goto done if !unsigned>
+jbe ._done
+# comment:fp stack unchanged by fallthrough
+
+# qhasm: start:
+._start:
+
+# qhasm:   out_stack = out
+# asm 1: movl <out=int32#6,>out_stack=stack32#6
+# asm 2: movl <out=%edi,>out_stack=20(%esp)
+movl %edi,20(%esp)
+
+# qhasm:   bytes_stack = bytes
+# asm 1: movl <bytes=int32#3,>bytes_stack=stack32#7
+# asm 2: movl <bytes=%edx,>bytes_stack=24(%esp)
+movl %edx,24(%esp)
+
+# qhasm:   in4 = *(uint32 *) (k + 12)
+# asm 1: movl 12(<k=int32#7),>in4=int32#1
+# asm 2: movl 12(<k=%ebp),>in4=%eax
+movl 12(%ebp),%eax
+
+# qhasm:   in12 = *(uint32 *) (k + 20)
+# asm 1: movl 20(<k=int32#7),>in12=int32#2
+# asm 2: movl 20(<k=%ebp),>in12=%ecx
+movl 20(%ebp),%ecx
+
+# qhasm:   ((uint32 *)&x3)[0] = in4
+# asm 1: movl <in4=int32#1,>x3=stack128#1
+# asm 2: movl <in4=%eax,>x3=32(%esp)
+movl %eax,32(%esp)
+
+# qhasm:   ((uint32 *)&x1)[0] = in12
+# asm 1: movl <in12=int32#2,>x1=stack128#2
+# asm 2: movl <in12=%ecx,>x1=48(%esp)
+movl %ecx,48(%esp)
+
+# qhasm:   in0 = 1634760805
+# asm 1: mov  $1634760805,>in0=int32#1
+# asm 2: mov  $1634760805,>in0=%eax
+mov  $1634760805,%eax
+
+# qhasm:   in8 = 0
+# asm 1: mov  $0,>in8=int32#2
+# asm 2: mov  $0,>in8=%ecx
+mov  $0,%ecx
+
+# qhasm:   ((uint32 *)&x0)[0] = in0
+# asm 1: movl <in0=int32#1,>x0=stack128#3
+# asm 2: movl <in0=%eax,>x0=64(%esp)
+movl %eax,64(%esp)
+
+# qhasm:   ((uint32 *)&x2)[0] = in8
+# asm 1: movl <in8=int32#2,>x2=stack128#4
+# asm 2: movl <in8=%ecx,>x2=80(%esp)
+movl %ecx,80(%esp)
+
+# qhasm:   in6 = *(uint32 *) (iv + 0)
+# asm 1: movl 0(<iv=int32#4),>in6=int32#1
+# asm 2: movl 0(<iv=%ebx),>in6=%eax
+movl 0(%ebx),%eax
+
+# qhasm:   in7 = *(uint32 *) (iv + 4)
+# asm 1: movl 4(<iv=int32#4),>in7=int32#2
+# asm 2: movl 4(<iv=%ebx),>in7=%ecx
+movl 4(%ebx),%ecx
+
+# qhasm:   ((uint32 *)&x1)[2] = in6
+# asm 1: movl <in6=int32#1,8+<x1=stack128#2
+# asm 2: movl <in6=%eax,8+<x1=48(%esp)
+movl %eax,8+48(%esp)
+
+# qhasm:   ((uint32 *)&x2)[3] = in7
+# asm 1: movl <in7=int32#2,12+<x2=stack128#4
+# asm 2: movl <in7=%ecx,12+<x2=80(%esp)
+movl %ecx,12+80(%esp)
+
+# qhasm:   in9 = 0
+# asm 1: mov  $0,>in9=int32#1
+# asm 2: mov  $0,>in9=%eax
+mov  $0,%eax
+
+# qhasm:   in10 = 2036477234
+# asm 1: mov  $2036477234,>in10=int32#2
+# asm 2: mov  $2036477234,>in10=%ecx
+mov  $2036477234,%ecx
+
+# qhasm:   ((uint32 *)&x3)[1] = in9
+# asm 1: movl <in9=int32#1,4+<x3=stack128#1
+# asm 2: movl <in9=%eax,4+<x3=32(%esp)
+movl %eax,4+32(%esp)
+
+# qhasm:   ((uint32 *)&x0)[2] = in10
+# asm 1: movl <in10=int32#2,8+<x0=stack128#3
+# asm 2: movl <in10=%ecx,8+<x0=64(%esp)
+movl %ecx,8+64(%esp)
+
+# qhasm:   in1 = *(uint32 *) (k + 0)
+# asm 1: movl 0(<k=int32#7),>in1=int32#1
+# asm 2: movl 0(<k=%ebp),>in1=%eax
+movl 0(%ebp),%eax
+
+# qhasm:   in2 = *(uint32 *) (k + 4)
+# asm 1: movl 4(<k=int32#7),>in2=int32#2
+# asm 2: movl 4(<k=%ebp),>in2=%ecx
+movl 4(%ebp),%ecx
+
+# qhasm:   in3 = *(uint32 *) (k + 8)
+# asm 1: movl 8(<k=int32#7),>in3=int32#3
+# asm 2: movl 8(<k=%ebp),>in3=%edx
+movl 8(%ebp),%edx
+
+# qhasm:   in5 = 857760878
+# asm 1: mov  $857760878,>in5=int32#4
+# asm 2: mov  $857760878,>in5=%ebx
+mov  $857760878,%ebx
+
+# qhasm:   ((uint32 *)&x1)[1] = in1
+# asm 1: movl <in1=int32#1,4+<x1=stack128#2
+# asm 2: movl <in1=%eax,4+<x1=48(%esp)
+movl %eax,4+48(%esp)
+
+# qhasm:   ((uint32 *)&x2)[2] = in2
+# asm 1: movl <in2=int32#2,8+<x2=stack128#4
+# asm 2: movl <in2=%ecx,8+<x2=80(%esp)
+movl %ecx,8+80(%esp)
+
+# qhasm:   ((uint32 *)&x3)[3] = in3
+# asm 1: movl <in3=int32#3,12+<x3=stack128#1
+# asm 2: movl <in3=%edx,12+<x3=32(%esp)
+movl %edx,12+32(%esp)
+
+# qhasm:   ((uint32 *)&x0)[1] = in5
+# asm 1: movl <in5=int32#4,4+<x0=stack128#3
+# asm 2: movl <in5=%ebx,4+<x0=64(%esp)
+movl %ebx,4+64(%esp)
+
+# qhasm:   in11 = *(uint32 *) (k + 16)
+# asm 1: movl 16(<k=int32#7),>in11=int32#1
+# asm 2: movl 16(<k=%ebp),>in11=%eax
+movl 16(%ebp),%eax
+
+# qhasm:   in13 = *(uint32 *) (k + 24)
+# asm 1: movl 24(<k=int32#7),>in13=int32#2
+# asm 2: movl 24(<k=%ebp),>in13=%ecx
+movl 24(%ebp),%ecx
+
+# qhasm:   in14 = *(uint32 *) (k + 28)
+# asm 1: movl 28(<k=int32#7),>in14=int32#3
+# asm 2: movl 28(<k=%ebp),>in14=%edx
+movl 28(%ebp),%edx
+
+# qhasm:   in15 = 1797285236
+# asm 1: mov  $1797285236,>in15=int32#4
+# asm 2: mov  $1797285236,>in15=%ebx
+mov  $1797285236,%ebx
+
+# qhasm:   ((uint32 *)&x1)[3] = in11
+# asm 1: movl <in11=int32#1,12+<x1=stack128#2
+# asm 2: movl <in11=%eax,12+<x1=48(%esp)
+movl %eax,12+48(%esp)
+
+# qhasm:   ((uint32 *)&x2)[1] = in13
+# asm 1: movl <in13=int32#2,4+<x2=stack128#4
+# asm 2: movl <in13=%ecx,4+<x2=80(%esp)
+movl %ecx,4+80(%esp)
+
+# qhasm:   ((uint32 *)&x3)[2] = in14
+# asm 1: movl <in14=int32#3,8+<x3=stack128#1
+# asm 2: movl <in14=%edx,8+<x3=32(%esp)
+movl %edx,8+32(%esp)
+
+# qhasm:   ((uint32 *)&x0)[3] = in15
+# asm 1: movl <in15=int32#4,12+<x0=stack128#3
+# asm 2: movl <in15=%ebx,12+<x0=64(%esp)
+movl %ebx,12+64(%esp)
+
+# qhasm:   bytes = bytes_stack
+# asm 1: movl <bytes_stack=stack32#7,>bytes=int32#1
+# asm 2: movl <bytes_stack=24(%esp),>bytes=%eax
+movl 24(%esp),%eax
+
+# qhasm:                               unsigned<? bytes - 256
+# asm 1: cmp  $256,<bytes=int32#1
+# asm 2: cmp  $256,<bytes=%eax
+cmp  $256,%eax
+# comment:fp stack unchanged by jump
+
+# qhasm:   goto bytesbetween1and255 if unsigned<
+jb ._bytesbetween1and255
+
+# qhasm:   z0 = x0
+# asm 1: movdqa <x0=stack128#3,>z0=int6464#1
+# asm 2: movdqa <x0=64(%esp),>z0=%xmm0
+movdqa 64(%esp),%xmm0
+
+# qhasm:   z5 = z0[1,1,1,1]
+# asm 1: pshufd $0x55,<z0=int6464#1,>z5=int6464#2
+# asm 2: pshufd $0x55,<z0=%xmm0,>z5=%xmm1
+pshufd $0x55,%xmm0,%xmm1
+
+# qhasm:   z10 = z0[2,2,2,2]
+# asm 1: pshufd $0xaa,<z0=int6464#1,>z10=int6464#3
+# asm 2: pshufd $0xaa,<z0=%xmm0,>z10=%xmm2
+pshufd $0xaa,%xmm0,%xmm2
+
+# qhasm:   z15 = z0[3,3,3,3]
+# asm 1: pshufd $0xff,<z0=int6464#1,>z15=int6464#4
+# asm 2: pshufd $0xff,<z0=%xmm0,>z15=%xmm3
+pshufd $0xff,%xmm0,%xmm3
+
+# qhasm:   z0 = z0[0,0,0,0]
+# asm 1: pshufd $0x00,<z0=int6464#1,>z0=int6464#1
+# asm 2: pshufd $0x00,<z0=%xmm0,>z0=%xmm0
+pshufd $0x00,%xmm0,%xmm0
+
+# qhasm:   orig5 = z5
+# asm 1: movdqa <z5=int6464#2,>orig5=stack128#5
+# asm 2: movdqa <z5=%xmm1,>orig5=96(%esp)
+movdqa %xmm1,96(%esp)
+
+# qhasm:   orig10 = z10
+# asm 1: movdqa <z10=int6464#3,>orig10=stack128#6
+# asm 2: movdqa <z10=%xmm2,>orig10=112(%esp)
+movdqa %xmm2,112(%esp)
+
+# qhasm:   orig15 = z15
+# asm 1: movdqa <z15=int6464#4,>orig15=stack128#7
+# asm 2: movdqa <z15=%xmm3,>orig15=128(%esp)
+movdqa %xmm3,128(%esp)
+
+# qhasm:   orig0 = z0
+# asm 1: movdqa <z0=int6464#1,>orig0=stack128#8
+# asm 2: movdqa <z0=%xmm0,>orig0=144(%esp)
+movdqa %xmm0,144(%esp)
+
+# qhasm:   z1 = x1
+# asm 1: movdqa <x1=stack128#2,>z1=int6464#1
+# asm 2: movdqa <x1=48(%esp),>z1=%xmm0
+movdqa 48(%esp),%xmm0
+
+# qhasm:   z6 = z1[2,2,2,2]
+# asm 1: pshufd $0xaa,<z1=int6464#1,>z6=int6464#2
+# asm 2: pshufd $0xaa,<z1=%xmm0,>z6=%xmm1
+pshufd $0xaa,%xmm0,%xmm1
+
+# qhasm:   z11 = z1[3,3,3,3]
+# asm 1: pshufd $0xff,<z1=int6464#1,>z11=int6464#3
+# asm 2: pshufd $0xff,<z1=%xmm0,>z11=%xmm2
+pshufd $0xff,%xmm0,%xmm2
+
+# qhasm:   z12 = z1[0,0,0,0]
+# asm 1: pshufd $0x00,<z1=int6464#1,>z12=int6464#4
+# asm 2: pshufd $0x00,<z1=%xmm0,>z12=%xmm3
+pshufd $0x00,%xmm0,%xmm3
+
+# qhasm:   z1 = z1[1,1,1,1]
+# asm 1: pshufd $0x55,<z1=int6464#1,>z1=int6464#1
+# asm 2: pshufd $0x55,<z1=%xmm0,>z1=%xmm0
+pshufd $0x55,%xmm0,%xmm0
+
+# qhasm:   orig6 = z6
+# asm 1: movdqa <z6=int6464#2,>orig6=stack128#9
+# asm 2: movdqa <z6=%xmm1,>orig6=160(%esp)
+movdqa %xmm1,160(%esp)
+
+# qhasm:   orig11 = z11
+# asm 1: movdqa <z11=int6464#3,>orig11=stack128#10
+# asm 2: movdqa <z11=%xmm2,>orig11=176(%esp)
+movdqa %xmm2,176(%esp)
+
+# qhasm:   orig12 = z12
+# asm 1: movdqa <z12=int6464#4,>orig12=stack128#11
+# asm 2: movdqa <z12=%xmm3,>orig12=192(%esp)
+movdqa %xmm3,192(%esp)
+
+# qhasm:   orig1 = z1
+# asm 1: movdqa <z1=int6464#1,>orig1=stack128#12
+# asm 2: movdqa <z1=%xmm0,>orig1=208(%esp)
+movdqa %xmm0,208(%esp)
+
+# qhasm:   z2 = x2
+# asm 1: movdqa <x2=stack128#4,>z2=int6464#1
+# asm 2: movdqa <x2=80(%esp),>z2=%xmm0
+movdqa 80(%esp),%xmm0
+
+# qhasm:   z7 = z2[3,3,3,3]
+# asm 1: pshufd $0xff,<z2=int6464#1,>z7=int6464#2
+# asm 2: pshufd $0xff,<z2=%xmm0,>z7=%xmm1
+pshufd $0xff,%xmm0,%xmm1
+
+# qhasm:   z13 = z2[1,1,1,1]
+# asm 1: pshufd $0x55,<z2=int6464#1,>z13=int6464#3
+# asm 2: pshufd $0x55,<z2=%xmm0,>z13=%xmm2
+pshufd $0x55,%xmm0,%xmm2
+
+# qhasm:   z2 = z2[2,2,2,2]
+# asm 1: pshufd $0xaa,<z2=int6464#1,>z2=int6464#1
+# asm 2: pshufd $0xaa,<z2=%xmm0,>z2=%xmm0
+pshufd $0xaa,%xmm0,%xmm0
+
+# qhasm:   orig7 = z7
+# asm 1: movdqa <z7=int6464#2,>orig7=stack128#13
+# asm 2: movdqa <z7=%xmm1,>orig7=224(%esp)
+movdqa %xmm1,224(%esp)
+
+# qhasm:   orig13 = z13
+# asm 1: movdqa <z13=int6464#3,>orig13=stack128#14
+# asm 2: movdqa <z13=%xmm2,>orig13=240(%esp)
+movdqa %xmm2,240(%esp)
+
+# qhasm:   orig2 = z2
+# asm 1: movdqa <z2=int6464#1,>orig2=stack128#15
+# asm 2: movdqa <z2=%xmm0,>orig2=256(%esp)
+movdqa %xmm0,256(%esp)
+
+# qhasm:   z3 = x3
+# asm 1: movdqa <x3=stack128#1,>z3=int6464#1
+# asm 2: movdqa <x3=32(%esp),>z3=%xmm0
+movdqa 32(%esp),%xmm0
+
+# qhasm:   z4 = z3[0,0,0,0]
+# asm 1: pshufd $0x00,<z3=int6464#1,>z4=int6464#2
+# asm 2: pshufd $0x00,<z3=%xmm0,>z4=%xmm1
+pshufd $0x00,%xmm0,%xmm1
+
+# qhasm:   z14 = z3[2,2,2,2]
+# asm 1: pshufd $0xaa,<z3=int6464#1,>z14=int6464#3
+# asm 2: pshufd $0xaa,<z3=%xmm0,>z14=%xmm2
+pshufd $0xaa,%xmm0,%xmm2
+
+# qhasm:   z3 = z3[3,3,3,3]
+# asm 1: pshufd $0xff,<z3=int6464#1,>z3=int6464#1
+# asm 2: pshufd $0xff,<z3=%xmm0,>z3=%xmm0
+pshufd $0xff,%xmm0,%xmm0
+
+# qhasm:   orig4 = z4
+# asm 1: movdqa <z4=int6464#2,>orig4=stack128#16
+# asm 2: movdqa <z4=%xmm1,>orig4=272(%esp)
+movdqa %xmm1,272(%esp)
+
+# qhasm:   orig14 = z14
+# asm 1: movdqa <z14=int6464#3,>orig14=stack128#17
+# asm 2: movdqa <z14=%xmm2,>orig14=288(%esp)
+movdqa %xmm2,288(%esp)
+
+# qhasm:   orig3 = z3
+# asm 1: movdqa <z3=int6464#1,>orig3=stack128#18
+# asm 2: movdqa <z3=%xmm0,>orig3=304(%esp)
+movdqa %xmm0,304(%esp)
+
+# qhasm: bytesatleast256:
+._bytesatleast256:
+
+# qhasm:   in8 = ((uint32 *)&x2)[0]
+# asm 1: movl <x2=stack128#4,>in8=int32#2
+# asm 2: movl <x2=80(%esp),>in8=%ecx
+movl 80(%esp),%ecx
+
+# qhasm:   in9 = ((uint32 *)&x3)[1]
+# asm 1: movl 4+<x3=stack128#1,>in9=int32#3
+# asm 2: movl 4+<x3=32(%esp),>in9=%edx
+movl 4+32(%esp),%edx
+
+# qhasm:   ((uint32 *) &orig8)[0] = in8
+# asm 1: movl <in8=int32#2,>orig8=stack128#19
+# asm 2: movl <in8=%ecx,>orig8=320(%esp)
+movl %ecx,320(%esp)
+
+# qhasm:   ((uint32 *) &orig9)[0] = in9
+# asm 1: movl <in9=int32#3,>orig9=stack128#20
+# asm 2: movl <in9=%edx,>orig9=336(%esp)
+movl %edx,336(%esp)
+
+# qhasm:   carry? in8 += 1
+# asm 1: add  $1,<in8=int32#2
+# asm 2: add  $1,<in8=%ecx
+add  $1,%ecx
+
+# qhasm:   in9 += 0 + carry
+# asm 1: adc $0,<in9=int32#3
+# asm 2: adc $0,<in9=%edx
+adc $0,%edx
+
+# qhasm:   ((uint32 *) &orig8)[1] = in8
+# asm 1: movl <in8=int32#2,4+<orig8=stack128#19
+# asm 2: movl <in8=%ecx,4+<orig8=320(%esp)
+movl %ecx,4+320(%esp)
+
+# qhasm:   ((uint32 *) &orig9)[1] = in9
+# asm 1: movl <in9=int32#3,4+<orig9=stack128#20
+# asm 2: movl <in9=%edx,4+<orig9=336(%esp)
+movl %edx,4+336(%esp)
+
+# qhasm:   carry? in8 += 1
+# asm 1: add  $1,<in8=int32#2
+# asm 2: add  $1,<in8=%ecx
+add  $1,%ecx
+
+# qhasm:   in9 += 0 + carry
+# asm 1: adc $0,<in9=int32#3
+# asm 2: adc $0,<in9=%edx
+adc $0,%edx
+
+# qhasm:   ((uint32 *) &orig8)[2] = in8
+# asm 1: movl <in8=int32#2,8+<orig8=stack128#19
+# asm 2: movl <in8=%ecx,8+<orig8=320(%esp)
+movl %ecx,8+320(%esp)
+
+# qhasm:   ((uint32 *) &orig9)[2] = in9
+# asm 1: movl <in9=int32#3,8+<orig9=stack128#20
+# asm 2: movl <in9=%edx,8+<orig9=336(%esp)
+movl %edx,8+336(%esp)
+
+# qhasm:   carry? in8 += 1
+# asm 1: add  $1,<in8=int32#2
+# asm 2: add  $1,<in8=%ecx
+add  $1,%ecx
+
+# qhasm:   in9 += 0 + carry
+# asm 1: adc $0,<in9=int32#3
+# asm 2: adc $0,<in9=%edx
+adc $0,%edx
+
+# qhasm:   ((uint32 *) &orig8)[3] = in8
+# asm 1: movl <in8=int32#2,12+<orig8=stack128#19
+# asm 2: movl <in8=%ecx,12+<orig8=320(%esp)
+movl %ecx,12+320(%esp)
+
+# qhasm:   ((uint32 *) &orig9)[3] = in9
+# asm 1: movl <in9=int32#3,12+<orig9=stack128#20
+# asm 2: movl <in9=%edx,12+<orig9=336(%esp)
+movl %edx,12+336(%esp)
+
+# qhasm:   carry? in8 += 1
+# asm 1: add  $1,<in8=int32#2
+# asm 2: add  $1,<in8=%ecx
+add  $1,%ecx
+
+# qhasm:   in9 += 0 + carry
+# asm 1: adc $0,<in9=int32#3
+# asm 2: adc $0,<in9=%edx
+adc $0,%edx
+
+# qhasm:   ((uint32 *)&x2)[0] = in8
+# asm 1: movl <in8=int32#2,>x2=stack128#4
+# asm 2: movl <in8=%ecx,>x2=80(%esp)
+movl %ecx,80(%esp)
+
+# qhasm:   ((uint32 *)&x3)[1] = in9
+# asm 1: movl <in9=int32#3,4+<x3=stack128#1
+# asm 2: movl <in9=%edx,4+<x3=32(%esp)
+movl %edx,4+32(%esp)
+
+# qhasm:   bytes_stack = bytes
+# asm 1: movl <bytes=int32#1,>bytes_stack=stack32#7
+# asm 2: movl <bytes=%eax,>bytes_stack=24(%esp)
+movl %eax,24(%esp)
+
+# qhasm: i = 20
+# asm 1: mov  $20,>i=int32#1
+# asm 2: mov  $20,>i=%eax
+mov  $20,%eax
+
+# qhasm:   z5 = orig5
+# asm 1: movdqa <orig5=stack128#5,>z5=int6464#1
+# asm 2: movdqa <orig5=96(%esp),>z5=%xmm0
+movdqa 96(%esp),%xmm0
+
+# qhasm:   z10 = orig10
+# asm 1: movdqa <orig10=stack128#6,>z10=int6464#2
+# asm 2: movdqa <orig10=112(%esp),>z10=%xmm1
+movdqa 112(%esp),%xmm1
+
+# qhasm:   z15 = orig15
+# asm 1: movdqa <orig15=stack128#7,>z15=int6464#3
+# asm 2: movdqa <orig15=128(%esp),>z15=%xmm2
+movdqa 128(%esp),%xmm2
+
+# qhasm:   z14 = orig14
+# asm 1: movdqa <orig14=stack128#17,>z14=int6464#4
+# asm 2: movdqa <orig14=288(%esp),>z14=%xmm3
+movdqa 288(%esp),%xmm3
+
+# qhasm:   z3 = orig3
+# asm 1: movdqa <orig3=stack128#18,>z3=int6464#5
+# asm 2: movdqa <orig3=304(%esp),>z3=%xmm4
+movdqa 304(%esp),%xmm4
+
+# qhasm:   z6 = orig6
+# asm 1: movdqa <orig6=stack128#9,>z6=int6464#6
+# asm 2: movdqa <orig6=160(%esp),>z6=%xmm5
+movdqa 160(%esp),%xmm5
+
+# qhasm:   z11 = orig11
+# asm 1: movdqa <orig11=stack128#10,>z11=int6464#7
+# asm 2: movdqa <orig11=176(%esp),>z11=%xmm6
+movdqa 176(%esp),%xmm6
+
+# qhasm:   z1 = orig1
+# asm 1: movdqa <orig1=stack128#12,>z1=int6464#8
+# asm 2: movdqa <orig1=208(%esp),>z1=%xmm7
+movdqa 208(%esp),%xmm7
+
+# qhasm:   z5_stack = z5
+# asm 1: movdqa <z5=int6464#1,>z5_stack=stack128#21
+# asm 2: movdqa <z5=%xmm0,>z5_stack=352(%esp)
+movdqa %xmm0,352(%esp)
+
+# qhasm:   z10_stack = z10
+# asm 1: movdqa <z10=int6464#2,>z10_stack=stack128#22
+# asm 2: movdqa <z10=%xmm1,>z10_stack=368(%esp)
+movdqa %xmm1,368(%esp)
+
+# qhasm:   z15_stack = z15
+# asm 1: movdqa <z15=int6464#3,>z15_stack=stack128#23
+# asm 2: movdqa <z15=%xmm2,>z15_stack=384(%esp)
+movdqa %xmm2,384(%esp)
+
+# qhasm:   z14_stack = z14
+# asm 1: movdqa <z14=int6464#4,>z14_stack=stack128#24
+# asm 2: movdqa <z14=%xmm3,>z14_stack=400(%esp)
+movdqa %xmm3,400(%esp)
+
+# qhasm:   z3_stack = z3
+# asm 1: movdqa <z3=int6464#5,>z3_stack=stack128#25
+# asm 2: movdqa <z3=%xmm4,>z3_stack=416(%esp)
+movdqa %xmm4,416(%esp)
+
+# qhasm:   z6_stack = z6
+# asm 1: movdqa <z6=int6464#6,>z6_stack=stack128#26
+# asm 2: movdqa <z6=%xmm5,>z6_stack=432(%esp)
+movdqa %xmm5,432(%esp)
+
+# qhasm:   z11_stack = z11
+# asm 1: movdqa <z11=int6464#7,>z11_stack=stack128#27
+# asm 2: movdqa <z11=%xmm6,>z11_stack=448(%esp)
+movdqa %xmm6,448(%esp)
+
+# qhasm:   z1_stack = z1
+# asm 1: movdqa <z1=int6464#8,>z1_stack=stack128#28
+# asm 2: movdqa <z1=%xmm7,>z1_stack=464(%esp)
+movdqa %xmm7,464(%esp)
+
+# qhasm:   z7 = orig7
+# asm 1: movdqa <orig7=stack128#13,>z7=int6464#5
+# asm 2: movdqa <orig7=224(%esp),>z7=%xmm4
+movdqa 224(%esp),%xmm4
+
+# qhasm:   z13 = orig13
+# asm 1: movdqa <orig13=stack128#14,>z13=int6464#6
+# asm 2: movdqa <orig13=240(%esp),>z13=%xmm5
+movdqa 240(%esp),%xmm5
+
+# qhasm:   z2 = orig2
+# asm 1: movdqa <orig2=stack128#15,>z2=int6464#7
+# asm 2: movdqa <orig2=256(%esp),>z2=%xmm6
+movdqa 256(%esp),%xmm6
+
+# qhasm:   z9 = orig9
+# asm 1: movdqa <orig9=stack128#20,>z9=int6464#8
+# asm 2: movdqa <orig9=336(%esp),>z9=%xmm7
+movdqa 336(%esp),%xmm7
+
+# qhasm:                   p = orig0
+# asm 1: movdqa <orig0=stack128#8,>p=int6464#1
+# asm 2: movdqa <orig0=144(%esp),>p=%xmm0
+movdqa 144(%esp),%xmm0
+
+# qhasm:                   t = orig12
+# asm 1: movdqa <orig12=stack128#11,>t=int6464#3
+# asm 2: movdqa <orig12=192(%esp),>t=%xmm2
+movdqa 192(%esp),%xmm2
+
+# qhasm:                   q = orig4
+# asm 1: movdqa <orig4=stack128#16,>q=int6464#4
+# asm 2: movdqa <orig4=272(%esp),>q=%xmm3
+movdqa 272(%esp),%xmm3
+
+# qhasm:                   r = orig8
+# asm 1: movdqa <orig8=stack128#19,>r=int6464#2
+# asm 2: movdqa <orig8=320(%esp),>r=%xmm1
+movdqa 320(%esp),%xmm1
+
+# qhasm:   z7_stack = z7
+# asm 1: movdqa <z7=int6464#5,>z7_stack=stack128#29
+# asm 2: movdqa <z7=%xmm4,>z7_stack=480(%esp)
+movdqa %xmm4,480(%esp)
+
+# qhasm:   z13_stack = z13
+# asm 1: movdqa <z13=int6464#6,>z13_stack=stack128#30
+# asm 2: movdqa <z13=%xmm5,>z13_stack=496(%esp)
+movdqa %xmm5,496(%esp)
+
+# qhasm:   z2_stack = z2
+# asm 1: movdqa <z2=int6464#7,>z2_stack=stack128#31
+# asm 2: movdqa <z2=%xmm6,>z2_stack=512(%esp)
+movdqa %xmm6,512(%esp)
+
+# qhasm:   z9_stack = z9
+# asm 1: movdqa <z9=int6464#8,>z9_stack=stack128#32
+# asm 2: movdqa <z9=%xmm7,>z9_stack=528(%esp)
+movdqa %xmm7,528(%esp)
+
+# qhasm:   z0_stack = p
+# asm 1: movdqa <p=int6464#1,>z0_stack=stack128#33
+# asm 2: movdqa <p=%xmm0,>z0_stack=544(%esp)
+movdqa %xmm0,544(%esp)
+
+# qhasm:   z12_stack = t
+# asm 1: movdqa <t=int6464#3,>z12_stack=stack128#34
+# asm 2: movdqa <t=%xmm2,>z12_stack=560(%esp)
+movdqa %xmm2,560(%esp)
+
+# qhasm:   z4_stack = q
+# asm 1: movdqa <q=int6464#4,>z4_stack=stack128#35
+# asm 2: movdqa <q=%xmm3,>z4_stack=576(%esp)
+movdqa %xmm3,576(%esp)
+
+# qhasm:   z8_stack = r
+# asm 1: movdqa <r=int6464#2,>z8_stack=stack128#36
+# asm 2: movdqa <r=%xmm1,>z8_stack=592(%esp)
+movdqa %xmm1,592(%esp)
+
+# qhasm: mainloop1:
+._mainloop1:
+
+# qhasm:                   assign xmm0 to p
+
+# qhasm:                   assign xmm1 to r
+
+# qhasm:                   assign xmm2 to t
+
+# qhasm:                   assign xmm3 to q
+
+# qhasm:                   s = t
+# asm 1: movdqa <t=int6464#3,>s=int6464#7
+# asm 2: movdqa <t=%xmm2,>s=%xmm6
+movdqa %xmm2,%xmm6
+
+# qhasm: uint32323232      t += p
+# asm 1: paddd <p=int6464#1,<t=int6464#3
+# asm 2: paddd <p=%xmm0,<t=%xmm2
+paddd %xmm0,%xmm2
+
+# qhasm:                   u = t
+# asm 1: movdqa <t=int6464#3,>u=int6464#5
+# asm 2: movdqa <t=%xmm2,>u=%xmm4
+movdqa %xmm2,%xmm4
+
+# qhasm: uint32323232      t >>= 25
+# asm 1: psrld $25,<t=int6464#3
+# asm 2: psrld $25,<t=%xmm2
+psrld $25,%xmm2
+
+# qhasm:                   q ^= t
+# asm 1: pxor  <t=int6464#3,<q=int6464#4
+# asm 2: pxor  <t=%xmm2,<q=%xmm3
+pxor  %xmm2,%xmm3
+
+# qhasm: uint32323232      u <<= 7
+# asm 1: pslld $7,<u=int6464#5
+# asm 2: pslld $7,<u=%xmm4
+pslld $7,%xmm4
+
+# qhasm:                   q ^= u
+# asm 1: pxor  <u=int6464#5,<q=int6464#4
+# asm 2: pxor  <u=%xmm4,<q=%xmm3
+pxor  %xmm4,%xmm3
+
+# qhasm:                   z4_stack = q
+# asm 1: movdqa <q=int6464#4,>z4_stack=stack128#33
+# asm 2: movdqa <q=%xmm3,>z4_stack=544(%esp)
+movdqa %xmm3,544(%esp)
+
+# qhasm:                   t = p
+# asm 1: movdqa <p=int6464#1,>t=int6464#3
+# asm 2: movdqa <p=%xmm0,>t=%xmm2
+movdqa %xmm0,%xmm2
+
+# qhasm: uint32323232      t += q
+# asm 1: paddd <q=int6464#4,<t=int6464#3
+# asm 2: paddd <q=%xmm3,<t=%xmm2
+paddd %xmm3,%xmm2
+
+# qhasm:                   u = t
+# asm 1: movdqa <t=int6464#3,>u=int6464#5
+# asm 2: movdqa <t=%xmm2,>u=%xmm4
+movdqa %xmm2,%xmm4
+
+# qhasm: uint32323232      t >>= 23
+# asm 1: psrld $23,<t=int6464#3
+# asm 2: psrld $23,<t=%xmm2
+psrld $23,%xmm2
+
+# qhasm:                   r ^= t
+# asm 1: pxor  <t=int6464#3,<r=int6464#2
+# asm 2: pxor  <t=%xmm2,<r=%xmm1
+pxor  %xmm2,%xmm1
+
+# qhasm: uint32323232      u <<= 9
+# asm 1: pslld $9,<u=int6464#5
+# asm 2: pslld $9,<u=%xmm4
+pslld $9,%xmm4
+
+# qhasm:                   r ^= u
+# asm 1: pxor  <u=int6464#5,<r=int6464#2
+# asm 2: pxor  <u=%xmm4,<r=%xmm1
+pxor  %xmm4,%xmm1
+
+# qhasm:                   z8_stack = r
+# asm 1: movdqa <r=int6464#2,>z8_stack=stack128#34
+# asm 2: movdqa <r=%xmm1,>z8_stack=560(%esp)
+movdqa %xmm1,560(%esp)
+
+# qhasm: uint32323232      q += r
+# asm 1: paddd <r=int6464#2,<q=int6464#4
+# asm 2: paddd <r=%xmm1,<q=%xmm3
+paddd %xmm1,%xmm3
+
+# qhasm:                   u = q
+# asm 1: movdqa <q=int6464#4,>u=int6464#3
+# asm 2: movdqa <q=%xmm3,>u=%xmm2
+movdqa %xmm3,%xmm2
+
+# qhasm: uint32323232      q >>= 19
+# asm 1: psrld $19,<q=int6464#4
+# asm 2: psrld $19,<q=%xmm3
+psrld $19,%xmm3
+
+# qhasm:                   s ^= q
+# asm 1: pxor  <q=int6464#4,<s=int6464#7
+# asm 2: pxor  <q=%xmm3,<s=%xmm6
+pxor  %xmm3,%xmm6
+
+# qhasm: uint32323232      u <<= 13
+# asm 1: pslld $13,<u=int6464#3
+# asm 2: pslld $13,<u=%xmm2
+pslld $13,%xmm2
+
+# qhasm:                   s ^= u
+# asm 1: pxor  <u=int6464#3,<s=int6464#7
+# asm 2: pxor  <u=%xmm2,<s=%xmm6
+pxor  %xmm2,%xmm6
+
+# qhasm:                                   mt = z1_stack
+# asm 1: movdqa <z1_stack=stack128#28,>mt=int6464#3
+# asm 2: movdqa <z1_stack=464(%esp),>mt=%xmm2
+movdqa 464(%esp),%xmm2
+
+# qhasm:                                   mp = z5_stack
+# asm 1: movdqa <z5_stack=stack128#21,>mp=int6464#5
+# asm 2: movdqa <z5_stack=352(%esp),>mp=%xmm4
+movdqa 352(%esp),%xmm4
+
+# qhasm:                                   mq = z9_stack
+# asm 1: movdqa <z9_stack=stack128#32,>mq=int6464#4
+# asm 2: movdqa <z9_stack=528(%esp),>mq=%xmm3
+movdqa 528(%esp),%xmm3
+
+# qhasm:                                   mr = z13_stack
+# asm 1: movdqa <z13_stack=stack128#30,>mr=int6464#6
+# asm 2: movdqa <z13_stack=496(%esp),>mr=%xmm5
+movdqa 496(%esp),%xmm5
+
+# qhasm:                   z12_stack = s
+# asm 1: movdqa <s=int6464#7,>z12_stack=stack128#30
+# asm 2: movdqa <s=%xmm6,>z12_stack=496(%esp)
+movdqa %xmm6,496(%esp)
+
+# qhasm: uint32323232      r += s
+# asm 1: paddd <s=int6464#7,<r=int6464#2
+# asm 2: paddd <s=%xmm6,<r=%xmm1
+paddd %xmm6,%xmm1
+
+# qhasm:                   u = r
+# asm 1: movdqa <r=int6464#2,>u=int6464#7
+# asm 2: movdqa <r=%xmm1,>u=%xmm6
+movdqa %xmm1,%xmm6
+
+# qhasm: uint32323232      r >>= 14
+# asm 1: psrld $14,<r=int6464#2
+# asm 2: psrld $14,<r=%xmm1
+psrld $14,%xmm1
+
+# qhasm:                   p ^= r
+# asm 1: pxor  <r=int6464#2,<p=int6464#1
+# asm 2: pxor  <r=%xmm1,<p=%xmm0
+pxor  %xmm1,%xmm0
+
+# qhasm: uint32323232      u <<= 18
+# asm 1: pslld $18,<u=int6464#7
+# asm 2: pslld $18,<u=%xmm6
+pslld $18,%xmm6
+
+# qhasm:                   p ^= u
+# asm 1: pxor  <u=int6464#7,<p=int6464#1
+# asm 2: pxor  <u=%xmm6,<p=%xmm0
+pxor  %xmm6,%xmm0
+
+# qhasm:                   z0_stack = p
+# asm 1: movdqa <p=int6464#1,>z0_stack=stack128#21
+# asm 2: movdqa <p=%xmm0,>z0_stack=352(%esp)
+movdqa %xmm0,352(%esp)
+
+# qhasm:                                   assign xmm2 to mt
+
+# qhasm:                                   assign xmm3 to mq
+
+# qhasm:                                   assign xmm4 to mp
+
+# qhasm:                                   assign xmm5 to mr
+
+# qhasm:                                   ms = mt
+# asm 1: movdqa <mt=int6464#3,>ms=int6464#7
+# asm 2: movdqa <mt=%xmm2,>ms=%xmm6
+movdqa %xmm2,%xmm6
+
+# qhasm: uint32323232                      mt += mp
+# asm 1: paddd <mp=int6464#5,<mt=int6464#3
+# asm 2: paddd <mp=%xmm4,<mt=%xmm2
+paddd %xmm4,%xmm2
+
+# qhasm:                                   mu = mt
+# asm 1: movdqa <mt=int6464#3,>mu=int6464#1
+# asm 2: movdqa <mt=%xmm2,>mu=%xmm0
+movdqa %xmm2,%xmm0
+
+# qhasm: uint32323232                      mt >>= 25
+# asm 1: psrld $25,<mt=int6464#3
+# asm 2: psrld $25,<mt=%xmm2
+psrld $25,%xmm2
+
+# qhasm:                                   mq ^= mt
+# asm 1: pxor  <mt=int6464#3,<mq=int6464#4
+# asm 2: pxor  <mt=%xmm2,<mq=%xmm3
+pxor  %xmm2,%xmm3
+
+# qhasm: uint32323232                      mu <<= 7
+# asm 1: pslld $7,<mu=int6464#1
+# asm 2: pslld $7,<mu=%xmm0
+pslld $7,%xmm0
+
+# qhasm:                                   mq ^= mu
+# asm 1: pxor  <mu=int6464#1,<mq=int6464#4
+# asm 2: pxor  <mu=%xmm0,<mq=%xmm3
+pxor  %xmm0,%xmm3
+
+# qhasm:                                   z9_stack = mq
+# asm 1: movdqa <mq=int6464#4,>z9_stack=stack128#32
+# asm 2: movdqa <mq=%xmm3,>z9_stack=528(%esp)
+movdqa %xmm3,528(%esp)
+
+# qhasm:                                   mt = mp
+# asm 1: movdqa <mp=int6464#5,>mt=int6464#1
+# asm 2: movdqa <mp=%xmm4,>mt=%xmm0
+movdqa %xmm4,%xmm0
+
+# qhasm: uint32323232                      mt += mq
+# asm 1: paddd <mq=int6464#4,<mt=int6464#1
+# asm 2: paddd <mq=%xmm3,<mt=%xmm0
+paddd %xmm3,%xmm0
+
+# qhasm:                                   mu = mt
+# asm 1: movdqa <mt=int6464#1,>mu=int6464#2
+# asm 2: movdqa <mt=%xmm0,>mu=%xmm1
+movdqa %xmm0,%xmm1
+
+# qhasm: uint32323232                      mt >>= 23
+# asm 1: psrld $23,<mt=int6464#1
+# asm 2: psrld $23,<mt=%xmm0
+psrld $23,%xmm0
+
+# qhasm:                                   mr ^= mt
+# asm 1: pxor  <mt=int6464#1,<mr=int6464#6
+# asm 2: pxor  <mt=%xmm0,<mr=%xmm5
+pxor  %xmm0,%xmm5
+
+# qhasm: uint32323232                      mu <<= 9
+# asm 1: pslld $9,<mu=int6464#2
+# asm 2: pslld $9,<mu=%xmm1
+pslld $9,%xmm1
+
+# qhasm:                                   mr ^= mu
+# asm 1: pxor  <mu=int6464#2,<mr=int6464#6
+# asm 2: pxor  <mu=%xmm1,<mr=%xmm5
+pxor  %xmm1,%xmm5
+
+# qhasm:                                   z13_stack = mr
+# asm 1: movdqa <mr=int6464#6,>z13_stack=stack128#35
+# asm 2: movdqa <mr=%xmm5,>z13_stack=576(%esp)
+movdqa %xmm5,576(%esp)
+
+# qhasm: uint32323232                      mq += mr
+# asm 1: paddd <mr=int6464#6,<mq=int6464#4
+# asm 2: paddd <mr=%xmm5,<mq=%xmm3
+paddd %xmm5,%xmm3
+
+# qhasm:                                   mu = mq
+# asm 1: movdqa <mq=int6464#4,>mu=int6464#1
+# asm 2: movdqa <mq=%xmm3,>mu=%xmm0
+movdqa %xmm3,%xmm0
+
+# qhasm: uint32323232                      mq >>= 19
+# asm 1: psrld $19,<mq=int6464#4
+# asm 2: psrld $19,<mq=%xmm3
+psrld $19,%xmm3
+
+# qhasm:                                   ms ^= mq
+# asm 1: pxor  <mq=int6464#4,<ms=int6464#7
+# asm 2: pxor  <mq=%xmm3,<ms=%xmm6
+pxor  %xmm3,%xmm6
+
+# qhasm: uint32323232                      mu <<= 13
+# asm 1: pslld $13,<mu=int6464#1
+# asm 2: pslld $13,<mu=%xmm0
+pslld $13,%xmm0
+
+# qhasm:                                   ms ^= mu
+# asm 1: pxor  <mu=int6464#1,<ms=int6464#7
+# asm 2: pxor  <mu=%xmm0,<ms=%xmm6
+pxor  %xmm0,%xmm6
+
+# qhasm:                                                   t = z6_stack
+# asm 1: movdqa <z6_stack=stack128#26,>t=int6464#3
+# asm 2: movdqa <z6_stack=432(%esp),>t=%xmm2
+movdqa 432(%esp),%xmm2
+
+# qhasm:                                                   p = z10_stack
+# asm 1: movdqa <z10_stack=stack128#22,>p=int6464#1
+# asm 2: movdqa <z10_stack=368(%esp),>p=%xmm0
+movdqa 368(%esp),%xmm0
+
+# qhasm:                                                   q = z14_stack
+# asm 1: movdqa <z14_stack=stack128#24,>q=int6464#4
+# asm 2: movdqa <z14_stack=400(%esp),>q=%xmm3
+movdqa 400(%esp),%xmm3
+
+# qhasm:                                                   r = z2_stack
+# asm 1: movdqa <z2_stack=stack128#31,>r=int6464#2
+# asm 2: movdqa <z2_stack=512(%esp),>r=%xmm1
+movdqa 512(%esp),%xmm1
+
+# qhasm:                                   z1_stack = ms
+# asm 1: movdqa <ms=int6464#7,>z1_stack=stack128#22
+# asm 2: movdqa <ms=%xmm6,>z1_stack=368(%esp)
+movdqa %xmm6,368(%esp)
+
+# qhasm: uint32323232                      mr += ms
+# asm 1: paddd <ms=int6464#7,<mr=int6464#6
+# asm 2: paddd <ms=%xmm6,<mr=%xmm5
+paddd %xmm6,%xmm5
+
+# qhasm:                                   mu = mr
+# asm 1: movdqa <mr=int6464#6,>mu=int6464#7
+# asm 2: movdqa <mr=%xmm5,>mu=%xmm6
+movdqa %xmm5,%xmm6
+
+# qhasm: uint32323232                      mr >>= 14
+# asm 1: psrld $14,<mr=int6464#6
+# asm 2: psrld $14,<mr=%xmm5
+psrld $14,%xmm5
+
+# qhasm:                                   mp ^= mr
+# asm 1: pxor  <mr=int6464#6,<mp=int6464#5
+# asm 2: pxor  <mr=%xmm5,<mp=%xmm4
+pxor  %xmm5,%xmm4
+
+# qhasm: uint32323232                      mu <<= 18
+# asm 1: pslld $18,<mu=int6464#7
+# asm 2: pslld $18,<mu=%xmm6
+pslld $18,%xmm6
+
+# qhasm:                                   mp ^= mu
+# asm 1: pxor  <mu=int6464#7,<mp=int6464#5
+# asm 2: pxor  <mu=%xmm6,<mp=%xmm4
+pxor  %xmm6,%xmm4
+
+# qhasm:                                   z5_stack = mp
+# asm 1: movdqa <mp=int6464#5,>z5_stack=stack128#24
+# asm 2: movdqa <mp=%xmm4,>z5_stack=400(%esp)
+movdqa %xmm4,400(%esp)
+
+# qhasm:                                                   assign xmm0 to p
+
+# qhasm:                                                   assign xmm1 to r
+
+# qhasm:                                                   assign xmm2 to t
+
+# qhasm:                                                   assign xmm3 to q
+
+# qhasm:                                                   s = t
+# asm 1: movdqa <t=int6464#3,>s=int6464#7
+# asm 2: movdqa <t=%xmm2,>s=%xmm6
+movdqa %xmm2,%xmm6
+
+# qhasm: uint32323232                                      t += p
+# asm 1: paddd <p=int6464#1,<t=int6464#3
+# asm 2: paddd <p=%xmm0,<t=%xmm2
+paddd %xmm0,%xmm2
+
+# qhasm:                                                   u = t
+# asm 1: movdqa <t=int6464#3,>u=int6464#5
+# asm 2: movdqa <t=%xmm2,>u=%xmm4
+movdqa %xmm2,%xmm4
+
+# qhasm: uint32323232                                      t >>= 25
+# asm 1: psrld $25,<t=int6464#3
+# asm 2: psrld $25,<t=%xmm2
+psrld $25,%xmm2
+
+# qhasm:                                                   q ^= t
+# asm 1: pxor  <t=int6464#3,<q=int6464#4
+# asm 2: pxor  <t=%xmm2,<q=%xmm3
+pxor  %xmm2,%xmm3
+
+# qhasm: uint32323232                                      u <<= 7
+# asm 1: pslld $7,<u=int6464#5
+# asm 2: pslld $7,<u=%xmm4
+pslld $7,%xmm4
+
+# qhasm:                                                   q ^= u
+# asm 1: pxor  <u=int6464#5,<q=int6464#4
+# asm 2: pxor  <u=%xmm4,<q=%xmm3
+pxor  %xmm4,%xmm3
+
+# qhasm:                                                   z14_stack = q
+# asm 1: movdqa <q=int6464#4,>z14_stack=stack128#36
+# asm 2: movdqa <q=%xmm3,>z14_stack=592(%esp)
+movdqa %xmm3,592(%esp)
+
+# qhasm:                                                   t = p
+# asm 1: movdqa <p=int6464#1,>t=int6464#3
+# asm 2: movdqa <p=%xmm0,>t=%xmm2
+movdqa %xmm0,%xmm2
+
+# qhasm: uint32323232                                      t += q
+# asm 1: paddd <q=int6464#4,<t=int6464#3
+# asm 2: paddd <q=%xmm3,<t=%xmm2
+paddd %xmm3,%xmm2
+
+# qhasm:                                                   u = t
+# asm 1: movdqa <t=int6464#3,>u=int6464#5
+# asm 2: movdqa <t=%xmm2,>u=%xmm4
+movdqa %xmm2,%xmm4
+
+# qhasm: uint32323232                                      t >>= 23
+# asm 1: psrld $23,<t=int6464#3
+# asm 2: psrld $23,<t=%xmm2
+psrld $23,%xmm2
+
+# qhasm:                                                   r ^= t
+# asm 1: pxor  <t=int6464#3,<r=int6464#2
+# asm 2: pxor  <t=%xmm2,<r=%xmm1
+pxor  %xmm2,%xmm1
+
+# qhasm: uint32323232                                      u <<= 9
+# asm 1: pslld $9,<u=int6464#5
+# asm 2: pslld $9,<u=%xmm4
+pslld $9,%xmm4
+
+# qhasm:                                                   r ^= u
+# asm 1: pxor  <u=int6464#5,<r=int6464#2
+# asm 2: pxor  <u=%xmm4,<r=%xmm1
+pxor  %xmm4,%xmm1
+
+# qhasm:                                                   z2_stack = r
+# asm 1: movdqa <r=int6464#2,>z2_stack=stack128#26
+# asm 2: movdqa <r=%xmm1,>z2_stack=432(%esp)
+movdqa %xmm1,432(%esp)
+
+# qhasm: uint32323232                                      q += r
+# asm 1: paddd <r=int6464#2,<q=int6464#4
+# asm 2: paddd <r=%xmm1,<q=%xmm3
+paddd %xmm1,%xmm3
+
+# qhasm:                                                   u = q
+# asm 1: movdqa <q=int6464#4,>u=int6464#3
+# asm 2: movdqa <q=%xmm3,>u=%xmm2
+movdqa %xmm3,%xmm2
+
+# qhasm: uint32323232                                      q >>= 19
+# asm 1: psrld $19,<q=int6464#4
+# asm 2: psrld $19,<q=%xmm3
+psrld $19,%xmm3
+
+# qhasm:                                                   s ^= q
+# asm 1: pxor  <q=int6464#4,<s=int6464#7
+# asm 2: pxor  <q=%xmm3,<s=%xmm6
+pxor  %xmm3,%xmm6
+
+# qhasm: uint32323232                                      u <<= 13
+# asm 1: pslld $13,<u=int6464#3
+# asm 2: pslld $13,<u=%xmm2
+pslld $13,%xmm2
+
+# qhasm:                                                   s ^= u
+# asm 1: pxor  <u=int6464#3,<s=int6464#7
+# asm 2: pxor  <u=%xmm2,<s=%xmm6
+pxor  %xmm2,%xmm6
+
+# qhasm:                                                                   mt = z11_stack
+# asm 1: movdqa <z11_stack=stack128#27,>mt=int6464#3
+# asm 2: movdqa <z11_stack=448(%esp),>mt=%xmm2
+movdqa 448(%esp),%xmm2
+
+# qhasm:                                                                   mp = z15_stack
+# asm 1: movdqa <z15_stack=stack128#23,>mp=int6464#5
+# asm 2: movdqa <z15_stack=384(%esp),>mp=%xmm4
+movdqa 384(%esp),%xmm4
+
+# qhasm:                                                                   mq = z3_stack
+# asm 1: movdqa <z3_stack=stack128#25,>mq=int6464#4
+# asm 2: movdqa <z3_stack=416(%esp),>mq=%xmm3
+movdqa 416(%esp),%xmm3
+
+# qhasm:                                                                   mr = z7_stack
+# asm 1: movdqa <z7_stack=stack128#29,>mr=int6464#6
+# asm 2: movdqa <z7_stack=480(%esp),>mr=%xmm5
+movdqa 480(%esp),%xmm5
+
+# qhasm:                                                   z6_stack = s
+# asm 1: movdqa <s=int6464#7,>z6_stack=stack128#23
+# asm 2: movdqa <s=%xmm6,>z6_stack=384(%esp)
+movdqa %xmm6,384(%esp)
+
+# qhasm: uint32323232                                      r += s
+# asm 1: paddd <s=int6464#7,<r=int6464#2
+# asm 2: paddd <s=%xmm6,<r=%xmm1
+paddd %xmm6,%xmm1
+
+# qhasm:                                                   u = r
+# asm 1: movdqa <r=int6464#2,>u=int6464#7
+# asm 2: movdqa <r=%xmm1,>u=%xmm6
+movdqa %xmm1,%xmm6
+
+# qhasm: uint32323232                                      r >>= 14
+# asm 1: psrld $14,<r=int6464#2
+# asm 2: psrld $14,<r=%xmm1
+psrld $14,%xmm1
+
+# qhasm:                                                   p ^= r
+# asm 1: pxor  <r=int6464#2,<p=int6464#1
+# asm 2: pxor  <r=%xmm1,<p=%xmm0
+pxor  %xmm1,%xmm0
+
+# qhasm: uint32323232                                      u <<= 18
+# asm 1: pslld $18,<u=int6464#7
+# asm 2: pslld $18,<u=%xmm6
+pslld $18,%xmm6
+
+# qhasm:                                                   p ^= u
+# asm 1: pxor  <u=int6464#7,<p=int6464#1
+# asm 2: pxor  <u=%xmm6,<p=%xmm0
+pxor  %xmm6,%xmm0
+
+# qhasm:                                                   z10_stack = p
+# asm 1: movdqa <p=int6464#1,>z10_stack=stack128#27
+# asm 2: movdqa <p=%xmm0,>z10_stack=448(%esp)
+movdqa %xmm0,448(%esp)
+
+# qhasm:                                                                   assign xmm2 to mt
+
+# qhasm:                                                                   assign xmm3 to mq
+
+# qhasm:                                                                   assign xmm4 to mp
+
+# qhasm:                                                                   assign xmm5 to mr
+
+# qhasm:                                                                   ms = mt
+# asm 1: movdqa <mt=int6464#3,>ms=int6464#7
+# asm 2: movdqa <mt=%xmm2,>ms=%xmm6
+movdqa %xmm2,%xmm6
+
+# qhasm: uint32323232                                                      mt += mp
+# asm 1: paddd <mp=int6464#5,<mt=int6464#3
+# asm 2: paddd <mp=%xmm4,<mt=%xmm2
+paddd %xmm4,%xmm2
+
+# qhasm:                                                                   mu = mt
+# asm 1: movdqa <mt=int6464#3,>mu=int6464#1
+# asm 2: movdqa <mt=%xmm2,>mu=%xmm0
+movdqa %xmm2,%xmm0
+
+# qhasm: uint32323232                                                      mt >>= 25
+# asm 1: psrld $25,<mt=int6464#3
+# asm 2: psrld $25,<mt=%xmm2
+psrld $25,%xmm2
+
+# qhasm:                                                                   mq ^= mt
+# asm 1: pxor  <mt=int6464#3,<mq=int6464#4
+# asm 2: pxor  <mt=%xmm2,<mq=%xmm3
+pxor  %xmm2,%xmm3
+
+# qhasm: uint32323232                                                      mu <<= 7
+# asm 1: pslld $7,<mu=int6464#1
+# asm 2: pslld $7,<mu=%xmm0
+pslld $7,%xmm0
+
+# qhasm:                                                                   mq ^= mu
+# asm 1: pxor  <mu=int6464#1,<mq=int6464#4
+# asm 2: pxor  <mu=%xmm0,<mq=%xmm3
+pxor  %xmm0,%xmm3
+
+# qhasm:                                                                   z3_stack = mq
+# asm 1: movdqa <mq=int6464#4,>z3_stack=stack128#25
+# asm 2: movdqa <mq=%xmm3,>z3_stack=416(%esp)
+movdqa %xmm3,416(%esp)
+
+# qhasm:                                                                   mt = mp
+# asm 1: movdqa <mp=int6464#5,>mt=int6464#1
+# asm 2: movdqa <mp=%xmm4,>mt=%xmm0
+movdqa %xmm4,%xmm0
+
+# qhasm: uint32323232                                                      mt += mq
+# asm 1: paddd <mq=int6464#4,<mt=int6464#1
+# asm 2: paddd <mq=%xmm3,<mt=%xmm0
+paddd %xmm3,%xmm0
+
+# qhasm:                                                                   mu = mt
+# asm 1: movdqa <mt=int6464#1,>mu=int6464#2
+# asm 2: movdqa <mt=%xmm0,>mu=%xmm1
+movdqa %xmm0,%xmm1
+
+# qhasm: uint32323232                                                      mt >>= 23
+# asm 1: psrld $23,<mt=int6464#1
+# asm 2: psrld $23,<mt=%xmm0
+psrld $23,%xmm0
+
+# qhasm:                                                                   mr ^= mt
+# asm 1: pxor  <mt=int6464#1,<mr=int6464#6
+# asm 2: pxor  <mt=%xmm0,<mr=%xmm5
+pxor  %xmm0,%xmm5
+
+# qhasm: uint32323232                                                      mu <<= 9
+# asm 1: pslld $9,<mu=int6464#2
+# asm 2: pslld $9,<mu=%xmm1
+pslld $9,%xmm1
+
+# qhasm:                                                                   mr ^= mu
+# asm 1: pxor  <mu=int6464#2,<mr=int6464#6
+# asm 2: pxor  <mu=%xmm1,<mr=%xmm5
+pxor  %xmm1,%xmm5
+
+# qhasm:                                                                   z7_stack = mr
+# asm 1: movdqa <mr=int6464#6,>z7_stack=stack128#29
+# asm 2: movdqa <mr=%xmm5,>z7_stack=480(%esp)
+movdqa %xmm5,480(%esp)
+
+# qhasm: uint32323232                                                      mq += mr
+# asm 1: paddd <mr=int6464#6,<mq=int6464#4
+# asm 2: paddd <mr=%xmm5,<mq=%xmm3
+paddd %xmm5,%xmm3
+
+# qhasm:                                                                   mu = mq
+# asm 1: movdqa <mq=int6464#4,>mu=int6464#1
+# asm 2: movdqa <mq=%xmm3,>mu=%xmm0
+movdqa %xmm3,%xmm0
+
+# qhasm: uint32323232                                                      mq >>= 19
+# asm 1: psrld $19,<mq=int6464#4
+# asm 2: psrld $19,<mq=%xmm3
+psrld $19,%xmm3
+
+# qhasm:                                                                   ms ^= mq
+# asm 1: pxor  <mq=int6464#4,<ms=int6464#7
+# asm 2: pxor  <mq=%xmm3,<ms=%xmm6
+pxor  %xmm3,%xmm6
+
+# qhasm: uint32323232                                                      mu <<= 13
+# asm 1: pslld $13,<mu=int6464#1
+# asm 2: pslld $13,<mu=%xmm0
+pslld $13,%xmm0
+
+# qhasm:                                                                   ms ^= mu
+# asm 1: pxor  <mu=int6464#1,<ms=int6464#7
+# asm 2: pxor  <mu=%xmm0,<ms=%xmm6
+pxor  %xmm0,%xmm6
+
+# qhasm:                   t = z3_stack
+# asm 1: movdqa <z3_stack=stack128#25,>t=int6464#3
+# asm 2: movdqa <z3_stack=416(%esp),>t=%xmm2
+movdqa 416(%esp),%xmm2
+
+# qhasm:                   p = z0_stack
+# asm 1: movdqa <z0_stack=stack128#21,>p=int6464#1
+# asm 2: movdqa <z0_stack=352(%esp),>p=%xmm0
+movdqa 352(%esp),%xmm0
+
+# qhasm:                   q = z1_stack
+# asm 1: movdqa <z1_stack=stack128#22,>q=int6464#4
+# asm 2: movdqa <z1_stack=368(%esp),>q=%xmm3
+movdqa 368(%esp),%xmm3
+
+# qhasm:                   r = z2_stack
+# asm 1: movdqa <z2_stack=stack128#26,>r=int6464#2
+# asm 2: movdqa <z2_stack=432(%esp),>r=%xmm1
+movdqa 432(%esp),%xmm1
+
+# qhasm:                                                                   z11_stack = ms
+# asm 1: movdqa <ms=int6464#7,>z11_stack=stack128#21
+# asm 2: movdqa <ms=%xmm6,>z11_stack=352(%esp)
+movdqa %xmm6,352(%esp)
+
+# qhasm: uint32323232                                                      mr += ms
+# asm 1: paddd <ms=int6464#7,<mr=int6464#6
+# asm 2: paddd <ms=%xmm6,<mr=%xmm5
+paddd %xmm6,%xmm5
+
+# qhasm:                                                                   mu = mr
+# asm 1: movdqa <mr=int6464#6,>mu=int6464#7
+# asm 2: movdqa <mr=%xmm5,>mu=%xmm6
+movdqa %xmm5,%xmm6
+
+# qhasm: uint32323232                                                      mr >>= 14
+# asm 1: psrld $14,<mr=int6464#6
+# asm 2: psrld $14,<mr=%xmm5
+psrld $14,%xmm5
+
+# qhasm:                                                                   mp ^= mr
+# asm 1: pxor  <mr=int6464#6,<mp=int6464#5
+# asm 2: pxor  <mr=%xmm5,<mp=%xmm4
+pxor  %xmm5,%xmm4
+
+# qhasm: uint32323232                                                      mu <<= 18
+# asm 1: pslld $18,<mu=int6464#7
+# asm 2: pslld $18,<mu=%xmm6
+pslld $18,%xmm6
+
+# qhasm:                                                                   mp ^= mu
+# asm 1: pxor  <mu=int6464#7,<mp=int6464#5
+# asm 2: pxor  <mu=%xmm6,<mp=%xmm4
+pxor  %xmm6,%xmm4
+
+# qhasm:                                                                   z15_stack = mp
+# asm 1: movdqa <mp=int6464#5,>z15_stack=stack128#22
+# asm 2: movdqa <mp=%xmm4,>z15_stack=368(%esp)
+movdqa %xmm4,368(%esp)
+
+# qhasm:                   assign xmm0 to p
+
+# qhasm:                   assign xmm1 to r
+
+# qhasm:                   assign xmm2 to t
+
+# qhasm:                   assign xmm3 to q
+
+# qhasm:                   s = t
+# asm 1: movdqa <t=int6464#3,>s=int6464#7
+# asm 2: movdqa <t=%xmm2,>s=%xmm6
+movdqa %xmm2,%xmm6
+
+# qhasm: uint32323232      t += p
+# asm 1: paddd <p=int6464#1,<t=int6464#3
+# asm 2: paddd <p=%xmm0,<t=%xmm2
+paddd %xmm0,%xmm2
+
+# qhasm:                   u = t
+# asm 1: movdqa <t=int6464#3,>u=int6464#5
+# asm 2: movdqa <t=%xmm2,>u=%xmm4
+movdqa %xmm2,%xmm4
+
+# qhasm: uint32323232      t >>= 25
+# asm 1: psrld $25,<t=int6464#3
+# asm 2: psrld $25,<t=%xmm2
+psrld $25,%xmm2
+
+# qhasm:                   q ^= t
+# asm 1: pxor  <t=int6464#3,<q=int6464#4
+# asm 2: pxor  <t=%xmm2,<q=%xmm3
+pxor  %xmm2,%xmm3
+
+# qhasm: uint32323232      u <<= 7
+# asm 1: pslld $7,<u=int6464#5
+# asm 2: pslld $7,<u=%xmm4
+pslld $7,%xmm4
+
+# qhasm:                   q ^= u
+# asm 1: pxor  <u=int6464#5,<q=int6464#4
+# asm 2: pxor  <u=%xmm4,<q=%xmm3
+pxor  %xmm4,%xmm3
+
+# qhasm:                   z1_stack = q
+# asm 1: movdqa <q=int6464#4,>z1_stack=stack128#28
+# asm 2: movdqa <q=%xmm3,>z1_stack=464(%esp)
+movdqa %xmm3,464(%esp)
+
+# qhasm:                   t = p
+# asm 1: movdqa <p=int6464#1,>t=int6464#3
+# asm 2: movdqa <p=%xmm0,>t=%xmm2
+movdqa %xmm0,%xmm2
+
+# qhasm: uint32323232      t += q
+# asm 1: paddd <q=int6464#4,<t=int6464#3
+# asm 2: paddd <q=%xmm3,<t=%xmm2
+paddd %xmm3,%xmm2
+
+# qhasm:                   u = t
+# asm 1: movdqa <t=int6464#3,>u=int6464#5
+# asm 2: movdqa <t=%xmm2,>u=%xmm4
+movdqa %xmm2,%xmm4
+
+# qhasm: uint32323232      t >>= 23
+# asm 1: psrld $23,<t=int6464#3
+# asm 2: psrld $23,<t=%xmm2
+psrld $23,%xmm2
+
+# qhasm:                   r ^= t
+# asm 1: pxor  <t=int6464#3,<r=int6464#2
+# asm 2: pxor  <t=%xmm2,<r=%xmm1
+pxor  %xmm2,%xmm1
+
+# qhasm: uint32323232      u <<= 9
+# asm 1: pslld $9,<u=int6464#5
+# asm 2: pslld $9,<u=%xmm4
+pslld $9,%xmm4
+
+# qhasm:                   r ^= u
+# asm 1: pxor  <u=int6464#5,<r=int6464#2
+# asm 2: pxor  <u=%xmm4,<r=%xmm1
+pxor  %xmm4,%xmm1
+
+# qhasm:                   z2_stack = r
+# asm 1: movdqa <r=int6464#2,>z2_stack=stack128#31
+# asm 2: movdqa <r=%xmm1,>z2_stack=512(%esp)
+movdqa %xmm1,512(%esp)
+
+# qhasm: uint32323232      q += r
+# asm 1: paddd <r=int6464#2,<q=int6464#4
+# asm 2: paddd <r=%xmm1,<q=%xmm3
+paddd %xmm1,%xmm3
+
+# qhasm:                   u = q
+# asm 1: movdqa <q=int6464#4,>u=int6464#3
+# asm 2: movdqa <q=%xmm3,>u=%xmm2
+movdqa %xmm3,%xmm2
+
+# qhasm: uint32323232      q >>= 19
+# asm 1: psrld $19,<q=int6464#4
+# asm 2: psrld $19,<q=%xmm3
+psrld $19,%xmm3
+
+# qhasm:                   s ^= q
+# asm 1: pxor  <q=int6464#4,<s=int6464#7
+# asm 2: pxor  <q=%xmm3,<s=%xmm6
+pxor  %xmm3,%xmm6
+
+# qhasm: uint32323232      u <<= 13
+# asm 1: pslld $13,<u=int6464#3
+# asm 2: pslld $13,<u=%xmm2
+pslld $13,%xmm2
+
+# qhasm:                   s ^= u
+# asm 1: pxor  <u=int6464#3,<s=int6464#7
+# asm 2: pxor  <u=%xmm2,<s=%xmm6
+pxor  %xmm2,%xmm6
+
+# qhasm:                                   mt = z4_stack
+# asm 1: movdqa <z4_stack=stack128#33,>mt=int6464#3
+# asm 2: movdqa <z4_stack=544(%esp),>mt=%xmm2
+movdqa 544(%esp),%xmm2
+
+# qhasm:                                   mp = z5_stack
+# asm 1: movdqa <z5_stack=stack128#24,>mp=int6464#5
+# asm 2: movdqa <z5_stack=400(%esp),>mp=%xmm4
+movdqa 400(%esp),%xmm4
+
+# qhasm:                                   mq = z6_stack
+# asm 1: movdqa <z6_stack=stack128#23,>mq=int6464#4
+# asm 2: movdqa <z6_stack=384(%esp),>mq=%xmm3
+movdqa 384(%esp),%xmm3
+
+# qhasm:                                   mr = z7_stack
+# asm 1: movdqa <z7_stack=stack128#29,>mr=int6464#6
+# asm 2: movdqa <z7_stack=480(%esp),>mr=%xmm5
+movdqa 480(%esp),%xmm5
+
+# qhasm:                   z3_stack = s
+# asm 1: movdqa <s=int6464#7,>z3_stack=stack128#25
+# asm 2: movdqa <s=%xmm6,>z3_stack=416(%esp)
+movdqa %xmm6,416(%esp)
+
+# qhasm: uint32323232      r += s
+# asm 1: paddd <s=int6464#7,<r=int6464#2
+# asm 2: paddd <s=%xmm6,<r=%xmm1
+paddd %xmm6,%xmm1
+
+# qhasm:                   u = r
+# asm 1: movdqa <r=int6464#2,>u=int6464#7
+# asm 2: movdqa <r=%xmm1,>u=%xmm6
+movdqa %xmm1,%xmm6
+
+# qhasm: uint32323232      r >>= 14
+# asm 1: psrld $14,<r=int6464#2
+# asm 2: psrld $14,<r=%xmm1
+psrld $14,%xmm1
+
+# qhasm:                   p ^= r
+# asm 1: pxor  <r=int6464#2,<p=int6464#1
+# asm 2: pxor  <r=%xmm1,<p=%xmm0
+pxor  %xmm1,%xmm0
+
+# qhasm: uint32323232      u <<= 18
+# asm 1: pslld $18,<u=int6464#7
+# asm 2: pslld $18,<u=%xmm6
+pslld $18,%xmm6
+
+# qhasm:                   p ^= u
+# asm 1: pxor  <u=int6464#7,<p=int6464#1
+# asm 2: pxor  <u=%xmm6,<p=%xmm0
+pxor  %xmm6,%xmm0
+
+# qhasm:                   z0_stack = p
+# asm 1: movdqa <p=int6464#1,>z0_stack=stack128#33
+# asm 2: movdqa <p=%xmm0,>z0_stack=544(%esp)
+movdqa %xmm0,544(%esp)
+
+# qhasm:                                   assign xmm2 to mt
+
+# qhasm:                                   assign xmm3 to mq
+
+# qhasm:                                   assign xmm4 to mp
+
+# qhasm:                                   assign xmm5 to mr
+
+# qhasm:                                   ms = mt
+# asm 1: movdqa <mt=int6464#3,>ms=int6464#7
+# asm 2: movdqa <mt=%xmm2,>ms=%xmm6
+movdqa %xmm2,%xmm6
+
+# qhasm: uint32323232                      mt += mp
+# asm 1: paddd <mp=int6464#5,<mt=int6464#3
+# asm 2: paddd <mp=%xmm4,<mt=%xmm2
+paddd %xmm4,%xmm2
+
+# qhasm:                                   mu = mt
+# asm 1: movdqa <mt=int6464#3,>mu=int6464#1
+# asm 2: movdqa <mt=%xmm2,>mu=%xmm0
+movdqa %xmm2,%xmm0
+
+# qhasm: uint32323232                      mt >>= 25
+# asm 1: psrld $25,<mt=int6464#3
+# asm 2: psrld $25,<mt=%xmm2
+psrld $25,%xmm2
+
+# qhasm:                                   mq ^= mt
+# asm 1: pxor  <mt=int6464#3,<mq=int6464#4
+# asm 2: pxor  <mt=%xmm2,<mq=%xmm3
+pxor  %xmm2,%xmm3
+
+# qhasm: uint32323232                      mu <<= 7
+# asm 1: pslld $7,<mu=int6464#1
+# asm 2: pslld $7,<mu=%xmm0
+pslld $7,%xmm0
+
+# qhasm:                                   mq ^= mu
+# asm 1: pxor  <mu=int6464#1,<mq=int6464#4
+# asm 2: pxor  <mu=%xmm0,<mq=%xmm3
+pxor  %xmm0,%xmm3
+
+# qhasm:                                   z6_stack = mq
+# asm 1: movdqa <mq=int6464#4,>z6_stack=stack128#26
+# asm 2: movdqa <mq=%xmm3,>z6_stack=432(%esp)
+movdqa %xmm3,432(%esp)
+
+# qhasm:                                   mt = mp
+# asm 1: movdqa <mp=int6464#5,>mt=int6464#1
+# asm 2: movdqa <mp=%xmm4,>mt=%xmm0
+movdqa %xmm4,%xmm0
+
+# qhasm: uint32323232                      mt += mq
+# asm 1: paddd <mq=int6464#4,<mt=int6464#1
+# asm 2: paddd <mq=%xmm3,<mt=%xmm0
+paddd %xmm3,%xmm0
+
+# qhasm:                                   mu = mt
+# asm 1: movdqa <mt=int6464#1,>mu=int6464#2
+# asm 2: movdqa <mt=%xmm0,>mu=%xmm1
+movdqa %xmm0,%xmm1
+
+# qhasm: uint32323232                      mt >>= 23
+# asm 1: psrld $23,<mt=int6464#1
+# asm 2: psrld $23,<mt=%xmm0
+psrld $23,%xmm0
+
+# qhasm:                                   mr ^= mt
+# asm 1: pxor  <mt=int6464#1,<mr=int6464#6
+# asm 2: pxor  <mt=%xmm0,<mr=%xmm5
+pxor  %xmm0,%xmm5
+
+# qhasm: uint32323232                      mu <<= 9
+# asm 1: pslld $9,<mu=int6464#2
+# asm 2: pslld $9,<mu=%xmm1
+pslld $9,%xmm1
+
+# qhasm:                                   mr ^= mu
+# asm 1: pxor  <mu=int6464#2,<mr=int6464#6
+# asm 2: pxor  <mu=%xmm1,<mr=%xmm5
+pxor  %xmm1,%xmm5
+
+# qhasm:                                   z7_stack = mr
+# asm 1: movdqa <mr=int6464#6,>z7_stack=stack128#29
+# asm 2: movdqa <mr=%xmm5,>z7_stack=480(%esp)
+movdqa %xmm5,480(%esp)
+
+# qhasm: uint32323232                      mq += mr
+# asm 1: paddd <mr=int6464#6,<mq=int6464#4
+# asm 2: paddd <mr=%xmm5,<mq=%xmm3
+paddd %xmm5,%xmm3
+
+# qhasm:                                   mu = mq
+# asm 1: movdqa <mq=int6464#4,>mu=int6464#1
+# asm 2: movdqa <mq=%xmm3,>mu=%xmm0
+movdqa %xmm3,%xmm0
+
+# qhasm: uint32323232                      mq >>= 19
+# asm 1: psrld $19,<mq=int6464#4
+# asm 2: psrld $19,<mq=%xmm3
+psrld $19,%xmm3
+
+# qhasm:                                   ms ^= mq
+# asm 1: pxor  <mq=int6464#4,<ms=int6464#7
+# asm 2: pxor  <mq=%xmm3,<ms=%xmm6
+pxor  %xmm3,%xmm6
+
+# qhasm: uint32323232                      mu <<= 13
+# asm 1: pslld $13,<mu=int6464#1
+# asm 2: pslld $13,<mu=%xmm0
+pslld $13,%xmm0
+
+# qhasm:                                   ms ^= mu
+# asm 1: pxor  <mu=int6464#1,<ms=int6464#7
+# asm 2: pxor  <mu=%xmm0,<ms=%xmm6
+pxor  %xmm0,%xmm6
+
+# qhasm:                                                   t = z9_stack
+# asm 1: movdqa <z9_stack=stack128#32,>t=int6464#3
+# asm 2: movdqa <z9_stack=528(%esp),>t=%xmm2
+movdqa 528(%esp),%xmm2
+
+# qhasm:                                                   p = z10_stack
+# asm 1: movdqa <z10_stack=stack128#27,>p=int6464#1
+# asm 2: movdqa <z10_stack=448(%esp),>p=%xmm0
+movdqa 448(%esp),%xmm0
+
+# qhasm:                                                   q = z11_stack
+# asm 1: movdqa <z11_stack=stack128#21,>q=int6464#4
+# asm 2: movdqa <z11_stack=352(%esp),>q=%xmm3
+movdqa 352(%esp),%xmm3
+
+# qhasm:                                                   r = z8_stack
+# asm 1: movdqa <z8_stack=stack128#34,>r=int6464#2
+# asm 2: movdqa <z8_stack=560(%esp),>r=%xmm1
+movdqa 560(%esp),%xmm1
+
+# qhasm:                                   z4_stack = ms
+# asm 1: movdqa <ms=int6464#7,>z4_stack=stack128#34
+# asm 2: movdqa <ms=%xmm6,>z4_stack=560(%esp)
+movdqa %xmm6,560(%esp)
+
+# qhasm: uint32323232                      mr += ms
+# asm 1: paddd <ms=int6464#7,<mr=int6464#6
+# asm 2: paddd <ms=%xmm6,<mr=%xmm5
+paddd %xmm6,%xmm5
+
+# qhasm:                                   mu = mr
+# asm 1: movdqa <mr=int6464#6,>mu=int6464#7
+# asm 2: movdqa <mr=%xmm5,>mu=%xmm6
+movdqa %xmm5,%xmm6
+
+# qhasm: uint32323232                      mr >>= 14
+# asm 1: psrld $14,<mr=int6464#6
+# asm 2: psrld $14,<mr=%xmm5
+psrld $14,%xmm5
+
+# qhasm:                                   mp ^= mr
+# asm 1: pxor  <mr=int6464#6,<mp=int6464#5
+# asm 2: pxor  <mr=%xmm5,<mp=%xmm4
+pxor  %xmm5,%xmm4
+
+# qhasm: uint32323232                      mu <<= 18
+# asm 1: pslld $18,<mu=int6464#7
+# asm 2: pslld $18,<mu=%xmm6
+pslld $18,%xmm6
+
+# qhasm:                                   mp ^= mu
+# asm 1: pxor  <mu=int6464#7,<mp=int6464#5
+# asm 2: pxor  <mu=%xmm6,<mp=%xmm4
+pxor  %xmm6,%xmm4
+
+# qhasm:                                   z5_stack = mp
+# asm 1: movdqa <mp=int6464#5,>z5_stack=stack128#21
+# asm 2: movdqa <mp=%xmm4,>z5_stack=352(%esp)
+movdqa %xmm4,352(%esp)
+
+# qhasm:                                                   assign xmm0 to p
+
+# qhasm:                                                   assign xmm1 to r
+
+# qhasm:                                                   assign xmm2 to t
+
+# qhasm:                                                   assign xmm3 to q
+
+# qhasm:                                                   s = t
+# asm 1: movdqa <t=int6464#3,>s=int6464#7
+# asm 2: movdqa <t=%xmm2,>s=%xmm6
+movdqa %xmm2,%xmm6
+
+# qhasm: uint32323232                                      t += p
+# asm 1: paddd <p=int6464#1,<t=int6464#3
+# asm 2: paddd <p=%xmm0,<t=%xmm2
+paddd %xmm0,%xmm2
+
+# qhasm:                                                   u = t
+# asm 1: movdqa <t=int6464#3,>u=int6464#5
+# asm 2: movdqa <t=%xmm2,>u=%xmm4
+movdqa %xmm2,%xmm4
+
+# qhasm: uint32323232                                      t >>= 25
+# asm 1: psrld $25,<t=int6464#3
+# asm 2: psrld $25,<t=%xmm2
+psrld $25,%xmm2
+
+# qhasm:                                                   q ^= t
+# asm 1: pxor  <t=int6464#3,<q=int6464#4
+# asm 2: pxor  <t=%xmm2,<q=%xmm3
+pxor  %xmm2,%xmm3
+
+# qhasm: uint32323232                                      u <<= 7
+# asm 1: pslld $7,<u=int6464#5
+# asm 2: pslld $7,<u=%xmm4
+pslld $7,%xmm4
+
+# qhasm:                                                   q ^= u
+# asm 1: pxor  <u=int6464#5,<q=int6464#4
+# asm 2: pxor  <u=%xmm4,<q=%xmm3
+pxor  %xmm4,%xmm3
+
+# qhasm:                                                   z11_stack = q
+# asm 1: movdqa <q=int6464#4,>z11_stack=stack128#27
+# asm 2: movdqa <q=%xmm3,>z11_stack=448(%esp)
+movdqa %xmm3,448(%esp)
+
+# qhasm:                                                   t = p
+# asm 1: movdqa <p=int6464#1,>t=int6464#3
+# asm 2: movdqa <p=%xmm0,>t=%xmm2
+movdqa %xmm0,%xmm2
+
+# qhasm: uint32323232                                      t += q
+# asm 1: paddd <q=int6464#4,<t=int6464#3
+# asm 2: paddd <q=%xmm3,<t=%xmm2
+paddd %xmm3,%xmm2
+
+# qhasm:                                                   u = t
+# asm 1: movdqa <t=int6464#3,>u=int6464#5
+# asm 2: movdqa <t=%xmm2,>u=%xmm4
+movdqa %xmm2,%xmm4
+
+# qhasm: uint32323232                                      t >>= 23
+# asm 1: psrld $23,<t=int6464#3
+# asm 2: psrld $23,<t=%xmm2
+psrld $23,%xmm2
+
+# qhasm:                                                   r ^= t
+# asm 1: pxor  <t=int6464#3,<r=int6464#2
+# asm 2: pxor  <t=%xmm2,<r=%xmm1
+pxor  %xmm2,%xmm1
+
+# qhasm: uint32323232                                      u <<= 9
+# asm 1: pslld $9,<u=int6464#5
+# asm 2: pslld $9,<u=%xmm4
+pslld $9,%xmm4
+
+# qhasm:                                                   r ^= u
+# asm 1: pxor  <u=int6464#5,<r=int6464#2
+# asm 2: pxor  <u=%xmm4,<r=%xmm1
+pxor  %xmm4,%xmm1
+
+# qhasm:                                                   z8_stack = r
+# asm 1: movdqa <r=int6464#2,>z8_stack=stack128#37
+# asm 2: movdqa <r=%xmm1,>z8_stack=608(%esp)
+movdqa %xmm1,608(%esp)
+
+# qhasm: uint32323232                                      q += r
+# asm 1: paddd <r=int6464#2,<q=int6464#4
+# asm 2: paddd <r=%xmm1,<q=%xmm3
+paddd %xmm1,%xmm3
+
+# qhasm:                                                   u = q
+# asm 1: movdqa <q=int6464#4,>u=int6464#3
+# asm 2: movdqa <q=%xmm3,>u=%xmm2
+movdqa %xmm3,%xmm2
+
+# qhasm: uint32323232                                      q >>= 19
+# asm 1: psrld $19,<q=int6464#4
+# asm 2: psrld $19,<q=%xmm3
+psrld $19,%xmm3
+
+# qhasm:                                                   s ^= q
+# asm 1: pxor  <q=int6464#4,<s=int6464#7
+# asm 2: pxor  <q=%xmm3,<s=%xmm6
+pxor  %xmm3,%xmm6
+
+# qhasm: uint32323232                                      u <<= 13
+# asm 1: pslld $13,<u=int6464#3
+# asm 2: pslld $13,<u=%xmm2
+pslld $13,%xmm2
+
+# qhasm:                                                   s ^= u
+# asm 1: pxor  <u=int6464#3,<s=int6464#7
+# asm 2: pxor  <u=%xmm2,<s=%xmm6
+pxor  %xmm2,%xmm6
+
+# qhasm:                                                                   mt = z14_stack
+# asm 1: movdqa <z14_stack=stack128#36,>mt=int6464#3
+# asm 2: movdqa <z14_stack=592(%esp),>mt=%xmm2
+movdqa 592(%esp),%xmm2
+
+# qhasm:                                                                   mp = z15_stack
+# asm 1: movdqa <z15_stack=stack128#22,>mp=int6464#5
+# asm 2: movdqa <z15_stack=368(%esp),>mp=%xmm4
+movdqa 368(%esp),%xmm4
+
+# qhasm:                                                                   mq = z12_stack
+# asm 1: movdqa <z12_stack=stack128#30,>mq=int6464#4
+# asm 2: movdqa <z12_stack=496(%esp),>mq=%xmm3
+movdqa 496(%esp),%xmm3
+
+# qhasm:                                                                   mr = z13_stack
+# asm 1: movdqa <z13_stack=stack128#35,>mr=int6464#6
+# asm 2: movdqa <z13_stack=576(%esp),>mr=%xmm5
+movdqa 576(%esp),%xmm5
+
+# qhasm:                                                   z9_stack = s
+# asm 1: movdqa <s=int6464#7,>z9_stack=stack128#32
+# asm 2: movdqa <s=%xmm6,>z9_stack=528(%esp)
+movdqa %xmm6,528(%esp)
+
+# qhasm: uint32323232                                      r += s
+# asm 1: paddd <s=int6464#7,<r=int6464#2
+# asm 2: paddd <s=%xmm6,<r=%xmm1
+paddd %xmm6,%xmm1
+
+# qhasm:                                                   u = r
+# asm 1: movdqa <r=int6464#2,>u=int6464#7
+# asm 2: movdqa <r=%xmm1,>u=%xmm6
+movdqa %xmm1,%xmm6
+
+# qhasm: uint32323232                                      r >>= 14
+# asm 1: psrld $14,<r=int6464#2
+# asm 2: psrld $14,<r=%xmm1
+psrld $14,%xmm1
+
+# qhasm:                                                   p ^= r
+# asm 1: pxor  <r=int6464#2,<p=int6464#1
+# asm 2: pxor  <r=%xmm1,<p=%xmm0
+pxor  %xmm1,%xmm0
+
+# qhasm: uint32323232                                      u <<= 18
+# asm 1: pslld $18,<u=int6464#7
+# asm 2: pslld $18,<u=%xmm6
+pslld $18,%xmm6
+
+# qhasm:                                                   p ^= u
+# asm 1: pxor  <u=int6464#7,<p=int6464#1
+# asm 2: pxor  <u=%xmm6,<p=%xmm0
+pxor  %xmm6,%xmm0
+
+# qhasm:                                                   z10_stack = p
+# asm 1: movdqa <p=int6464#1,>z10_stack=stack128#22
+# asm 2: movdqa <p=%xmm0,>z10_stack=368(%esp)
+movdqa %xmm0,368(%esp)
+
+# qhasm:                                                                   assign xmm2 to mt
+
+# qhasm:                                                                   assign xmm3 to mq
+
+# qhasm:                                                                   assign xmm4 to mp
+
+# qhasm:                                                                   assign xmm5 to mr
+
+# qhasm:                                                                   ms = mt
+# asm 1: movdqa <mt=int6464#3,>ms=int6464#7
+# asm 2: movdqa <mt=%xmm2,>ms=%xmm6
+movdqa %xmm2,%xmm6
+
+# qhasm: uint32323232                                                      mt += mp
+# asm 1: paddd <mp=int6464#5,<mt=int6464#3
+# asm 2: paddd <mp=%xmm4,<mt=%xmm2
+paddd %xmm4,%xmm2
+
+# qhasm:                                                                   mu = mt
+# asm 1: movdqa <mt=int6464#3,>mu=int6464#1
+# asm 2: movdqa <mt=%xmm2,>mu=%xmm0
+movdqa %xmm2,%xmm0
+
+# qhasm: uint32323232                                                      mt >>= 25
+# asm 1: psrld $25,<mt=int6464#3
+# asm 2: psrld $25,<mt=%xmm2
+psrld $25,%xmm2
+
+# qhasm:                                                                   mq ^= mt
+# asm 1: pxor  <mt=int6464#3,<mq=int6464#4
+# asm 2: pxor  <mt=%xmm2,<mq=%xmm3
+pxor  %xmm2,%xmm3
+
+# qhasm: uint32323232                                                      mu <<= 7
+# asm 1: pslld $7,<mu=int6464#1
+# asm 2: pslld $7,<mu=%xmm0
+pslld $7,%xmm0
+
+# qhasm:                                                                   mq ^= mu
+# asm 1: pxor  <mu=int6464#1,<mq=int6464#4
+# asm 2: pxor  <mu=%xmm0,<mq=%xmm3
+pxor  %xmm0,%xmm3
+
+# qhasm:                                                                   z12_stack = mq
+# asm 1: movdqa <mq=int6464#4,>z12_stack=stack128#35
+# asm 2: movdqa <mq=%xmm3,>z12_stack=576(%esp)
+movdqa %xmm3,576(%esp)
+
+# qhasm:                                                                   mt = mp
+# asm 1: movdqa <mp=int6464#5,>mt=int6464#1
+# asm 2: movdqa <mp=%xmm4,>mt=%xmm0
+movdqa %xmm4,%xmm0
+
+# qhasm: uint32323232                                                      mt += mq
+# asm 1: paddd <mq=int6464#4,<mt=int6464#1
+# asm 2: paddd <mq=%xmm3,<mt=%xmm0
+paddd %xmm3,%xmm0
+
+# qhasm:                                                                   mu = mt
+# asm 1: movdqa <mt=int6464#1,>mu=int6464#2
+# asm 2: movdqa <mt=%xmm0,>mu=%xmm1
+movdqa %xmm0,%xmm1
+
+# qhasm: uint32323232                                                      mt >>= 23
+# asm 1: psrld $23,<mt=int6464#1
+# asm 2: psrld $23,<mt=%xmm0
+psrld $23,%xmm0
+
+# qhasm:                                                                   mr ^= mt
+# asm 1: pxor  <mt=int6464#1,<mr=int6464#6
+# asm 2: pxor  <mt=%xmm0,<mr=%xmm5
+pxor  %xmm0,%xmm5
+
+# qhasm: uint32323232                                                      mu <<= 9
+# asm 1: pslld $9,<mu=int6464#2
+# asm 2: pslld $9,<mu=%xmm1
+pslld $9,%xmm1
+
+# qhasm:                                                                   mr ^= mu
+# asm 1: pxor  <mu=int6464#2,<mr=int6464#6
+# asm 2: pxor  <mu=%xmm1,<mr=%xmm5
+pxor  %xmm1,%xmm5
+
+# qhasm:                                                                   z13_stack = mr
+# asm 1: movdqa <mr=int6464#6,>z13_stack=stack128#30
+# asm 2: movdqa <mr=%xmm5,>z13_stack=496(%esp)
+movdqa %xmm5,496(%esp)
+
+# qhasm: uint32323232                                                      mq += mr
+# asm 1: paddd <mr=int6464#6,<mq=int6464#4
+# asm 2: paddd <mr=%xmm5,<mq=%xmm3
+paddd %xmm5,%xmm3
+
+# qhasm:                                                                   mu = mq
+# asm 1: movdqa <mq=int6464#4,>mu=int6464#1
+# asm 2: movdqa <mq=%xmm3,>mu=%xmm0
+movdqa %xmm3,%xmm0
+
+# qhasm: uint32323232                                                      mq >>= 19
+# asm 1: psrld $19,<mq=int6464#4
+# asm 2: psrld $19,<mq=%xmm3
+psrld $19,%xmm3
+
+# qhasm:                                                                   ms ^= mq
+# asm 1: pxor  <mq=int6464#4,<ms=int6464#7
+# asm 2: pxor  <mq=%xmm3,<ms=%xmm6
+pxor  %xmm3,%xmm6
+
+# qhasm: uint32323232                                                      mu <<= 13
+# asm 1: pslld $13,<mu=int6464#1
+# asm 2: pslld $13,<mu=%xmm0
+pslld $13,%xmm0
+
+# qhasm:                                                                   ms ^= mu
+# asm 1: pxor  <mu=int6464#1,<ms=int6464#7
+# asm 2: pxor  <mu=%xmm0,<ms=%xmm6
+pxor  %xmm0,%xmm6
+
+# qhasm:                   t = z12_stack
+# asm 1: movdqa <z12_stack=stack128#35,>t=int6464#3
+# asm 2: movdqa <z12_stack=576(%esp),>t=%xmm2
+movdqa 576(%esp),%xmm2
+
+# qhasm:                   p = z0_stack
+# asm 1: movdqa <z0_stack=stack128#33,>p=int6464#1
+# asm 2: movdqa <z0_stack=544(%esp),>p=%xmm0
+movdqa 544(%esp),%xmm0
+
+# qhasm:                   q = z4_stack
+# asm 1: movdqa <z4_stack=stack128#34,>q=int6464#4
+# asm 2: movdqa <z4_stack=560(%esp),>q=%xmm3
+movdqa 560(%esp),%xmm3
+
+# qhasm:                   r = z8_stack
+# asm 1: movdqa <z8_stack=stack128#37,>r=int6464#2
+# asm 2: movdqa <z8_stack=608(%esp),>r=%xmm1
+movdqa 608(%esp),%xmm1
+
+# qhasm:                                                                   z14_stack = ms
+# asm 1: movdqa <ms=int6464#7,>z14_stack=stack128#24
+# asm 2: movdqa <ms=%xmm6,>z14_stack=400(%esp)
+movdqa %xmm6,400(%esp)
+
+# qhasm: uint32323232                                                      mr += ms
+# asm 1: paddd <ms=int6464#7,<mr=int6464#6
+# asm 2: paddd <ms=%xmm6,<mr=%xmm5
+paddd %xmm6,%xmm5
+
+# qhasm:                                                                   mu = mr
+# asm 1: movdqa <mr=int6464#6,>mu=int6464#7
+# asm 2: movdqa <mr=%xmm5,>mu=%xmm6
+movdqa %xmm5,%xmm6
+
+# qhasm: uint32323232                                                      mr >>= 14
+# asm 1: psrld $14,<mr=int6464#6
+# asm 2: psrld $14,<mr=%xmm5
+psrld $14,%xmm5
+
+# qhasm:                                                                   mp ^= mr
+# asm 1: pxor  <mr=int6464#6,<mp=int6464#5
+# asm 2: pxor  <mr=%xmm5,<mp=%xmm4
+pxor  %xmm5,%xmm4
+
+# qhasm: uint32323232                                                      mu <<= 18
+# asm 1: pslld $18,<mu=int6464#7
+# asm 2: pslld $18,<mu=%xmm6
+pslld $18,%xmm6
+
+# qhasm:                                                                   mp ^= mu
+# asm 1: pxor  <mu=int6464#7,<mp=int6464#5
+# asm 2: pxor  <mu=%xmm6,<mp=%xmm4
+pxor  %xmm6,%xmm4
+
+# qhasm:                                                                   z15_stack = mp
+# asm 1: movdqa <mp=int6464#5,>z15_stack=stack128#23
+# asm 2: movdqa <mp=%xmm4,>z15_stack=384(%esp)
+movdqa %xmm4,384(%esp)
+
+# qhasm:                   unsigned>? i -= 2
+# asm 1: sub  $2,<i=int32#1
+# asm 2: sub  $2,<i=%eax
+sub  $2,%eax
+# comment:fp stack unchanged by jump
+
+# qhasm: goto mainloop1 if unsigned>
+ja ._mainloop1
+
+# qhasm:   out = out_stack
+# asm 1: movl <out_stack=stack32#6,>out=int32#6
+# asm 2: movl <out_stack=20(%esp),>out=%edi
+movl 20(%esp),%edi
+
+# qhasm:   z0 = z0_stack
+# asm 1: movdqa <z0_stack=stack128#33,>z0=int6464#1
+# asm 2: movdqa <z0_stack=544(%esp),>z0=%xmm0
+movdqa 544(%esp),%xmm0
+
+# qhasm:   z1 = z1_stack
+# asm 1: movdqa <z1_stack=stack128#28,>z1=int6464#2
+# asm 2: movdqa <z1_stack=464(%esp),>z1=%xmm1
+movdqa 464(%esp),%xmm1
+
+# qhasm:   z2 = z2_stack
+# asm 1: movdqa <z2_stack=stack128#31,>z2=int6464#3
+# asm 2: movdqa <z2_stack=512(%esp),>z2=%xmm2
+movdqa 512(%esp),%xmm2
+
+# qhasm:   z3 = z3_stack
+# asm 1: movdqa <z3_stack=stack128#25,>z3=int6464#4
+# asm 2: movdqa <z3_stack=416(%esp),>z3=%xmm3
+movdqa 416(%esp),%xmm3
+
+# qhasm:   uint32323232 z0 += orig0
+# asm 1: paddd <orig0=stack128#8,<z0=int6464#1
+# asm 2: paddd <orig0=144(%esp),<z0=%xmm0
+paddd 144(%esp),%xmm0
+
+# qhasm:   uint32323232 z1 += orig1
+# asm 1: paddd <orig1=stack128#12,<z1=int6464#2
+# asm 2: paddd <orig1=208(%esp),<z1=%xmm1
+paddd 208(%esp),%xmm1
+
+# qhasm:   uint32323232 z2 += orig2
+# asm 1: paddd <orig2=stack128#15,<z2=int6464#3
+# asm 2: paddd <orig2=256(%esp),<z2=%xmm2
+paddd 256(%esp),%xmm2
+
+# qhasm:   uint32323232 z3 += orig3
+# asm 1: paddd <orig3=stack128#18,<z3=int6464#4
+# asm 2: paddd <orig3=304(%esp),<z3=%xmm3
+paddd 304(%esp),%xmm3
+
+# qhasm:   in0 = z0
+# asm 1: movd   <z0=int6464#1,>in0=int32#1
+# asm 2: movd   <z0=%xmm0,>in0=%eax
+movd   %xmm0,%eax
+
+# qhasm:   in1 = z1
+# asm 1: movd   <z1=int6464#2,>in1=int32#2
+# asm 2: movd   <z1=%xmm1,>in1=%ecx
+movd   %xmm1,%ecx
+
+# qhasm:   in2 = z2
+# asm 1: movd   <z2=int6464#3,>in2=int32#3
+# asm 2: movd   <z2=%xmm2,>in2=%edx
+movd   %xmm2,%edx
+
+# qhasm:   in3 = z3
+# asm 1: movd   <z3=int6464#4,>in3=int32#4
+# asm 2: movd   <z3=%xmm3,>in3=%ebx
+movd   %xmm3,%ebx
+
+# qhasm:   z0 <<<= 96
+# asm 1: pshufd $0x39,<z0=int6464#1,<z0=int6464#1
+# asm 2: pshufd $0x39,<z0=%xmm0,<z0=%xmm0
+pshufd $0x39,%xmm0,%xmm0
+
+# qhasm:   z1 <<<= 96
+# asm 1: pshufd $0x39,<z1=int6464#2,<z1=int6464#2
+# asm 2: pshufd $0x39,<z1=%xmm1,<z1=%xmm1
+pshufd $0x39,%xmm1,%xmm1
+
+# qhasm:   z2 <<<= 96
+# asm 1: pshufd $0x39,<z2=int6464#3,<z2=int6464#3
+# asm 2: pshufd $0x39,<z2=%xmm2,<z2=%xmm2
+pshufd $0x39,%xmm2,%xmm2
+
+# qhasm:   z3 <<<= 96
+# asm 1: pshufd $0x39,<z3=int6464#4,<z3=int6464#4
+# asm 2: pshufd $0x39,<z3=%xmm3,<z3=%xmm3
+pshufd $0x39,%xmm3,%xmm3
+
+# qhasm:   in0 ^= *(uint32 *) (m + 0)
+# asm 1: xorl 0(<m=int32#5),<in0=int32#1
+# asm 2: xorl 0(<m=%esi),<in0=%eax
+xorl 0(%esi),%eax
+
+# qhasm:   in1 ^= *(uint32 *) (m + 4)
+# asm 1: xorl 4(<m=int32#5),<in1=int32#2
+# asm 2: xorl 4(<m=%esi),<in1=%ecx
+xorl 4(%esi),%ecx
+
+# qhasm:   in2 ^= *(uint32 *) (m + 8)
+# asm 1: xorl 8(<m=int32#5),<in2=int32#3
+# asm 2: xorl 8(<m=%esi),<in2=%edx
+xorl 8(%esi),%edx
+
+# qhasm:   in3 ^= *(uint32 *) (m + 12)
+# asm 1: xorl 12(<m=int32#5),<in3=int32#4
+# asm 2: xorl 12(<m=%esi),<in3=%ebx
+xorl 12(%esi),%ebx
+
+# qhasm:   *(uint32 *) (out + 0) = in0
+# asm 1: movl <in0=int32#1,0(<out=int32#6)
+# asm 2: movl <in0=%eax,0(<out=%edi)
+movl %eax,0(%edi)
+
+# qhasm:   *(uint32 *) (out + 4) = in1
+# asm 1: movl <in1=int32#2,4(<out=int32#6)
+# asm 2: movl <in1=%ecx,4(<out=%edi)
+movl %ecx,4(%edi)
+
+# qhasm:   *(uint32 *) (out + 8) = in2
+# asm 1: movl <in2=int32#3,8(<out=int32#6)
+# asm 2: movl <in2=%edx,8(<out=%edi)
+movl %edx,8(%edi)
+
+# qhasm:   *(uint32 *) (out + 12) = in3
+# asm 1: movl <in3=int32#4,12(<out=int32#6)
+# asm 2: movl <in3=%ebx,12(<out=%edi)
+movl %ebx,12(%edi)
+
+# qhasm:   in0 = z0
+# asm 1: movd   <z0=int6464#1,>in0=int32#1
+# asm 2: movd   <z0=%xmm0,>in0=%eax
+movd   %xmm0,%eax
+
+# qhasm:   in1 = z1
+# asm 1: movd   <z1=int6464#2,>in1=int32#2
+# asm 2: movd   <z1=%xmm1,>in1=%ecx
+movd   %xmm1,%ecx
+
+# qhasm:   in2 = z2
+# asm 1: movd   <z2=int6464#3,>in2=int32#3
+# asm 2: movd   <z2=%xmm2,>in2=%edx
+movd   %xmm2,%edx
+
+# qhasm:   in3 = z3
+# asm 1: movd   <z3=int6464#4,>in3=int32#4
+# asm 2: movd   <z3=%xmm3,>in3=%ebx
+movd   %xmm3,%ebx
+
+# qhasm:   z0 <<<= 96
+# asm 1: pshufd $0x39,<z0=int6464#1,<z0=int6464#1
+# asm 2: pshufd $0x39,<z0=%xmm0,<z0=%xmm0
+pshufd $0x39,%xmm0,%xmm0
+
+# qhasm:   z1 <<<= 96
+# asm 1: pshufd $0x39,<z1=int6464#2,<z1=int6464#2
+# asm 2: pshufd $0x39,<z1=%xmm1,<z1=%xmm1
+pshufd $0x39,%xmm1,%xmm1
+
+# qhasm:   z2 <<<= 96
+# asm 1: pshufd $0x39,<z2=int6464#3,<z2=int6464#3
+# asm 2: pshufd $0x39,<z2=%xmm2,<z2=%xmm2
+pshufd $0x39,%xmm2,%xmm2
+
+# qhasm:   z3 <<<= 96
+# asm 1: pshufd $0x39,<z3=int6464#4,<z3=int6464#4
+# asm 2: pshufd $0x39,<z3=%xmm3,<z3=%xmm3
+pshufd $0x39,%xmm3,%xmm3
+
+# qhasm:   in0 ^= *(uint32 *) (m + 64)
+# asm 1: xorl 64(<m=int32#5),<in0=int32#1
+# asm 2: xorl 64(<m=%esi),<in0=%eax
+xorl 64(%esi),%eax
+
+# qhasm:   in1 ^= *(uint32 *) (m + 68)
+# asm 1: xorl 68(<m=int32#5),<in1=int32#2
+# asm 2: xorl 68(<m=%esi),<in1=%ecx
+xorl 68(%esi),%ecx
+
+# qhasm:   in2 ^= *(uint32 *) (m + 72)
+# asm 1: xorl 72(<m=int32#5),<in2=int32#3
+# asm 2: xorl 72(<m=%esi),<in2=%edx
+xorl 72(%esi),%edx
+
+# qhasm:   in3 ^= *(uint32 *) (m + 76)
+# asm 1: xorl 76(<m=int32#5),<in3=int32#4
+# asm 2: xorl 76(<m=%esi),<in3=%ebx
+xorl 76(%esi),%ebx
+
+# qhasm:   *(uint32 *) (out + 64) = in0
+# asm 1: movl <in0=int32#1,64(<out=int32#6)
+# asm 2: movl <in0=%eax,64(<out=%edi)
+movl %eax,64(%edi)
+
+# qhasm:   *(uint32 *) (out + 68) = in1
+# asm 1: movl <in1=int32#2,68(<out=int32#6)
+# asm 2: movl <in1=%ecx,68(<out=%edi)
+movl %ecx,68(%edi)
+
+# qhasm:   *(uint32 *) (out + 72) = in2
+# asm 1: movl <in2=int32#3,72(<out=int32#6)
+# asm 2: movl <in2=%edx,72(<out=%edi)
+movl %edx,72(%edi)
+
+# qhasm:   *(uint32 *) (out + 76) = in3
+# asm 1: movl <in3=int32#4,76(<out=int32#6)
+# asm 2: movl <in3=%ebx,76(<out=%edi)
+movl %ebx,76(%edi)
+
+# qhasm:   in0 = z0
+# asm 1: movd   <z0=int6464#1,>in0=int32#1
+# asm 2: movd   <z0=%xmm0,>in0=%eax
+movd   %xmm0,%eax
+
+# qhasm:   in1 = z1
+# asm 1: movd   <z1=int6464#2,>in1=int32#2
+# asm 2: movd   <z1=%xmm1,>in1=%ecx
+movd   %xmm1,%ecx
+
+# qhasm:   in2 = z2
+# asm 1: movd   <z2=int6464#3,>in2=int32#3
+# asm 2: movd   <z2=%xmm2,>in2=%edx
+movd   %xmm2,%edx
+
+# qhasm:   in3 = z3
+# asm 1: movd   <z3=int6464#4,>in3=int32#4
+# asm 2: movd   <z3=%xmm3,>in3=%ebx
+movd   %xmm3,%ebx
+
+# qhasm:   z0 <<<= 96
+# asm 1: pshufd $0x39,<z0=int6464#1,<z0=int6464#1
+# asm 2: pshufd $0x39,<z0=%xmm0,<z0=%xmm0
+pshufd $0x39,%xmm0,%xmm0
+
+# qhasm:   z1 <<<= 96
+# asm 1: pshufd $0x39,<z1=int6464#2,<z1=int6464#2
+# asm 2: pshufd $0x39,<z1=%xmm1,<z1=%xmm1
+pshufd $0x39,%xmm1,%xmm1
+
+# qhasm:   z2 <<<= 96
+# asm 1: pshufd $0x39,<z2=int6464#3,<z2=int6464#3
+# asm 2: pshufd $0x39,<z2=%xmm2,<z2=%xmm2
+pshufd $0x39,%xmm2,%xmm2
+
+# qhasm:   z3 <<<= 96
+# asm 1: pshufd $0x39,<z3=int6464#4,<z3=int6464#4
+# asm 2: pshufd $0x39,<z3=%xmm3,<z3=%xmm3
+pshufd $0x39,%xmm3,%xmm3
+
+# qhasm:   in0 ^= *(uint32 *) (m + 128)
+# asm 1: xorl 128(<m=int32#5),<in0=int32#1
+# asm 2: xorl 128(<m=%esi),<in0=%eax
+xorl 128(%esi),%eax
+
+# qhasm:   in1 ^= *(uint32 *) (m + 132)
+# asm 1: xorl 132(<m=int32#5),<in1=int32#2
+# asm 2: xorl 132(<m=%esi),<in1=%ecx
+xorl 132(%esi),%ecx
+
+# qhasm:   in2 ^= *(uint32 *) (m + 136)
+# asm 1: xorl 136(<m=int32#5),<in2=int32#3
+# asm 2: xorl 136(<m=%esi),<in2=%edx
+xorl 136(%esi),%edx
+
+# qhasm:   in3 ^= *(uint32 *) (m + 140)
+# asm 1: xorl 140(<m=int32#5),<in3=int32#4
+# asm 2: xorl 140(<m=%esi),<in3=%ebx
+xorl 140(%esi),%ebx
+
+# qhasm:   *(uint32 *) (out + 128) = in0
+# asm 1: movl <in0=int32#1,128(<out=int32#6)
+# asm 2: movl <in0=%eax,128(<out=%edi)
+movl %eax,128(%edi)
+
+# qhasm:   *(uint32 *) (out + 132) = in1
+# asm 1: movl <in1=int32#2,132(<out=int32#6)
+# asm 2: movl <in1=%ecx,132(<out=%edi)
+movl %ecx,132(%edi)
+
+# qhasm:   *(uint32 *) (out + 136) = in2
+# asm 1: movl <in2=int32#3,136(<out=int32#6)
+# asm 2: movl <in2=%edx,136(<out=%edi)
+movl %edx,136(%edi)
+
+# qhasm:   *(uint32 *) (out + 140) = in3
+# asm 1: movl <in3=int32#4,140(<out=int32#6)
+# asm 2: movl <in3=%ebx,140(<out=%edi)
+movl %ebx,140(%edi)
+
+# qhasm:   in0 = z0
+# asm 1: movd   <z0=int6464#1,>in0=int32#1
+# asm 2: movd   <z0=%xmm0,>in0=%eax
+movd   %xmm0,%eax
+
+# qhasm:   in1 = z1
+# asm 1: movd   <z1=int6464#2,>in1=int32#2
+# asm 2: movd   <z1=%xmm1,>in1=%ecx
+movd   %xmm1,%ecx
+
+# qhasm:   in2 = z2
+# asm 1: movd   <z2=int6464#3,>in2=int32#3
+# asm 2: movd   <z2=%xmm2,>in2=%edx
+movd   %xmm2,%edx
+
+# qhasm:   in3 = z3
+# asm 1: movd   <z3=int6464#4,>in3=int32#4
+# asm 2: movd   <z3=%xmm3,>in3=%ebx
+movd   %xmm3,%ebx
+
+# qhasm:   in0 ^= *(uint32 *) (m + 192)
+# asm 1: xorl 192(<m=int32#5),<in0=int32#1
+# asm 2: xorl 192(<m=%esi),<in0=%eax
+xorl 192(%esi),%eax
+
+# qhasm:   in1 ^= *(uint32 *) (m + 196)
+# asm 1: xorl 196(<m=int32#5),<in1=int32#2
+# asm 2: xorl 196(<m=%esi),<in1=%ecx
+xorl 196(%esi),%ecx
+
+# qhasm:   in2 ^= *(uint32 *) (m + 200)
+# asm 1: xorl 200(<m=int32#5),<in2=int32#3
+# asm 2: xorl 200(<m=%esi),<in2=%edx
+xorl 200(%esi),%edx
+
+# qhasm:   in3 ^= *(uint32 *) (m + 204)
+# asm 1: xorl 204(<m=int32#5),<in3=int32#4
+# asm 2: xorl 204(<m=%esi),<in3=%ebx
+xorl 204(%esi),%ebx
+
+# qhasm:   *(uint32 *) (out + 192) = in0
+# asm 1: movl <in0=int32#1,192(<out=int32#6)
+# asm 2: movl <in0=%eax,192(<out=%edi)
+movl %eax,192(%edi)
+
+# qhasm:   *(uint32 *) (out + 196) = in1
+# asm 1: movl <in1=int32#2,196(<out=int32#6)
+# asm 2: movl <in1=%ecx,196(<out=%edi)
+movl %ecx,196(%edi)
+
+# qhasm:   *(uint32 *) (out + 200) = in2
+# asm 1: movl <in2=int32#3,200(<out=int32#6)
+# asm 2: movl <in2=%edx,200(<out=%edi)
+movl %edx,200(%edi)
+
+# qhasm:   *(uint32 *) (out + 204) = in3
+# asm 1: movl <in3=int32#4,204(<out=int32#6)
+# asm 2: movl <in3=%ebx,204(<out=%edi)
+movl %ebx,204(%edi)
+
+# qhasm:   z4 = z4_stack
+# asm 1: movdqa <z4_stack=stack128#34,>z4=int6464#1
+# asm 2: movdqa <z4_stack=560(%esp),>z4=%xmm0
+movdqa 560(%esp),%xmm0
+
+# qhasm:   z5 = z5_stack
+# asm 1: movdqa <z5_stack=stack128#21,>z5=int6464#2
+# asm 2: movdqa <z5_stack=352(%esp),>z5=%xmm1
+movdqa 352(%esp),%xmm1
+
+# qhasm:   z6 = z6_stack
+# asm 1: movdqa <z6_stack=stack128#26,>z6=int6464#3
+# asm 2: movdqa <z6_stack=432(%esp),>z6=%xmm2
+movdqa 432(%esp),%xmm2
+
+# qhasm:   z7 = z7_stack
+# asm 1: movdqa <z7_stack=stack128#29,>z7=int6464#4
+# asm 2: movdqa <z7_stack=480(%esp),>z7=%xmm3
+movdqa 480(%esp),%xmm3
+
+# qhasm:   uint32323232 z4 += orig4
+# asm 1: paddd <orig4=stack128#16,<z4=int6464#1
+# asm 2: paddd <orig4=272(%esp),<z4=%xmm0
+paddd 272(%esp),%xmm0
+
+# qhasm:   uint32323232 z5 += orig5
+# asm 1: paddd <orig5=stack128#5,<z5=int6464#2
+# asm 2: paddd <orig5=96(%esp),<z5=%xmm1
+paddd 96(%esp),%xmm1
+
+# qhasm:   uint32323232 z6 += orig6
+# asm 1: paddd <orig6=stack128#9,<z6=int6464#3
+# asm 2: paddd <orig6=160(%esp),<z6=%xmm2
+paddd 160(%esp),%xmm2
+
+# qhasm:   uint32323232 z7 += orig7
+# asm 1: paddd <orig7=stack128#13,<z7=int6464#4
+# asm 2: paddd <orig7=224(%esp),<z7=%xmm3
+paddd 224(%esp),%xmm3
+
+# qhasm:   in4 = z4
+# asm 1: movd   <z4=int6464#1,>in4=int32#1
+# asm 2: movd   <z4=%xmm0,>in4=%eax
+movd   %xmm0,%eax
+
+# qhasm:   in5 = z5
+# asm 1: movd   <z5=int6464#2,>in5=int32#2
+# asm 2: movd   <z5=%xmm1,>in5=%ecx
+movd   %xmm1,%ecx
+
+# qhasm:   in6 = z6
+# asm 1: movd   <z6=int6464#3,>in6=int32#3
+# asm 2: movd   <z6=%xmm2,>in6=%edx
+movd   %xmm2,%edx
+
+# qhasm:   in7 = z7
+# asm 1: movd   <z7=int6464#4,>in7=int32#4
+# asm 2: movd   <z7=%xmm3,>in7=%ebx
+movd   %xmm3,%ebx
+
+# qhasm:   z4 <<<= 96
+# asm 1: pshufd $0x39,<z4=int6464#1,<z4=int6464#1
+# asm 2: pshufd $0x39,<z4=%xmm0,<z4=%xmm0
+pshufd $0x39,%xmm0,%xmm0
+
+# qhasm:   z5 <<<= 96
+# asm 1: pshufd $0x39,<z5=int6464#2,<z5=int6464#2
+# asm 2: pshufd $0x39,<z5=%xmm1,<z5=%xmm1
+pshufd $0x39,%xmm1,%xmm1
+
+# qhasm:   z6 <<<= 96
+# asm 1: pshufd $0x39,<z6=int6464#3,<z6=int6464#3
+# asm 2: pshufd $0x39,<z6=%xmm2,<z6=%xmm2
+pshufd $0x39,%xmm2,%xmm2
+
+# qhasm:   z7 <<<= 96
+# asm 1: pshufd $0x39,<z7=int6464#4,<z7=int6464#4
+# asm 2: pshufd $0x39,<z7=%xmm3,<z7=%xmm3
+pshufd $0x39,%xmm3,%xmm3
+
+# qhasm:   in4 ^= *(uint32 *) (m + 16)
+# asm 1: xorl 16(<m=int32#5),<in4=int32#1
+# asm 2: xorl 16(<m=%esi),<in4=%eax
+xorl 16(%esi),%eax
+
+# qhasm:   in5 ^= *(uint32 *) (m + 20)
+# asm 1: xorl 20(<m=int32#5),<in5=int32#2
+# asm 2: xorl 20(<m=%esi),<in5=%ecx
+xorl 20(%esi),%ecx
+
+# qhasm:   in6 ^= *(uint32 *) (m + 24)
+# asm 1: xorl 24(<m=int32#5),<in6=int32#3
+# asm 2: xorl 24(<m=%esi),<in6=%edx
+xorl 24(%esi),%edx
+
+# qhasm:   in7 ^= *(uint32 *) (m + 28)
+# asm 1: xorl 28(<m=int32#5),<in7=int32#4
+# asm 2: xorl 28(<m=%esi),<in7=%ebx
+xorl 28(%esi),%ebx
+
+# qhasm:   *(uint32 *) (out + 16) = in4
+# asm 1: movl <in4=int32#1,16(<out=int32#6)
+# asm 2: movl <in4=%eax,16(<out=%edi)
+movl %eax,16(%edi)
+
+# qhasm:   *(uint32 *) (out + 20) = in5
+# asm 1: movl <in5=int32#2,20(<out=int32#6)
+# asm 2: movl <in5=%ecx,20(<out=%edi)
+movl %ecx,20(%edi)
+
+# qhasm:   *(uint32 *) (out + 24) = in6
+# asm 1: movl <in6=int32#3,24(<out=int32#6)
+# asm 2: movl <in6=%edx,24(<out=%edi)
+movl %edx,24(%edi)
+
+# qhasm:   *(uint32 *) (out + 28) = in7
+# asm 1: movl <in7=int32#4,28(<out=int32#6)
+# asm 2: movl <in7=%ebx,28(<out=%edi)
+movl %ebx,28(%edi)
+
+# qhasm:   in4 = z4
+# asm 1: movd   <z4=int6464#1,>in4=int32#1
+# asm 2: movd   <z4=%xmm0,>in4=%eax
+movd   %xmm0,%eax
+
+# qhasm:   in5 = z5
+# asm 1: movd   <z5=int6464#2,>in5=int32#2
+# asm 2: movd   <z5=%xmm1,>in5=%ecx
+movd   %xmm1,%ecx
+
+# qhasm:   in6 = z6
+# asm 1: movd   <z6=int6464#3,>in6=int32#3
+# asm 2: movd   <z6=%xmm2,>in6=%edx
+movd   %xmm2,%edx
+
+# qhasm:   in7 = z7
+# asm 1: movd   <z7=int6464#4,>in7=int32#4
+# asm 2: movd   <z7=%xmm3,>in7=%ebx
+movd   %xmm3,%ebx
+
+# qhasm:   z4 <<<= 96
+# asm 1: pshufd $0x39,<z4=int6464#1,<z4=int6464#1
+# asm 2: pshufd $0x39,<z4=%xmm0,<z4=%xmm0
+pshufd $0x39,%xmm0,%xmm0
+
+# qhasm:   z5 <<<= 96
+# asm 1: pshufd $0x39,<z5=int6464#2,<z5=int6464#2
+# asm 2: pshufd $0x39,<z5=%xmm1,<z5=%xmm1
+pshufd $0x39,%xmm1,%xmm1
+
+# qhasm:   z6 <<<= 96
+# asm 1: pshufd $0x39,<z6=int6464#3,<z6=int6464#3
+# asm 2: pshufd $0x39,<z6=%xmm2,<z6=%xmm2
+pshufd $0x39,%xmm2,%xmm2
+
+# qhasm:   z7 <<<= 96
+# asm 1: pshufd $0x39,<z7=int6464#4,<z7=int6464#4
+# asm 2: pshufd $0x39,<z7=%xmm3,<z7=%xmm3
+pshufd $0x39,%xmm3,%xmm3
+
+# qhasm:   in4 ^= *(uint32 *) (m + 80)
+# asm 1: xorl 80(<m=int32#5),<in4=int32#1
+# asm 2: xorl 80(<m=%esi),<in4=%eax
+xorl 80(%esi),%eax
+
+# qhasm:   in5 ^= *(uint32 *) (m + 84)
+# asm 1: xorl 84(<m=int32#5),<in5=int32#2
+# asm 2: xorl 84(<m=%esi),<in5=%ecx
+xorl 84(%esi),%ecx
+
+# qhasm:   in6 ^= *(uint32 *) (m + 88)
+# asm 1: xorl 88(<m=int32#5),<in6=int32#3
+# asm 2: xorl 88(<m=%esi),<in6=%edx
+xorl 88(%esi),%edx
+
+# qhasm:   in7 ^= *(uint32 *) (m + 92)
+# asm 1: xorl 92(<m=int32#5),<in7=int32#4
+# asm 2: xorl 92(<m=%esi),<in7=%ebx
+xorl 92(%esi),%ebx
+
+# qhasm:   *(uint32 *) (out + 80) = in4
+# asm 1: movl <in4=int32#1,80(<out=int32#6)
+# asm 2: movl <in4=%eax,80(<out=%edi)
+movl %eax,80(%edi)
+
+# qhasm:   *(uint32 *) (out + 84) = in5
+# asm 1: movl <in5=int32#2,84(<out=int32#6)
+# asm 2: movl <in5=%ecx,84(<out=%edi)
+movl %ecx,84(%edi)
+
+# qhasm:   *(uint32 *) (out + 88) = in6
+# asm 1: movl <in6=int32#3,88(<out=int32#6)
+# asm 2: movl <in6=%edx,88(<out=%edi)
+movl %edx,88(%edi)
+
+# qhasm:   *(uint32 *) (out + 92) = in7
+# asm 1: movl <in7=int32#4,92(<out=int32#6)
+# asm 2: movl <in7=%ebx,92(<out=%edi)
+movl %ebx,92(%edi)
+
+# qhasm:   in4 = z4
+# asm 1: movd   <z4=int6464#1,>in4=int32#1
+# asm 2: movd   <z4=%xmm0,>in4=%eax
+movd   %xmm0,%eax
+
+# qhasm:   in5 = z5
+# asm 1: movd   <z5=int6464#2,>in5=int32#2
+# asm 2: movd   <z5=%xmm1,>in5=%ecx
+movd   %xmm1,%ecx
+
+# qhasm:   in6 = z6
+# asm 1: movd   <z6=int6464#3,>in6=int32#3
+# asm 2: movd   <z6=%xmm2,>in6=%edx
+movd   %xmm2,%edx
+
+# qhasm:   in7 = z7
+# asm 1: movd   <z7=int6464#4,>in7=int32#4
+# asm 2: movd   <z7=%xmm3,>in7=%ebx
+movd   %xmm3,%ebx
+
+# qhasm:   z4 <<<= 96
+# asm 1: pshufd $0x39,<z4=int6464#1,<z4=int6464#1
+# asm 2: pshufd $0x39,<z4=%xmm0,<z4=%xmm0
+pshufd $0x39,%xmm0,%xmm0
+
+# qhasm:   z5 <<<= 96
+# asm 1: pshufd $0x39,<z5=int6464#2,<z5=int6464#2
+# asm 2: pshufd $0x39,<z5=%xmm1,<z5=%xmm1
+pshufd $0x39,%xmm1,%xmm1
+
+# qhasm:   z6 <<<= 96
+# asm 1: pshufd $0x39,<z6=int6464#3,<z6=int6464#3
+# asm 2: pshufd $0x39,<z6=%xmm2,<z6=%xmm2
+pshufd $0x39,%xmm2,%xmm2
+
+# qhasm:   z7 <<<= 96
+# asm 1: pshufd $0x39,<z7=int6464#4,<z7=int6464#4
+# asm 2: pshufd $0x39,<z7=%xmm3,<z7=%xmm3
+pshufd $0x39,%xmm3,%xmm3
+
+# qhasm:   in4 ^= *(uint32 *) (m + 144)
+# asm 1: xorl 144(<m=int32#5),<in4=int32#1
+# asm 2: xorl 144(<m=%esi),<in4=%eax
+xorl 144(%esi),%eax
+
+# qhasm:   in5 ^= *(uint32 *) (m + 148)
+# asm 1: xorl 148(<m=int32#5),<in5=int32#2
+# asm 2: xorl 148(<m=%esi),<in5=%ecx
+xorl 148(%esi),%ecx
+
+# qhasm:   in6 ^= *(uint32 *) (m + 152)
+# asm 1: xorl 152(<m=int32#5),<in6=int32#3
+# asm 2: xorl 152(<m=%esi),<in6=%edx
+xorl 152(%esi),%edx
+
+# qhasm:   in7 ^= *(uint32 *) (m + 156)
+# asm 1: xorl 156(<m=int32#5),<in7=int32#4
+# asm 2: xorl 156(<m=%esi),<in7=%ebx
+xorl 156(%esi),%ebx
+
+# qhasm:   *(uint32 *) (out + 144) = in4
+# asm 1: movl <in4=int32#1,144(<out=int32#6)
+# asm 2: movl <in4=%eax,144(<out=%edi)
+movl %eax,144(%edi)
+
+# qhasm:   *(uint32 *) (out + 148) = in5
+# asm 1: movl <in5=int32#2,148(<out=int32#6)
+# asm 2: movl <in5=%ecx,148(<out=%edi)
+movl %ecx,148(%edi)
+
+# qhasm:   *(uint32 *) (out + 152) = in6
+# asm 1: movl <in6=int32#3,152(<out=int32#6)
+# asm 2: movl <in6=%edx,152(<out=%edi)
+movl %edx,152(%edi)
+
+# qhasm:   *(uint32 *) (out + 156) = in7
+# asm 1: movl <in7=int32#4,156(<out=int32#6)
+# asm 2: movl <in7=%ebx,156(<out=%edi)
+movl %ebx,156(%edi)
+
+# qhasm:   in4 = z4
+# asm 1: movd   <z4=int6464#1,>in4=int32#1
+# asm 2: movd   <z4=%xmm0,>in4=%eax
+movd   %xmm0,%eax
+
+# qhasm:   in5 = z5
+# asm 1: movd   <z5=int6464#2,>in5=int32#2
+# asm 2: movd   <z5=%xmm1,>in5=%ecx
+movd   %xmm1,%ecx
+
+# qhasm:   in6 = z6
+# asm 1: movd   <z6=int6464#3,>in6=int32#3
+# asm 2: movd   <z6=%xmm2,>in6=%edx
+movd   %xmm2,%edx
+
+# qhasm:   in7 = z7
+# asm 1: movd   <z7=int6464#4,>in7=int32#4
+# asm 2: movd   <z7=%xmm3,>in7=%ebx
+movd   %xmm3,%ebx
+
+# qhasm:   in4 ^= *(uint32 *) (m + 208)
+# asm 1: xorl 208(<m=int32#5),<in4=int32#1
+# asm 2: xorl 208(<m=%esi),<in4=%eax
+xorl 208(%esi),%eax
+
+# qhasm:   in5 ^= *(uint32 *) (m + 212)
+# asm 1: xorl 212(<m=int32#5),<in5=int32#2
+# asm 2: xorl 212(<m=%esi),<in5=%ecx
+xorl 212(%esi),%ecx
+
+# qhasm:   in6 ^= *(uint32 *) (m + 216)
+# asm 1: xorl 216(<m=int32#5),<in6=int32#3
+# asm 2: xorl 216(<m=%esi),<in6=%edx
+xorl 216(%esi),%edx
+
+# qhasm:   in7 ^= *(uint32 *) (m + 220)
+# asm 1: xorl 220(<m=int32#5),<in7=int32#4
+# asm 2: xorl 220(<m=%esi),<in7=%ebx
+xorl 220(%esi),%ebx
+
+# qhasm:   *(uint32 *) (out + 208) = in4
+# asm 1: movl <in4=int32#1,208(<out=int32#6)
+# asm 2: movl <in4=%eax,208(<out=%edi)
+movl %eax,208(%edi)
+
+# qhasm:   *(uint32 *) (out + 212) = in5
+# asm 1: movl <in5=int32#2,212(<out=int32#6)
+# asm 2: movl <in5=%ecx,212(<out=%edi)
+movl %ecx,212(%edi)
+
+# qhasm:   *(uint32 *) (out + 216) = in6
+# asm 1: movl <in6=int32#3,216(<out=int32#6)
+# asm 2: movl <in6=%edx,216(<out=%edi)
+movl %edx,216(%edi)
+
+# qhasm:   *(uint32 *) (out + 220) = in7
+# asm 1: movl <in7=int32#4,220(<out=int32#6)
+# asm 2: movl <in7=%ebx,220(<out=%edi)
+movl %ebx,220(%edi)
+
+# qhasm:   z8 = z8_stack
+# asm 1: movdqa <z8_stack=stack128#37,>z8=int6464#1
+# asm 2: movdqa <z8_stack=608(%esp),>z8=%xmm0
+movdqa 608(%esp),%xmm0
+
+# qhasm:   z9 = z9_stack
+# asm 1: movdqa <z9_stack=stack128#32,>z9=int6464#2
+# asm 2: movdqa <z9_stack=528(%esp),>z9=%xmm1
+movdqa 528(%esp),%xmm1
+
+# qhasm:   z10 = z10_stack
+# asm 1: movdqa <z10_stack=stack128#22,>z10=int6464#3
+# asm 2: movdqa <z10_stack=368(%esp),>z10=%xmm2
+movdqa 368(%esp),%xmm2
+
+# qhasm:   z11 = z11_stack
+# asm 1: movdqa <z11_stack=stack128#27,>z11=int6464#4
+# asm 2: movdqa <z11_stack=448(%esp),>z11=%xmm3
+movdqa 448(%esp),%xmm3
+
+# qhasm:   uint32323232 z8 += orig8
+# asm 1: paddd <orig8=stack128#19,<z8=int6464#1
+# asm 2: paddd <orig8=320(%esp),<z8=%xmm0
+paddd 320(%esp),%xmm0
+
+# qhasm:   uint32323232 z9 += orig9
+# asm 1: paddd <orig9=stack128#20,<z9=int6464#2
+# asm 2: paddd <orig9=336(%esp),<z9=%xmm1
+paddd 336(%esp),%xmm1
+
+# qhasm:   uint32323232 z10 += orig10
+# asm 1: paddd <orig10=stack128#6,<z10=int6464#3
+# asm 2: paddd <orig10=112(%esp),<z10=%xmm2
+paddd 112(%esp),%xmm2
+
+# qhasm:   uint32323232 z11 += orig11
+# asm 1: paddd <orig11=stack128#10,<z11=int6464#4
+# asm 2: paddd <orig11=176(%esp),<z11=%xmm3
+paddd 176(%esp),%xmm3
+
+# qhasm:   in8 = z8
+# asm 1: movd   <z8=int6464#1,>in8=int32#1
+# asm 2: movd   <z8=%xmm0,>in8=%eax
+movd   %xmm0,%eax
+
+# qhasm:   in9 = z9
+# asm 1: movd   <z9=int6464#2,>in9=int32#2
+# asm 2: movd   <z9=%xmm1,>in9=%ecx
+movd   %xmm1,%ecx
+
+# qhasm:   in10 = z10
+# asm 1: movd   <z10=int6464#3,>in10=int32#3
+# asm 2: movd   <z10=%xmm2,>in10=%edx
+movd   %xmm2,%edx
+
+# qhasm:   in11 = z11
+# asm 1: movd   <z11=int6464#4,>in11=int32#4
+# asm 2: movd   <z11=%xmm3,>in11=%ebx
+movd   %xmm3,%ebx
+
+# qhasm:   z8 <<<= 96
+# asm 1: pshufd $0x39,<z8=int6464#1,<z8=int6464#1
+# asm 2: pshufd $0x39,<z8=%xmm0,<z8=%xmm0
+pshufd $0x39,%xmm0,%xmm0
+
+# qhasm:   z9 <<<= 96
+# asm 1: pshufd $0x39,<z9=int6464#2,<z9=int6464#2
+# asm 2: pshufd $0x39,<z9=%xmm1,<z9=%xmm1
+pshufd $0x39,%xmm1,%xmm1
+
+# qhasm:   z10 <<<= 96
+# asm 1: pshufd $0x39,<z10=int6464#3,<z10=int6464#3
+# asm 2: pshufd $0x39,<z10=%xmm2,<z10=%xmm2
+pshufd $0x39,%xmm2,%xmm2
+
+# qhasm:   z11 <<<= 96
+# asm 1: pshufd $0x39,<z11=int6464#4,<z11=int6464#4
+# asm 2: pshufd $0x39,<z11=%xmm3,<z11=%xmm3
+pshufd $0x39,%xmm3,%xmm3
+
+# qhasm:   in8 ^= *(uint32 *) (m + 32)
+# asm 1: xorl 32(<m=int32#5),<in8=int32#1
+# asm 2: xorl 32(<m=%esi),<in8=%eax
+xorl 32(%esi),%eax
+
+# qhasm:   in9 ^= *(uint32 *) (m + 36)
+# asm 1: xorl 36(<m=int32#5),<in9=int32#2
+# asm 2: xorl 36(<m=%esi),<in9=%ecx
+xorl 36(%esi),%ecx
+
+# qhasm:   in10 ^= *(uint32 *) (m + 40)
+# asm 1: xorl 40(<m=int32#5),<in10=int32#3
+# asm 2: xorl 40(<m=%esi),<in10=%edx
+xorl 40(%esi),%edx
+
+# qhasm:   in11 ^= *(uint32 *) (m + 44)
+# asm 1: xorl 44(<m=int32#5),<in11=int32#4
+# asm 2: xorl 44(<m=%esi),<in11=%ebx
+xorl 44(%esi),%ebx
+
+# qhasm:   *(uint32 *) (out + 32) = in8
+# asm 1: movl <in8=int32#1,32(<out=int32#6)
+# asm 2: movl <in8=%eax,32(<out=%edi)
+movl %eax,32(%edi)
+
+# qhasm:   *(uint32 *) (out + 36) = in9
+# asm 1: movl <in9=int32#2,36(<out=int32#6)
+# asm 2: movl <in9=%ecx,36(<out=%edi)
+movl %ecx,36(%edi)
+
+# qhasm:   *(uint32 *) (out + 40) = in10
+# asm 1: movl <in10=int32#3,40(<out=int32#6)
+# asm 2: movl <in10=%edx,40(<out=%edi)
+movl %edx,40(%edi)
+
+# qhasm:   *(uint32 *) (out + 44) = in11
+# asm 1: movl <in11=int32#4,44(<out=int32#6)
+# asm 2: movl <in11=%ebx,44(<out=%edi)
+movl %ebx,44(%edi)
+
+# qhasm:   in8 = z8
+# asm 1: movd   <z8=int6464#1,>in8=int32#1
+# asm 2: movd   <z8=%xmm0,>in8=%eax
+movd   %xmm0,%eax
+
+# qhasm:   in9 = z9
+# asm 1: movd   <z9=int6464#2,>in9=int32#2
+# asm 2: movd   <z9=%xmm1,>in9=%ecx
+movd   %xmm1,%ecx
+
+# qhasm:   in10 = z10
+# asm 1: movd   <z10=int6464#3,>in10=int32#3
+# asm 2: movd   <z10=%xmm2,>in10=%edx
+movd   %xmm2,%edx
+
+# qhasm:   in11 = z11
+# asm 1: movd   <z11=int6464#4,>in11=int32#4
+# asm 2: movd   <z11=%xmm3,>in11=%ebx
+movd   %xmm3,%ebx
+
+# qhasm:   z8 <<<= 96
+# asm 1: pshufd $0x39,<z8=int6464#1,<z8=int6464#1
+# asm 2: pshufd $0x39,<z8=%xmm0,<z8=%xmm0
+pshufd $0x39,%xmm0,%xmm0
+
+# qhasm:   z9 <<<= 96
+# asm 1: pshufd $0x39,<z9=int6464#2,<z9=int6464#2
+# asm 2: pshufd $0x39,<z9=%xmm1,<z9=%xmm1
+pshufd $0x39,%xmm1,%xmm1
+
+# qhasm:   z10 <<<= 96
+# asm 1: pshufd $0x39,<z10=int6464#3,<z10=int6464#3
+# asm 2: pshufd $0x39,<z10=%xmm2,<z10=%xmm2
+pshufd $0x39,%xmm2,%xmm2
+
+# qhasm:   z11 <<<= 96
+# asm 1: pshufd $0x39,<z11=int6464#4,<z11=int6464#4
+# asm 2: pshufd $0x39,<z11=%xmm3,<z11=%xmm3
+pshufd $0x39,%xmm3,%xmm3
+
+# qhasm:   in8 ^= *(uint32 *) (m + 96)
+# asm 1: xorl 96(<m=int32#5),<in8=int32#1
+# asm 2: xorl 96(<m=%esi),<in8=%eax
+xorl 96(%esi),%eax
+
+# qhasm:   in9 ^= *(uint32 *) (m + 100)
+# asm 1: xorl 100(<m=int32#5),<in9=int32#2
+# asm 2: xorl 100(<m=%esi),<in9=%ecx
+xorl 100(%esi),%ecx
+
+# qhasm:   in10 ^= *(uint32 *) (m + 104)
+# asm 1: xorl 104(<m=int32#5),<in10=int32#3
+# asm 2: xorl 104(<m=%esi),<in10=%edx
+xorl 104(%esi),%edx
+
+# qhasm:   in11 ^= *(uint32 *) (m + 108)
+# asm 1: xorl 108(<m=int32#5),<in11=int32#4
+# asm 2: xorl 108(<m=%esi),<in11=%ebx
+xorl 108(%esi),%ebx
+
+# qhasm:   *(uint32 *) (out + 96) = in8
+# asm 1: movl <in8=int32#1,96(<out=int32#6)
+# asm 2: movl <in8=%eax,96(<out=%edi)
+movl %eax,96(%edi)
+
+# qhasm:   *(uint32 *) (out + 100) = in9
+# asm 1: movl <in9=int32#2,100(<out=int32#6)
+# asm 2: movl <in9=%ecx,100(<out=%edi)
+movl %ecx,100(%edi)
+
+# qhasm:   *(uint32 *) (out + 104) = in10
+# asm 1: movl <in10=int32#3,104(<out=int32#6)
+# asm 2: movl <in10=%edx,104(<out=%edi)
+movl %edx,104(%edi)
+
+# qhasm:   *(uint32 *) (out + 108) = in11
+# asm 1: movl <in11=int32#4,108(<out=int32#6)
+# asm 2: movl <in11=%ebx,108(<out=%edi)
+movl %ebx,108(%edi)
+
+# qhasm:   in8 = z8
+# asm 1: movd   <z8=int6464#1,>in8=int32#1
+# asm 2: movd   <z8=%xmm0,>in8=%eax
+movd   %xmm0,%eax
+
+# qhasm:   in9 = z9
+# asm 1: movd   <z9=int6464#2,>in9=int32#2
+# asm 2: movd   <z9=%xmm1,>in9=%ecx
+movd   %xmm1,%ecx
+
+# qhasm:   in10 = z10
+# asm 1: movd   <z10=int6464#3,>in10=int32#3
+# asm 2: movd   <z10=%xmm2,>in10=%edx
+movd   %xmm2,%edx
+
+# qhasm:   in11 = z11
+# asm 1: movd   <z11=int6464#4,>in11=int32#4
+# asm 2: movd   <z11=%xmm3,>in11=%ebx
+movd   %xmm3,%ebx
+
+# qhasm:   z8 <<<= 96
+# asm 1: pshufd $0x39,<z8=int6464#1,<z8=int6464#1
+# asm 2: pshufd $0x39,<z8=%xmm0,<z8=%xmm0
+pshufd $0x39,%xmm0,%xmm0
+
+# qhasm:   z9 <<<= 96
+# asm 1: pshufd $0x39,<z9=int6464#2,<z9=int6464#2
+# asm 2: pshufd $0x39,<z9=%xmm1,<z9=%xmm1
+pshufd $0x39,%xmm1,%xmm1
+
+# qhasm:   z10 <<<= 96
+# asm 1: pshufd $0x39,<z10=int6464#3,<z10=int6464#3
+# asm 2: pshufd $0x39,<z10=%xmm2,<z10=%xmm2
+pshufd $0x39,%xmm2,%xmm2
+
+# qhasm:   z11 <<<= 96
+# asm 1: pshufd $0x39,<z11=int6464#4,<z11=int6464#4
+# asm 2: pshufd $0x39,<z11=%xmm3,<z11=%xmm3
+pshufd $0x39,%xmm3,%xmm3
+
+# qhasm:   in8 ^= *(uint32 *) (m + 160)
+# asm 1: xorl 160(<m=int32#5),<in8=int32#1
+# asm 2: xorl 160(<m=%esi),<in8=%eax
+xorl 160(%esi),%eax
+
+# qhasm:   in9 ^= *(uint32 *) (m + 164)
+# asm 1: xorl 164(<m=int32#5),<in9=int32#2
+# asm 2: xorl 164(<m=%esi),<in9=%ecx
+xorl 164(%esi),%ecx
+
+# qhasm:   in10 ^= *(uint32 *) (m + 168)
+# asm 1: xorl 168(<m=int32#5),<in10=int32#3
+# asm 2: xorl 168(<m=%esi),<in10=%edx
+xorl 168(%esi),%edx
+
+# qhasm:   in11 ^= *(uint32 *) (m + 172)
+# asm 1: xorl 172(<m=int32#5),<in11=int32#4
+# asm 2: xorl 172(<m=%esi),<in11=%ebx
+xorl 172(%esi),%ebx
+
+# qhasm:   *(uint32 *) (out + 160) = in8
+# asm 1: movl <in8=int32#1,160(<out=int32#6)
+# asm 2: movl <in8=%eax,160(<out=%edi)
+movl %eax,160(%edi)
+
+# qhasm:   *(uint32 *) (out + 164) = in9
+# asm 1: movl <in9=int32#2,164(<out=int32#6)
+# asm 2: movl <in9=%ecx,164(<out=%edi)
+movl %ecx,164(%edi)
+
+# qhasm:   *(uint32 *) (out + 168) = in10
+# asm 1: movl <in10=int32#3,168(<out=int32#6)
+# asm 2: movl <in10=%edx,168(<out=%edi)
+movl %edx,168(%edi)
+
+# qhasm:   *(uint32 *) (out + 172) = in11
+# asm 1: movl <in11=int32#4,172(<out=int32#6)
+# asm 2: movl <in11=%ebx,172(<out=%edi)
+movl %ebx,172(%edi)
+
+# qhasm:   in8 = z8
+# asm 1: movd   <z8=int6464#1,>in8=int32#1
+# asm 2: movd   <z8=%xmm0,>in8=%eax
+movd   %xmm0,%eax
+
+# qhasm:   in9 = z9
+# asm 1: movd   <z9=int6464#2,>in9=int32#2
+# asm 2: movd   <z9=%xmm1,>in9=%ecx
+movd   %xmm1,%ecx
+
+# qhasm:   in10 = z10
+# asm 1: movd   <z10=int6464#3,>in10=int32#3
+# asm 2: movd   <z10=%xmm2,>in10=%edx
+movd   %xmm2,%edx
+
+# qhasm:   in11 = z11
+# asm 1: movd   <z11=int6464#4,>in11=int32#4
+# asm 2: movd   <z11=%xmm3,>in11=%ebx
+movd   %xmm3,%ebx
+
+# qhasm:   in8 ^= *(uint32 *) (m + 224)
+# asm 1: xorl 224(<m=int32#5),<in8=int32#1
+# asm 2: xorl 224(<m=%esi),<in8=%eax
+xorl 224(%esi),%eax
+
+# qhasm:   in9 ^= *(uint32 *) (m + 228)
+# asm 1: xorl 228(<m=int32#5),<in9=int32#2
+# asm 2: xorl 228(<m=%esi),<in9=%ecx
+xorl 228(%esi),%ecx
+
+# qhasm:   in10 ^= *(uint32 *) (m + 232)
+# asm 1: xorl 232(<m=int32#5),<in10=int32#3
+# asm 2: xorl 232(<m=%esi),<in10=%edx
+xorl 232(%esi),%edx
+
+# qhasm:   in11 ^= *(uint32 *) (m + 236)
+# asm 1: xorl 236(<m=int32#5),<in11=int32#4
+# asm 2: xorl 236(<m=%esi),<in11=%ebx
+xorl 236(%esi),%ebx
+
+# qhasm:   *(uint32 *) (out + 224) = in8
+# asm 1: movl <in8=int32#1,224(<out=int32#6)
+# asm 2: movl <in8=%eax,224(<out=%edi)
+movl %eax,224(%edi)
+
+# qhasm:   *(uint32 *) (out + 228) = in9
+# asm 1: movl <in9=int32#2,228(<out=int32#6)
+# asm 2: movl <in9=%ecx,228(<out=%edi)
+movl %ecx,228(%edi)
+
+# qhasm:   *(uint32 *) (out + 232) = in10
+# asm 1: movl <in10=int32#3,232(<out=int32#6)
+# asm 2: movl <in10=%edx,232(<out=%edi)
+movl %edx,232(%edi)
+
+# qhasm:   *(uint32 *) (out + 236) = in11
+# asm 1: movl <in11=int32#4,236(<out=int32#6)
+# asm 2: movl <in11=%ebx,236(<out=%edi)
+movl %ebx,236(%edi)
+
+# qhasm:   z12 = z12_stack
+# asm 1: movdqa <z12_stack=stack128#35,>z12=int6464#1
+# asm 2: movdqa <z12_stack=576(%esp),>z12=%xmm0
+movdqa 576(%esp),%xmm0
+
+# qhasm:   z13 = z13_stack
+# asm 1: movdqa <z13_stack=stack128#30,>z13=int6464#2
+# asm 2: movdqa <z13_stack=496(%esp),>z13=%xmm1
+movdqa 496(%esp),%xmm1
+
+# qhasm:   z14 = z14_stack
+# asm 1: movdqa <z14_stack=stack128#24,>z14=int6464#3
+# asm 2: movdqa <z14_stack=400(%esp),>z14=%xmm2
+movdqa 400(%esp),%xmm2
+
+# qhasm:   z15 = z15_stack
+# asm 1: movdqa <z15_stack=stack128#23,>z15=int6464#4
+# asm 2: movdqa <z15_stack=384(%esp),>z15=%xmm3
+movdqa 384(%esp),%xmm3
+
+# qhasm:   uint32323232 z12 += orig12
+# asm 1: paddd <orig12=stack128#11,<z12=int6464#1
+# asm 2: paddd <orig12=192(%esp),<z12=%xmm0
+paddd 192(%esp),%xmm0
+
+# qhasm:   uint32323232 z13 += orig13
+# asm 1: paddd <orig13=stack128#14,<z13=int6464#2
+# asm 2: paddd <orig13=240(%esp),<z13=%xmm1
+paddd 240(%esp),%xmm1
+
+# qhasm:   uint32323232 z14 += orig14
+# asm 1: paddd <orig14=stack128#17,<z14=int6464#3
+# asm 2: paddd <orig14=288(%esp),<z14=%xmm2
+paddd 288(%esp),%xmm2
+
+# qhasm:   uint32323232 z15 += orig15
+# asm 1: paddd <orig15=stack128#7,<z15=int6464#4
+# asm 2: paddd <orig15=128(%esp),<z15=%xmm3
+paddd 128(%esp),%xmm3
+
+# qhasm:   in12 = z12
+# asm 1: movd   <z12=int6464#1,>in12=int32#1
+# asm 2: movd   <z12=%xmm0,>in12=%eax
+movd   %xmm0,%eax
+
+# qhasm:   in13 = z13
+# asm 1: movd   <z13=int6464#2,>in13=int32#2
+# asm 2: movd   <z13=%xmm1,>in13=%ecx
+movd   %xmm1,%ecx
+
+# qhasm:   in14 = z14
+# asm 1: movd   <z14=int6464#3,>in14=int32#3
+# asm 2: movd   <z14=%xmm2,>in14=%edx
+movd   %xmm2,%edx
+
+# qhasm:   in15 = z15
+# asm 1: movd   <z15=int6464#4,>in15=int32#4
+# asm 2: movd   <z15=%xmm3,>in15=%ebx
+movd   %xmm3,%ebx
+
+# qhasm:   z12 <<<= 96
+# asm 1: pshufd $0x39,<z12=int6464#1,<z12=int6464#1
+# asm 2: pshufd $0x39,<z12=%xmm0,<z12=%xmm0
+pshufd $0x39,%xmm0,%xmm0
+
+# qhasm:   z13 <<<= 96
+# asm 1: pshufd $0x39,<z13=int6464#2,<z13=int6464#2
+# asm 2: pshufd $0x39,<z13=%xmm1,<z13=%xmm1
+pshufd $0x39,%xmm1,%xmm1
+
+# qhasm:   z14 <<<= 96
+# asm 1: pshufd $0x39,<z14=int6464#3,<z14=int6464#3
+# asm 2: pshufd $0x39,<z14=%xmm2,<z14=%xmm2
+pshufd $0x39,%xmm2,%xmm2
+
+# qhasm:   z15 <<<= 96
+# asm 1: pshufd $0x39,<z15=int6464#4,<z15=int6464#4
+# asm 2: pshufd $0x39,<z15=%xmm3,<z15=%xmm3
+pshufd $0x39,%xmm3,%xmm3
+
+# qhasm:   in12 ^= *(uint32 *) (m + 48)
+# asm 1: xorl 48(<m=int32#5),<in12=int32#1
+# asm 2: xorl 48(<m=%esi),<in12=%eax
+xorl 48(%esi),%eax
+
+# qhasm:   in13 ^= *(uint32 *) (m + 52)
+# asm 1: xorl 52(<m=int32#5),<in13=int32#2
+# asm 2: xorl 52(<m=%esi),<in13=%ecx
+xorl 52(%esi),%ecx
+
+# qhasm:   in14 ^= *(uint32 *) (m + 56)
+# asm 1: xorl 56(<m=int32#5),<in14=int32#3
+# asm 2: xorl 56(<m=%esi),<in14=%edx
+xorl 56(%esi),%edx
+
+# qhasm:   in15 ^= *(uint32 *) (m + 60)
+# asm 1: xorl 60(<m=int32#5),<in15=int32#4
+# asm 2: xorl 60(<m=%esi),<in15=%ebx
+xorl 60(%esi),%ebx
+
+# qhasm:   *(uint32 *) (out + 48) = in12
+# asm 1: movl <in12=int32#1,48(<out=int32#6)
+# asm 2: movl <in12=%eax,48(<out=%edi)
+movl %eax,48(%edi)
+
+# qhasm:   *(uint32 *) (out + 52) = in13
+# asm 1: movl <in13=int32#2,52(<out=int32#6)
+# asm 2: movl <in13=%ecx,52(<out=%edi)
+movl %ecx,52(%edi)
+
+# qhasm:   *(uint32 *) (out + 56) = in14
+# asm 1: movl <in14=int32#3,56(<out=int32#6)
+# asm 2: movl <in14=%edx,56(<out=%edi)
+movl %edx,56(%edi)
+
+# qhasm:   *(uint32 *) (out + 60) = in15
+# asm 1: movl <in15=int32#4,60(<out=int32#6)
+# asm 2: movl <in15=%ebx,60(<out=%edi)
+movl %ebx,60(%edi)
+
+# qhasm:   in12 = z12
+# asm 1: movd   <z12=int6464#1,>in12=int32#1
+# asm 2: movd   <z12=%xmm0,>in12=%eax
+movd   %xmm0,%eax
+
+# qhasm:   in13 = z13
+# asm 1: movd   <z13=int6464#2,>in13=int32#2
+# asm 2: movd   <z13=%xmm1,>in13=%ecx
+movd   %xmm1,%ecx
+
+# qhasm:   in14 = z14
+# asm 1: movd   <z14=int6464#3,>in14=int32#3
+# asm 2: movd   <z14=%xmm2,>in14=%edx
+movd   %xmm2,%edx
+
+# qhasm:   in15 = z15
+# asm 1: movd   <z15=int6464#4,>in15=int32#4
+# asm 2: movd   <z15=%xmm3,>in15=%ebx
+movd   %xmm3,%ebx
+
+# qhasm:   z12 <<<= 96
+# asm 1: pshufd $0x39,<z12=int6464#1,<z12=int6464#1
+# asm 2: pshufd $0x39,<z12=%xmm0,<z12=%xmm0
+pshufd $0x39,%xmm0,%xmm0
+
+# qhasm:   z13 <<<= 96
+# asm 1: pshufd $0x39,<z13=int6464#2,<z13=int6464#2
+# asm 2: pshufd $0x39,<z13=%xmm1,<z13=%xmm1
+pshufd $0x39,%xmm1,%xmm1
+
+# qhasm:   z14 <<<= 96
+# asm 1: pshufd $0x39,<z14=int6464#3,<z14=int6464#3
+# asm 2: pshufd $0x39,<z14=%xmm2,<z14=%xmm2
+pshufd $0x39,%xmm2,%xmm2
+
+# qhasm:   z15 <<<= 96
+# asm 1: pshufd $0x39,<z15=int6464#4,<z15=int6464#4
+# asm 2: pshufd $0x39,<z15=%xmm3,<z15=%xmm3
+pshufd $0x39,%xmm3,%xmm3
+
+# qhasm:   in12 ^= *(uint32 *) (m + 112)
+# asm 1: xorl 112(<m=int32#5),<in12=int32#1
+# asm 2: xorl 112(<m=%esi),<in12=%eax
+xorl 112(%esi),%eax
+
+# qhasm:   in13 ^= *(uint32 *) (m + 116)
+# asm 1: xorl 116(<m=int32#5),<in13=int32#2
+# asm 2: xorl 116(<m=%esi),<in13=%ecx
+xorl 116(%esi),%ecx
+
+# qhasm:   in14 ^= *(uint32 *) (m + 120)
+# asm 1: xorl 120(<m=int32#5),<in14=int32#3
+# asm 2: xorl 120(<m=%esi),<in14=%edx
+xorl 120(%esi),%edx
+
+# qhasm:   in15 ^= *(uint32 *) (m + 124)
+# asm 1: xorl 124(<m=int32#5),<in15=int32#4
+# asm 2: xorl 124(<m=%esi),<in15=%ebx
+xorl 124(%esi),%ebx
+
+# qhasm:   *(uint32 *) (out + 112) = in12
+# asm 1: movl <in12=int32#1,112(<out=int32#6)
+# asm 2: movl <in12=%eax,112(<out=%edi)
+movl %eax,112(%edi)
+
+# qhasm:   *(uint32 *) (out + 116) = in13
+# asm 1: movl <in13=int32#2,116(<out=int32#6)
+# asm 2: movl <in13=%ecx,116(<out=%edi)
+movl %ecx,116(%edi)
+
+# qhasm:   *(uint32 *) (out + 120) = in14
+# asm 1: movl <in14=int32#3,120(<out=int32#6)
+# asm 2: movl <in14=%edx,120(<out=%edi)
+movl %edx,120(%edi)
+
+# qhasm:   *(uint32 *) (out + 124) = in15
+# asm 1: movl <in15=int32#4,124(<out=int32#6)
+# asm 2: movl <in15=%ebx,124(<out=%edi)
+movl %ebx,124(%edi)
+
+# qhasm:   in12 = z12
+# asm 1: movd   <z12=int6464#1,>in12=int32#1
+# asm 2: movd   <z12=%xmm0,>in12=%eax
+movd   %xmm0,%eax
+
+# qhasm:   in13 = z13
+# asm 1: movd   <z13=int6464#2,>in13=int32#2
+# asm 2: movd   <z13=%xmm1,>in13=%ecx
+movd   %xmm1,%ecx
+
+# qhasm:   in14 = z14
+# asm 1: movd   <z14=int6464#3,>in14=int32#3
+# asm 2: movd   <z14=%xmm2,>in14=%edx
+movd   %xmm2,%edx
+
+# qhasm:   in15 = z15
+# asm 1: movd   <z15=int6464#4,>in15=int32#4
+# asm 2: movd   <z15=%xmm3,>in15=%ebx
+movd   %xmm3,%ebx
+
+# qhasm:   z12 <<<= 96
+# asm 1: pshufd $0x39,<z12=int6464#1,<z12=int6464#1
+# asm 2: pshufd $0x39,<z12=%xmm0,<z12=%xmm0
+pshufd $0x39,%xmm0,%xmm0
+
+# qhasm:   z13 <<<= 96
+# asm 1: pshufd $0x39,<z13=int6464#2,<z13=int6464#2
+# asm 2: pshufd $0x39,<z13=%xmm1,<z13=%xmm1
+pshufd $0x39,%xmm1,%xmm1
+
+# qhasm:   z14 <<<= 96
+# asm 1: pshufd $0x39,<z14=int6464#3,<z14=int6464#3
+# asm 2: pshufd $0x39,<z14=%xmm2,<z14=%xmm2
+pshufd $0x39,%xmm2,%xmm2
+
+# qhasm:   z15 <<<= 96
+# asm 1: pshufd $0x39,<z15=int6464#4,<z15=int6464#4
+# asm 2: pshufd $0x39,<z15=%xmm3,<z15=%xmm3
+pshufd $0x39,%xmm3,%xmm3
+
+# qhasm:   in12 ^= *(uint32 *) (m + 176)
+# asm 1: xorl 176(<m=int32#5),<in12=int32#1
+# asm 2: xorl 176(<m=%esi),<in12=%eax
+xorl 176(%esi),%eax
+
+# qhasm:   in13 ^= *(uint32 *) (m + 180)
+# asm 1: xorl 180(<m=int32#5),<in13=int32#2
+# asm 2: xorl 180(<m=%esi),<in13=%ecx
+xorl 180(%esi),%ecx
+
+# qhasm:   in14 ^= *(uint32 *) (m + 184)
+# asm 1: xorl 184(<m=int32#5),<in14=int32#3
+# asm 2: xorl 184(<m=%esi),<in14=%edx
+xorl 184(%esi),%edx
+
+# qhasm:   in15 ^= *(uint32 *) (m + 188)
+# asm 1: xorl 188(<m=int32#5),<in15=int32#4
+# asm 2: xorl 188(<m=%esi),<in15=%ebx
+xorl 188(%esi),%ebx
+
+# qhasm:   *(uint32 *) (out + 176) = in12
+# asm 1: movl <in12=int32#1,176(<out=int32#6)
+# asm 2: movl <in12=%eax,176(<out=%edi)
+movl %eax,176(%edi)
+
+# qhasm:   *(uint32 *) (out + 180) = in13
+# asm 1: movl <in13=int32#2,180(<out=int32#6)
+# asm 2: movl <in13=%ecx,180(<out=%edi)
+movl %ecx,180(%edi)
+
+# qhasm:   *(uint32 *) (out + 184) = in14
+# asm 1: movl <in14=int32#3,184(<out=int32#6)
+# asm 2: movl <in14=%edx,184(<out=%edi)
+movl %edx,184(%edi)
+
+# qhasm:   *(uint32 *) (out + 188) = in15
+# asm 1: movl <in15=int32#4,188(<out=int32#6)
+# asm 2: movl <in15=%ebx,188(<out=%edi)
+movl %ebx,188(%edi)
+
+# qhasm:   in12 = z12
+# asm 1: movd   <z12=int6464#1,>in12=int32#1
+# asm 2: movd   <z12=%xmm0,>in12=%eax
+movd   %xmm0,%eax
+
+# qhasm:   in13 = z13
+# asm 1: movd   <z13=int6464#2,>in13=int32#2
+# asm 2: movd   <z13=%xmm1,>in13=%ecx
+movd   %xmm1,%ecx
+
+# qhasm:   in14 = z14
+# asm 1: movd   <z14=int6464#3,>in14=int32#3
+# asm 2: movd   <z14=%xmm2,>in14=%edx
+movd   %xmm2,%edx
+
+# qhasm:   in15 = z15
+# asm 1: movd   <z15=int6464#4,>in15=int32#4
+# asm 2: movd   <z15=%xmm3,>in15=%ebx
+movd   %xmm3,%ebx
+
+# qhasm:   in12 ^= *(uint32 *) (m + 240)
+# asm 1: xorl 240(<m=int32#5),<in12=int32#1
+# asm 2: xorl 240(<m=%esi),<in12=%eax
+xorl 240(%esi),%eax
+
+# qhasm:   in13 ^= *(uint32 *) (m + 244)
+# asm 1: xorl 244(<m=int32#5),<in13=int32#2
+# asm 2: xorl 244(<m=%esi),<in13=%ecx
+xorl 244(%esi),%ecx
+
+# qhasm:   in14 ^= *(uint32 *) (m + 248)
+# asm 1: xorl 248(<m=int32#5),<in14=int32#3
+# asm 2: xorl 248(<m=%esi),<in14=%edx
+xorl 248(%esi),%edx
+
+# qhasm:   in15 ^= *(uint32 *) (m + 252)
+# asm 1: xorl 252(<m=int32#5),<in15=int32#4
+# asm 2: xorl 252(<m=%esi),<in15=%ebx
+xorl 252(%esi),%ebx
+
+# qhasm:   *(uint32 *) (out + 240) = in12
+# asm 1: movl <in12=int32#1,240(<out=int32#6)
+# asm 2: movl <in12=%eax,240(<out=%edi)
+movl %eax,240(%edi)
+
+# qhasm:   *(uint32 *) (out + 244) = in13
+# asm 1: movl <in13=int32#2,244(<out=int32#6)
+# asm 2: movl <in13=%ecx,244(<out=%edi)
+movl %ecx,244(%edi)
+
+# qhasm:   *(uint32 *) (out + 248) = in14
+# asm 1: movl <in14=int32#3,248(<out=int32#6)
+# asm 2: movl <in14=%edx,248(<out=%edi)
+movl %edx,248(%edi)
+
+# qhasm:   *(uint32 *) (out + 252) = in15
+# asm 1: movl <in15=int32#4,252(<out=int32#6)
+# asm 2: movl <in15=%ebx,252(<out=%edi)
+movl %ebx,252(%edi)
+
+# qhasm:   bytes = bytes_stack
+# asm 1: movl <bytes_stack=stack32#7,>bytes=int32#1
+# asm 2: movl <bytes_stack=24(%esp),>bytes=%eax
+movl 24(%esp),%eax
+
+# qhasm:   bytes -= 256
+# asm 1: sub  $256,<bytes=int32#1
+# asm 2: sub  $256,<bytes=%eax
+sub  $256,%eax
+
+# qhasm:   m += 256
+# asm 1: add  $256,<m=int32#5
+# asm 2: add  $256,<m=%esi
+add  $256,%esi
+
+# qhasm:   out += 256
+# asm 1: add  $256,<out=int32#6
+# asm 2: add  $256,<out=%edi
+add  $256,%edi
+
+# qhasm:   out_stack = out
+# asm 1: movl <out=int32#6,>out_stack=stack32#6
+# asm 2: movl <out=%edi,>out_stack=20(%esp)
+movl %edi,20(%esp)
+
+# qhasm:                            unsigned<? bytes - 256
+# asm 1: cmp  $256,<bytes=int32#1
+# asm 2: cmp  $256,<bytes=%eax
+cmp  $256,%eax
+# comment:fp stack unchanged by jump
+
+# qhasm:   goto bytesatleast256 if !unsigned<
+jae ._bytesatleast256
+
+# qhasm:                 unsigned>? bytes - 0
+# asm 1: cmp  $0,<bytes=int32#1
+# asm 2: cmp  $0,<bytes=%eax
+cmp  $0,%eax
+# comment:fp stack unchanged by jump
+
+# qhasm:   goto done if !unsigned>
+jbe ._done
+# comment:fp stack unchanged by fallthrough
+
+# qhasm: bytesbetween1and255:
+._bytesbetween1and255:
+
+# qhasm:                   unsigned<? bytes - 64
+# asm 1: cmp  $64,<bytes=int32#1
+# asm 2: cmp  $64,<bytes=%eax
+cmp  $64,%eax
+# comment:fp stack unchanged by jump
+
+# qhasm:   goto nocopy if !unsigned<
+jae ._nocopy
+
+# qhasm:     ctarget = out
+# asm 1: movl <out=int32#6,>ctarget=stack32#6
+# asm 2: movl <out=%edi,>ctarget=20(%esp)
+movl %edi,20(%esp)
+
+# qhasm:     out = &tmp
+# asm 1: leal <tmp=stack512#1,>out=int32#6
+# asm 2: leal <tmp=640(%esp),>out=%edi
+leal 640(%esp),%edi
+
+# qhasm:     i = bytes
+# asm 1: mov  <bytes=int32#1,>i=int32#2
+# asm 2: mov  <bytes=%eax,>i=%ecx
+mov  %eax,%ecx
+
+# qhasm:     while (i) { *out++ = *m++; --i }
+rep movsb
+
+# qhasm:     out = &tmp
+# asm 1: leal <tmp=stack512#1,>out=int32#6
+# asm 2: leal <tmp=640(%esp),>out=%edi
+leal 640(%esp),%edi
+
+# qhasm:     m = &tmp
+# asm 1: leal <tmp=stack512#1,>m=int32#5
+# asm 2: leal <tmp=640(%esp),>m=%esi
+leal 640(%esp),%esi
+# comment:fp stack unchanged by fallthrough
+
+# qhasm:   nocopy:
+._nocopy:
+
+# qhasm:   bytes_stack = bytes
+# asm 1: movl <bytes=int32#1,>bytes_stack=stack32#7
+# asm 2: movl <bytes=%eax,>bytes_stack=24(%esp)
+movl %eax,24(%esp)
+
+# qhasm: diag0 = x0
+# asm 1: movdqa <x0=stack128#3,>diag0=int6464#1
+# asm 2: movdqa <x0=64(%esp),>diag0=%xmm0
+movdqa 64(%esp),%xmm0
+
+# qhasm: diag1 = x1
+# asm 1: movdqa <x1=stack128#2,>diag1=int6464#2
+# asm 2: movdqa <x1=48(%esp),>diag1=%xmm1
+movdqa 48(%esp),%xmm1
+
+# qhasm: diag2 = x2
+# asm 1: movdqa <x2=stack128#4,>diag2=int6464#3
+# asm 2: movdqa <x2=80(%esp),>diag2=%xmm2
+movdqa 80(%esp),%xmm2
+
+# qhasm: diag3 = x3
+# asm 1: movdqa <x3=stack128#1,>diag3=int6464#4
+# asm 2: movdqa <x3=32(%esp),>diag3=%xmm3
+movdqa 32(%esp),%xmm3
+
+# qhasm:                     a0 = diag1
+# asm 1: movdqa <diag1=int6464#2,>a0=int6464#5
+# asm 2: movdqa <diag1=%xmm1,>a0=%xmm4
+movdqa %xmm1,%xmm4
+
+# qhasm: i = 20
+# asm 1: mov  $20,>i=int32#1
+# asm 2: mov  $20,>i=%eax
+mov  $20,%eax
+
+# qhasm: mainloop2:
+._mainloop2:
+
+# qhasm: uint32323232        a0 += diag0
+# asm 1: paddd <diag0=int6464#1,<a0=int6464#5
+# asm 2: paddd <diag0=%xmm0,<a0=%xmm4
+paddd %xmm0,%xmm4
+
+# qhasm:                                 a1 = diag0
+# asm 1: movdqa <diag0=int6464#1,>a1=int6464#6
+# asm 2: movdqa <diag0=%xmm0,>a1=%xmm5
+movdqa %xmm0,%xmm5
+
+# qhasm:                     b0 = a0
+# asm 1: movdqa <a0=int6464#5,>b0=int6464#7
+# asm 2: movdqa <a0=%xmm4,>b0=%xmm6
+movdqa %xmm4,%xmm6
+
+# qhasm: uint32323232        a0 <<= 7
+# asm 1: pslld $7,<a0=int6464#5
+# asm 2: pslld $7,<a0=%xmm4
+pslld $7,%xmm4
+
+# qhasm: uint32323232        b0 >>= 25
+# asm 1: psrld $25,<b0=int6464#7
+# asm 2: psrld $25,<b0=%xmm6
+psrld $25,%xmm6
+
+# qhasm:                 diag3 ^= a0
+# asm 1: pxor  <a0=int6464#5,<diag3=int6464#4
+# asm 2: pxor  <a0=%xmm4,<diag3=%xmm3
+pxor  %xmm4,%xmm3
+
+# qhasm:                 diag3 ^= b0
+# asm 1: pxor  <b0=int6464#7,<diag3=int6464#4
+# asm 2: pxor  <b0=%xmm6,<diag3=%xmm3
+pxor  %xmm6,%xmm3
+
+# qhasm: uint32323232                        a1 += diag3
+# asm 1: paddd <diag3=int6464#4,<a1=int6464#6
+# asm 2: paddd <diag3=%xmm3,<a1=%xmm5
+paddd %xmm3,%xmm5
+
+# qhasm:                                                 a2 = diag3
+# asm 1: movdqa <diag3=int6464#4,>a2=int6464#5
+# asm 2: movdqa <diag3=%xmm3,>a2=%xmm4
+movdqa %xmm3,%xmm4
+
+# qhasm:                                     b1 = a1
+# asm 1: movdqa <a1=int6464#6,>b1=int6464#7
+# asm 2: movdqa <a1=%xmm5,>b1=%xmm6
+movdqa %xmm5,%xmm6
+
+# qhasm: uint32323232                        a1 <<= 9
+# asm 1: pslld $9,<a1=int6464#6
+# asm 2: pslld $9,<a1=%xmm5
+pslld $9,%xmm5
+
+# qhasm: uint32323232                        b1 >>= 23
+# asm 1: psrld $23,<b1=int6464#7
+# asm 2: psrld $23,<b1=%xmm6
+psrld $23,%xmm6
+
+# qhasm:                                 diag2 ^= a1
+# asm 1: pxor  <a1=int6464#6,<diag2=int6464#3
+# asm 2: pxor  <a1=%xmm5,<diag2=%xmm2
+pxor  %xmm5,%xmm2
+
+# qhasm:                 diag3 <<<= 32
+# asm 1: pshufd $0x93,<diag3=int6464#4,<diag3=int6464#4
+# asm 2: pshufd $0x93,<diag3=%xmm3,<diag3=%xmm3
+pshufd $0x93,%xmm3,%xmm3
+
+# qhasm:                                 diag2 ^= b1
+# asm 1: pxor  <b1=int6464#7,<diag2=int6464#3
+# asm 2: pxor  <b1=%xmm6,<diag2=%xmm2
+pxor  %xmm6,%xmm2
+
+# qhasm: uint32323232                                        a2 += diag2
+# asm 1: paddd <diag2=int6464#3,<a2=int6464#5
+# asm 2: paddd <diag2=%xmm2,<a2=%xmm4
+paddd %xmm2,%xmm4
+
+# qhasm:                                                                 a3 = diag2
+# asm 1: movdqa <diag2=int6464#3,>a3=int6464#6
+# asm 2: movdqa <diag2=%xmm2,>a3=%xmm5
+movdqa %xmm2,%xmm5
+
+# qhasm:                                                     b2 = a2
+# asm 1: movdqa <a2=int6464#5,>b2=int6464#7
+# asm 2: movdqa <a2=%xmm4,>b2=%xmm6
+movdqa %xmm4,%xmm6
+
+# qhasm: uint32323232                                        a2 <<= 13
+# asm 1: pslld $13,<a2=int6464#5
+# asm 2: pslld $13,<a2=%xmm4
+pslld $13,%xmm4
+
+# qhasm: uint32323232                                        b2 >>= 19
+# asm 1: psrld $19,<b2=int6464#7
+# asm 2: psrld $19,<b2=%xmm6
+psrld $19,%xmm6
+
+# qhasm:                                                 diag1 ^= a2
+# asm 1: pxor  <a2=int6464#5,<diag1=int6464#2
+# asm 2: pxor  <a2=%xmm4,<diag1=%xmm1
+pxor  %xmm4,%xmm1
+
+# qhasm:                                 diag2 <<<= 64
+# asm 1: pshufd $0x4e,<diag2=int6464#3,<diag2=int6464#3
+# asm 2: pshufd $0x4e,<diag2=%xmm2,<diag2=%xmm2
+pshufd $0x4e,%xmm2,%xmm2
+
+# qhasm:                                                 diag1 ^= b2
+# asm 1: pxor  <b2=int6464#7,<diag1=int6464#2
+# asm 2: pxor  <b2=%xmm6,<diag1=%xmm1
+pxor  %xmm6,%xmm1
+
+# qhasm: uint32323232                                                        a3 += diag1
+# asm 1: paddd <diag1=int6464#2,<a3=int6464#6
+# asm 2: paddd <diag1=%xmm1,<a3=%xmm5
+paddd %xmm1,%xmm5
+
+# qhasm:                 a4 = diag3
+# asm 1: movdqa <diag3=int6464#4,>a4=int6464#5
+# asm 2: movdqa <diag3=%xmm3,>a4=%xmm4
+movdqa %xmm3,%xmm4
+
+# qhasm:                                                                     b3 = a3
+# asm 1: movdqa <a3=int6464#6,>b3=int6464#7
+# asm 2: movdqa <a3=%xmm5,>b3=%xmm6
+movdqa %xmm5,%xmm6
+
+# qhasm: uint32323232                                                        a3 <<= 18
+# asm 1: pslld $18,<a3=int6464#6
+# asm 2: pslld $18,<a3=%xmm5
+pslld $18,%xmm5
+
+# qhasm: uint32323232                                                        b3 >>= 14
+# asm 1: psrld $14,<b3=int6464#7
+# asm 2: psrld $14,<b3=%xmm6
+psrld $14,%xmm6
+
+# qhasm:                                                                 diag0 ^= a3
+# asm 1: pxor  <a3=int6464#6,<diag0=int6464#1
+# asm 2: pxor  <a3=%xmm5,<diag0=%xmm0
+pxor  %xmm5,%xmm0
+
+# qhasm:                                                 diag1 <<<= 96
+# asm 1: pshufd $0x39,<diag1=int6464#2,<diag1=int6464#2
+# asm 2: pshufd $0x39,<diag1=%xmm1,<diag1=%xmm1
+pshufd $0x39,%xmm1,%xmm1
+
+# qhasm:                                                                 diag0 ^= b3
+# asm 1: pxor  <b3=int6464#7,<diag0=int6464#1
+# asm 2: pxor  <b3=%xmm6,<diag0=%xmm0
+pxor  %xmm6,%xmm0
+
+# qhasm: uint32323232        a4 += diag0
+# asm 1: paddd <diag0=int6464#1,<a4=int6464#5
+# asm 2: paddd <diag0=%xmm0,<a4=%xmm4
+paddd %xmm0,%xmm4
+
+# qhasm:                                 a5 = diag0
+# asm 1: movdqa <diag0=int6464#1,>a5=int6464#6
+# asm 2: movdqa <diag0=%xmm0,>a5=%xmm5
+movdqa %xmm0,%xmm5
+
+# qhasm:                     b4 = a4
+# asm 1: movdqa <a4=int6464#5,>b4=int6464#7
+# asm 2: movdqa <a4=%xmm4,>b4=%xmm6
+movdqa %xmm4,%xmm6
+
+# qhasm: uint32323232        a4 <<= 7
+# asm 1: pslld $7,<a4=int6464#5
+# asm 2: pslld $7,<a4=%xmm4
+pslld $7,%xmm4
+
+# qhasm: uint32323232        b4 >>= 25
+# asm 1: psrld $25,<b4=int6464#7
+# asm 2: psrld $25,<b4=%xmm6
+psrld $25,%xmm6
+
+# qhasm:                 diag1 ^= a4
+# asm 1: pxor  <a4=int6464#5,<diag1=int6464#2
+# asm 2: pxor  <a4=%xmm4,<diag1=%xmm1
+pxor  %xmm4,%xmm1
+
+# qhasm:                 diag1 ^= b4
+# asm 1: pxor  <b4=int6464#7,<diag1=int6464#2
+# asm 2: pxor  <b4=%xmm6,<diag1=%xmm1
+pxor  %xmm6,%xmm1
+
+# qhasm: uint32323232                        a5 += diag1
+# asm 1: paddd <diag1=int6464#2,<a5=int6464#6
+# asm 2: paddd <diag1=%xmm1,<a5=%xmm5
+paddd %xmm1,%xmm5
+
+# qhasm:                                                 a6 = diag1
+# asm 1: movdqa <diag1=int6464#2,>a6=int6464#5
+# asm 2: movdqa <diag1=%xmm1,>a6=%xmm4
+movdqa %xmm1,%xmm4
+
+# qhasm:                                     b5 = a5
+# asm 1: movdqa <a5=int6464#6,>b5=int6464#7
+# asm 2: movdqa <a5=%xmm5,>b5=%xmm6
+movdqa %xmm5,%xmm6
+
+# qhasm: uint32323232                        a5 <<= 9
+# asm 1: pslld $9,<a5=int6464#6
+# asm 2: pslld $9,<a5=%xmm5
+pslld $9,%xmm5
+
+# qhasm: uint32323232                        b5 >>= 23
+# asm 1: psrld $23,<b5=int6464#7
+# asm 2: psrld $23,<b5=%xmm6
+psrld $23,%xmm6
+
+# qhasm:                                 diag2 ^= a5
+# asm 1: pxor  <a5=int6464#6,<diag2=int6464#3
+# asm 2: pxor  <a5=%xmm5,<diag2=%xmm2
+pxor  %xmm5,%xmm2
+
+# qhasm:                 diag1 <<<= 32
+# asm 1: pshufd $0x93,<diag1=int6464#2,<diag1=int6464#2
+# asm 2: pshufd $0x93,<diag1=%xmm1,<diag1=%xmm1
+pshufd $0x93,%xmm1,%xmm1
+
+# qhasm:                                 diag2 ^= b5
+# asm 1: pxor  <b5=int6464#7,<diag2=int6464#3
+# asm 2: pxor  <b5=%xmm6,<diag2=%xmm2
+pxor  %xmm6,%xmm2
+
+# qhasm: uint32323232                                        a6 += diag2
+# asm 1: paddd <diag2=int6464#3,<a6=int6464#5
+# asm 2: paddd <diag2=%xmm2,<a6=%xmm4
+paddd %xmm2,%xmm4
+
+# qhasm:                                                                 a7 = diag2
+# asm 1: movdqa <diag2=int6464#3,>a7=int6464#6
+# asm 2: movdqa <diag2=%xmm2,>a7=%xmm5
+movdqa %xmm2,%xmm5
+
+# qhasm:                                                     b6 = a6
+# asm 1: movdqa <a6=int6464#5,>b6=int6464#7
+# asm 2: movdqa <a6=%xmm4,>b6=%xmm6
+movdqa %xmm4,%xmm6
+
+# qhasm: uint32323232                                        a6 <<= 13
+# asm 1: pslld $13,<a6=int6464#5
+# asm 2: pslld $13,<a6=%xmm4
+pslld $13,%xmm4
+
+# qhasm: uint32323232                                        b6 >>= 19
+# asm 1: psrld $19,<b6=int6464#7
+# asm 2: psrld $19,<b6=%xmm6
+psrld $19,%xmm6
+
+# qhasm:                                                 diag3 ^= a6
+# asm 1: pxor  <a6=int6464#5,<diag3=int6464#4
+# asm 2: pxor  <a6=%xmm4,<diag3=%xmm3
+pxor  %xmm4,%xmm3
+
+# qhasm:                                 diag2 <<<= 64
+# asm 1: pshufd $0x4e,<diag2=int6464#3,<diag2=int6464#3
+# asm 2: pshufd $0x4e,<diag2=%xmm2,<diag2=%xmm2
+pshufd $0x4e,%xmm2,%xmm2
+
+# qhasm:                                                 diag3 ^= b6
+# asm 1: pxor  <b6=int6464#7,<diag3=int6464#4
+# asm 2: pxor  <b6=%xmm6,<diag3=%xmm3
+pxor  %xmm6,%xmm3
+
+# qhasm: uint32323232                                                        a7 += diag3
+# asm 1: paddd <diag3=int6464#4,<a7=int6464#6
+# asm 2: paddd <diag3=%xmm3,<a7=%xmm5
+paddd %xmm3,%xmm5
+
+# qhasm:                 a0 = diag1
+# asm 1: movdqa <diag1=int6464#2,>a0=int6464#5
+# asm 2: movdqa <diag1=%xmm1,>a0=%xmm4
+movdqa %xmm1,%xmm4
+
+# qhasm:                                                                     b7 = a7
+# asm 1: movdqa <a7=int6464#6,>b7=int6464#7
+# asm 2: movdqa <a7=%xmm5,>b7=%xmm6
+movdqa %xmm5,%xmm6
+
+# qhasm: uint32323232                                                        a7 <<= 18
+# asm 1: pslld $18,<a7=int6464#6
+# asm 2: pslld $18,<a7=%xmm5
+pslld $18,%xmm5
+
+# qhasm: uint32323232                                                        b7 >>= 14
+# asm 1: psrld $14,<b7=int6464#7
+# asm 2: psrld $14,<b7=%xmm6
+psrld $14,%xmm6
+
+# qhasm:                                                                 diag0 ^= a7
+# asm 1: pxor  <a7=int6464#6,<diag0=int6464#1
+# asm 2: pxor  <a7=%xmm5,<diag0=%xmm0
+pxor  %xmm5,%xmm0
+
+# qhasm:                                                 diag3 <<<= 96
+# asm 1: pshufd $0x39,<diag3=int6464#4,<diag3=int6464#4
+# asm 2: pshufd $0x39,<diag3=%xmm3,<diag3=%xmm3
+pshufd $0x39,%xmm3,%xmm3
+
+# qhasm:                                                                 diag0 ^= b7
+# asm 1: pxor  <b7=int6464#7,<diag0=int6464#1
+# asm 2: pxor  <b7=%xmm6,<diag0=%xmm0
+pxor  %xmm6,%xmm0
+
+# qhasm: uint32323232        a0 += diag0
+# asm 1: paddd <diag0=int6464#1,<a0=int6464#5
+# asm 2: paddd <diag0=%xmm0,<a0=%xmm4
+paddd %xmm0,%xmm4
+
+# qhasm:                                 a1 = diag0
+# asm 1: movdqa <diag0=int6464#1,>a1=int6464#6
+# asm 2: movdqa <diag0=%xmm0,>a1=%xmm5
+movdqa %xmm0,%xmm5
+
+# qhasm:                     b0 = a0
+# asm 1: movdqa <a0=int6464#5,>b0=int6464#7
+# asm 2: movdqa <a0=%xmm4,>b0=%xmm6
+movdqa %xmm4,%xmm6
+
+# qhasm: uint32323232        a0 <<= 7
+# asm 1: pslld $7,<a0=int6464#5
+# asm 2: pslld $7,<a0=%xmm4
+pslld $7,%xmm4
+
+# qhasm: uint32323232        b0 >>= 25
+# asm 1: psrld $25,<b0=int6464#7
+# asm 2: psrld $25,<b0=%xmm6
+psrld $25,%xmm6
+
+# qhasm:                 diag3 ^= a0
+# asm 1: pxor  <a0=int6464#5,<diag3=int6464#4
+# asm 2: pxor  <a0=%xmm4,<diag3=%xmm3
+pxor  %xmm4,%xmm3
+
+# qhasm:                 diag3 ^= b0
+# asm 1: pxor  <b0=int6464#7,<diag3=int6464#4
+# asm 2: pxor  <b0=%xmm6,<diag3=%xmm3
+pxor  %xmm6,%xmm3
+
+# qhasm: uint32323232                        a1 += diag3
+# asm 1: paddd <diag3=int6464#4,<a1=int6464#6
+# asm 2: paddd <diag3=%xmm3,<a1=%xmm5
+paddd %xmm3,%xmm5
+
+# qhasm:                                                 a2 = diag3
+# asm 1: movdqa <diag3=int6464#4,>a2=int6464#5
+# asm 2: movdqa <diag3=%xmm3,>a2=%xmm4
+movdqa %xmm3,%xmm4
+
+# qhasm:                                     b1 = a1
+# asm 1: movdqa <a1=int6464#6,>b1=int6464#7
+# asm 2: movdqa <a1=%xmm5,>b1=%xmm6
+movdqa %xmm5,%xmm6
+
+# qhasm: uint32323232                        a1 <<= 9
+# asm 1: pslld $9,<a1=int6464#6
+# asm 2: pslld $9,<a1=%xmm5
+pslld $9,%xmm5
+
+# qhasm: uint32323232                        b1 >>= 23
+# asm 1: psrld $23,<b1=int6464#7
+# asm 2: psrld $23,<b1=%xmm6
+psrld $23,%xmm6
+
+# qhasm:                                 diag2 ^= a1
+# asm 1: pxor  <a1=int6464#6,<diag2=int6464#3
+# asm 2: pxor  <a1=%xmm5,<diag2=%xmm2
+pxor  %xmm5,%xmm2
+
+# qhasm:                 diag3 <<<= 32
+# asm 1: pshufd $0x93,<diag3=int6464#4,<diag3=int6464#4
+# asm 2: pshufd $0x93,<diag3=%xmm3,<diag3=%xmm3
+pshufd $0x93,%xmm3,%xmm3
+
+# qhasm:                                 diag2 ^= b1
+# asm 1: pxor  <b1=int6464#7,<diag2=int6464#3
+# asm 2: pxor  <b1=%xmm6,<diag2=%xmm2
+pxor  %xmm6,%xmm2
+
+# qhasm: uint32323232                                        a2 += diag2
+# asm 1: paddd <diag2=int6464#3,<a2=int6464#5
+# asm 2: paddd <diag2=%xmm2,<a2=%xmm4
+paddd %xmm2,%xmm4
+
+# qhasm:                                                                 a3 = diag2
+# asm 1: movdqa <diag2=int6464#3,>a3=int6464#6
+# asm 2: movdqa <diag2=%xmm2,>a3=%xmm5
+movdqa %xmm2,%xmm5
+
+# qhasm:                                                     b2 = a2
+# asm 1: movdqa <a2=int6464#5,>b2=int6464#7
+# asm 2: movdqa <a2=%xmm4,>b2=%xmm6
+movdqa %xmm4,%xmm6
+
+# qhasm: uint32323232                                        a2 <<= 13
+# asm 1: pslld $13,<a2=int6464#5
+# asm 2: pslld $13,<a2=%xmm4
+pslld $13,%xmm4
+
+# qhasm: uint32323232                                        b2 >>= 19
+# asm 1: psrld $19,<b2=int6464#7
+# asm 2: psrld $19,<b2=%xmm6
+psrld $19,%xmm6
+
+# qhasm:                                                 diag1 ^= a2
+# asm 1: pxor  <a2=int6464#5,<diag1=int6464#2
+# asm 2: pxor  <a2=%xmm4,<diag1=%xmm1
+pxor  %xmm4,%xmm1
+
+# qhasm:                                 diag2 <<<= 64
+# asm 1: pshufd $0x4e,<diag2=int6464#3,<diag2=int6464#3
+# asm 2: pshufd $0x4e,<diag2=%xmm2,<diag2=%xmm2
+pshufd $0x4e,%xmm2,%xmm2
+
+# qhasm:                                                 diag1 ^= b2
+# asm 1: pxor  <b2=int6464#7,<diag1=int6464#2
+# asm 2: pxor  <b2=%xmm6,<diag1=%xmm1
+pxor  %xmm6,%xmm1
+
+# qhasm: uint32323232                                                        a3 += diag1
+# asm 1: paddd <diag1=int6464#2,<a3=int6464#6
+# asm 2: paddd <diag1=%xmm1,<a3=%xmm5
+paddd %xmm1,%xmm5
+
+# qhasm:                 a4 = diag3
+# asm 1: movdqa <diag3=int6464#4,>a4=int6464#5
+# asm 2: movdqa <diag3=%xmm3,>a4=%xmm4
+movdqa %xmm3,%xmm4
+
+# qhasm:                                                                     b3 = a3
+# asm 1: movdqa <a3=int6464#6,>b3=int6464#7
+# asm 2: movdqa <a3=%xmm5,>b3=%xmm6
+movdqa %xmm5,%xmm6
+
+# qhasm: uint32323232                                                        a3 <<= 18
+# asm 1: pslld $18,<a3=int6464#6
+# asm 2: pslld $18,<a3=%xmm5
+pslld $18,%xmm5
+
+# qhasm: uint32323232                                                        b3 >>= 14
+# asm 1: psrld $14,<b3=int6464#7
+# asm 2: psrld $14,<b3=%xmm6
+psrld $14,%xmm6
+
+# qhasm:                                                                 diag0 ^= a3
+# asm 1: pxor  <a3=int6464#6,<diag0=int6464#1
+# asm 2: pxor  <a3=%xmm5,<diag0=%xmm0
+pxor  %xmm5,%xmm0
+
+# qhasm:                                                 diag1 <<<= 96
+# asm 1: pshufd $0x39,<diag1=int6464#2,<diag1=int6464#2
+# asm 2: pshufd $0x39,<diag1=%xmm1,<diag1=%xmm1
+pshufd $0x39,%xmm1,%xmm1
+
+# qhasm:                                                                 diag0 ^= b3
+# asm 1: pxor  <b3=int6464#7,<diag0=int6464#1
+# asm 2: pxor  <b3=%xmm6,<diag0=%xmm0
+pxor  %xmm6,%xmm0
+
+# qhasm: uint32323232        a4 += diag0
+# asm 1: paddd <diag0=int6464#1,<a4=int6464#5
+# asm 2: paddd <diag0=%xmm0,<a4=%xmm4
+paddd %xmm0,%xmm4
+
+# qhasm:                                 a5 = diag0
+# asm 1: movdqa <diag0=int6464#1,>a5=int6464#6
+# asm 2: movdqa <diag0=%xmm0,>a5=%xmm5
+movdqa %xmm0,%xmm5
+
+# qhasm:                     b4 = a4
+# asm 1: movdqa <a4=int6464#5,>b4=int6464#7
+# asm 2: movdqa <a4=%xmm4,>b4=%xmm6
+movdqa %xmm4,%xmm6
+
+# qhasm: uint32323232        a4 <<= 7
+# asm 1: pslld $7,<a4=int6464#5
+# asm 2: pslld $7,<a4=%xmm4
+pslld $7,%xmm4
+
+# qhasm: uint32323232        b4 >>= 25
+# asm 1: psrld $25,<b4=int6464#7
+# asm 2: psrld $25,<b4=%xmm6
+psrld $25,%xmm6
+
+# qhasm:                 diag1 ^= a4
+# asm 1: pxor  <a4=int6464#5,<diag1=int6464#2
+# asm 2: pxor  <a4=%xmm4,<diag1=%xmm1
+pxor  %xmm4,%xmm1
+
+# qhasm:                 diag1 ^= b4
+# asm 1: pxor  <b4=int6464#7,<diag1=int6464#2
+# asm 2: pxor  <b4=%xmm6,<diag1=%xmm1
+pxor  %xmm6,%xmm1
+
+# qhasm: uint32323232                        a5 += diag1
+# asm 1: paddd <diag1=int6464#2,<a5=int6464#6
+# asm 2: paddd <diag1=%xmm1,<a5=%xmm5
+paddd %xmm1,%xmm5
+
+# qhasm:                                                 a6 = diag1
+# asm 1: movdqa <diag1=int6464#2,>a6=int6464#5
+# asm 2: movdqa <diag1=%xmm1,>a6=%xmm4
+movdqa %xmm1,%xmm4
+
+# qhasm:                                     b5 = a5
+# asm 1: movdqa <a5=int6464#6,>b5=int6464#7
+# asm 2: movdqa <a5=%xmm5,>b5=%xmm6
+movdqa %xmm5,%xmm6
+
+# qhasm: uint32323232                        a5 <<= 9
+# asm 1: pslld $9,<a5=int6464#6
+# asm 2: pslld $9,<a5=%xmm5
+pslld $9,%xmm5
+
+# qhasm: uint32323232                        b5 >>= 23
+# asm 1: psrld $23,<b5=int6464#7
+# asm 2: psrld $23,<b5=%xmm6
+psrld $23,%xmm6
+
+# qhasm:                                 diag2 ^= a5
+# asm 1: pxor  <a5=int6464#6,<diag2=int6464#3
+# asm 2: pxor  <a5=%xmm5,<diag2=%xmm2
+pxor  %xmm5,%xmm2
+
+# qhasm:                 diag1 <<<= 32
+# asm 1: pshufd $0x93,<diag1=int6464#2,<diag1=int6464#2
+# asm 2: pshufd $0x93,<diag1=%xmm1,<diag1=%xmm1
+pshufd $0x93,%xmm1,%xmm1
+
+# qhasm:                                 diag2 ^= b5
+# asm 1: pxor  <b5=int6464#7,<diag2=int6464#3
+# asm 2: pxor  <b5=%xmm6,<diag2=%xmm2
+pxor  %xmm6,%xmm2
+
+# qhasm: uint32323232                                        a6 += diag2
+# asm 1: paddd <diag2=int6464#3,<a6=int6464#5
+# asm 2: paddd <diag2=%xmm2,<a6=%xmm4
+paddd %xmm2,%xmm4
+
+# qhasm:                                                                 a7 = diag2
+# asm 1: movdqa <diag2=int6464#3,>a7=int6464#6
+# asm 2: movdqa <diag2=%xmm2,>a7=%xmm5
+movdqa %xmm2,%xmm5
+
+# qhasm:                                                     b6 = a6
+# asm 1: movdqa <a6=int6464#5,>b6=int6464#7
+# asm 2: movdqa <a6=%xmm4,>b6=%xmm6
+movdqa %xmm4,%xmm6
+
+# qhasm: uint32323232                                        a6 <<= 13
+# asm 1: pslld $13,<a6=int6464#5
+# asm 2: pslld $13,<a6=%xmm4
+pslld $13,%xmm4
+
+# qhasm: uint32323232                                        b6 >>= 19
+# asm 1: psrld $19,<b6=int6464#7
+# asm 2: psrld $19,<b6=%xmm6
+psrld $19,%xmm6
+
+# qhasm:                                                 diag3 ^= a6
+# asm 1: pxor  <a6=int6464#5,<diag3=int6464#4
+# asm 2: pxor  <a6=%xmm4,<diag3=%xmm3
+pxor  %xmm4,%xmm3
+
+# qhasm:                                 diag2 <<<= 64
+# asm 1: pshufd $0x4e,<diag2=int6464#3,<diag2=int6464#3
+# asm 2: pshufd $0x4e,<diag2=%xmm2,<diag2=%xmm2
+pshufd $0x4e,%xmm2,%xmm2
+
+# qhasm:                                                 diag3 ^= b6
+# asm 1: pxor  <b6=int6464#7,<diag3=int6464#4
+# asm 2: pxor  <b6=%xmm6,<diag3=%xmm3
+pxor  %xmm6,%xmm3
+
+# qhasm:                  unsigned>? i -= 4
+# asm 1: sub  $4,<i=int32#1
+# asm 2: sub  $4,<i=%eax
+sub  $4,%eax
+
+# qhasm: uint32323232                                                        a7 += diag3
+# asm 1: paddd <diag3=int6464#4,<a7=int6464#6
+# asm 2: paddd <diag3=%xmm3,<a7=%xmm5
+paddd %xmm3,%xmm5
+
+# qhasm:                 a0 = diag1
+# asm 1: movdqa <diag1=int6464#2,>a0=int6464#5
+# asm 2: movdqa <diag1=%xmm1,>a0=%xmm4
+movdqa %xmm1,%xmm4
+
+# qhasm:                                                                     b7 = a7
+# asm 1: movdqa <a7=int6464#6,>b7=int6464#7
+# asm 2: movdqa <a7=%xmm5,>b7=%xmm6
+movdqa %xmm5,%xmm6
+
+# qhasm: uint32323232                                                        a7 <<= 18
+# asm 1: pslld $18,<a7=int6464#6
+# asm 2: pslld $18,<a7=%xmm5
+pslld $18,%xmm5
+
+# qhasm:                 b0 = 0
+# asm 1: pxor   >b0=int6464#8,>b0=int6464#8
+# asm 2: pxor   >b0=%xmm7,>b0=%xmm7
+pxor   %xmm7,%xmm7
+
+# qhasm: uint32323232                                                        b7 >>= 14
+# asm 1: psrld $14,<b7=int6464#7
+# asm 2: psrld $14,<b7=%xmm6
+psrld $14,%xmm6
+
+# qhasm:                                                                 diag0 ^= a7
+# asm 1: pxor  <a7=int6464#6,<diag0=int6464#1
+# asm 2: pxor  <a7=%xmm5,<diag0=%xmm0
+pxor  %xmm5,%xmm0
+
+# qhasm:                                                 diag3 <<<= 96
+# asm 1: pshufd $0x39,<diag3=int6464#4,<diag3=int6464#4
+# asm 2: pshufd $0x39,<diag3=%xmm3,<diag3=%xmm3
+pshufd $0x39,%xmm3,%xmm3
+
+# qhasm:                                                                 diag0 ^= b7
+# asm 1: pxor  <b7=int6464#7,<diag0=int6464#1
+# asm 2: pxor  <b7=%xmm6,<diag0=%xmm0
+pxor  %xmm6,%xmm0
+# comment:fp stack unchanged by jump
+
+# qhasm: goto mainloop2 if unsigned>
+ja ._mainloop2
+
+# qhasm: uint32323232 diag0 += x0
+# asm 1: paddd <x0=stack128#3,<diag0=int6464#1
+# asm 2: paddd <x0=64(%esp),<diag0=%xmm0
+paddd 64(%esp),%xmm0
+
+# qhasm: uint32323232 diag1 += x1
+# asm 1: paddd <x1=stack128#2,<diag1=int6464#2
+# asm 2: paddd <x1=48(%esp),<diag1=%xmm1
+paddd 48(%esp),%xmm1
+
+# qhasm: uint32323232 diag2 += x2
+# asm 1: paddd <x2=stack128#4,<diag2=int6464#3
+# asm 2: paddd <x2=80(%esp),<diag2=%xmm2
+paddd 80(%esp),%xmm2
+
+# qhasm: uint32323232 diag3 += x3
+# asm 1: paddd <x3=stack128#1,<diag3=int6464#4
+# asm 2: paddd <x3=32(%esp),<diag3=%xmm3
+paddd 32(%esp),%xmm3
+
+# qhasm: in0 = diag0
+# asm 1: movd   <diag0=int6464#1,>in0=int32#1
+# asm 2: movd   <diag0=%xmm0,>in0=%eax
+movd   %xmm0,%eax
+
+# qhasm: in12 = diag1
+# asm 1: movd   <diag1=int6464#2,>in12=int32#2
+# asm 2: movd   <diag1=%xmm1,>in12=%ecx
+movd   %xmm1,%ecx
+
+# qhasm: in8 = diag2
+# asm 1: movd   <diag2=int6464#3,>in8=int32#3
+# asm 2: movd   <diag2=%xmm2,>in8=%edx
+movd   %xmm2,%edx
+
+# qhasm: in4 = diag3
+# asm 1: movd   <diag3=int6464#4,>in4=int32#4
+# asm 2: movd   <diag3=%xmm3,>in4=%ebx
+movd   %xmm3,%ebx
+
+# qhasm: diag0 <<<= 96
+# asm 1: pshufd $0x39,<diag0=int6464#1,<diag0=int6464#1
+# asm 2: pshufd $0x39,<diag0=%xmm0,<diag0=%xmm0
+pshufd $0x39,%xmm0,%xmm0
+
+# qhasm: diag1 <<<= 96
+# asm 1: pshufd $0x39,<diag1=int6464#2,<diag1=int6464#2
+# asm 2: pshufd $0x39,<diag1=%xmm1,<diag1=%xmm1
+pshufd $0x39,%xmm1,%xmm1
+
+# qhasm: diag2 <<<= 96
+# asm 1: pshufd $0x39,<diag2=int6464#3,<diag2=int6464#3
+# asm 2: pshufd $0x39,<diag2=%xmm2,<diag2=%xmm2
+pshufd $0x39,%xmm2,%xmm2
+
+# qhasm: diag3 <<<= 96
+# asm 1: pshufd $0x39,<diag3=int6464#4,<diag3=int6464#4
+# asm 2: pshufd $0x39,<diag3=%xmm3,<diag3=%xmm3
+pshufd $0x39,%xmm3,%xmm3
+
+# qhasm: in0 ^= *(uint32 *) (m + 0)
+# asm 1: xorl 0(<m=int32#5),<in0=int32#1
+# asm 2: xorl 0(<m=%esi),<in0=%eax
+xorl 0(%esi),%eax
+
+# qhasm: in12 ^= *(uint32 *) (m + 48)
+# asm 1: xorl 48(<m=int32#5),<in12=int32#2
+# asm 2: xorl 48(<m=%esi),<in12=%ecx
+xorl 48(%esi),%ecx
+
+# qhasm: in8 ^= *(uint32 *) (m + 32)
+# asm 1: xorl 32(<m=int32#5),<in8=int32#3
+# asm 2: xorl 32(<m=%esi),<in8=%edx
+xorl 32(%esi),%edx
+
+# qhasm: in4 ^= *(uint32 *) (m + 16)
+# asm 1: xorl 16(<m=int32#5),<in4=int32#4
+# asm 2: xorl 16(<m=%esi),<in4=%ebx
+xorl 16(%esi),%ebx
+
+# qhasm: *(uint32 *) (out + 0) = in0
+# asm 1: movl <in0=int32#1,0(<out=int32#6)
+# asm 2: movl <in0=%eax,0(<out=%edi)
+movl %eax,0(%edi)
+
+# qhasm: *(uint32 *) (out + 48) = in12
+# asm 1: movl <in12=int32#2,48(<out=int32#6)
+# asm 2: movl <in12=%ecx,48(<out=%edi)
+movl %ecx,48(%edi)
+
+# qhasm: *(uint32 *) (out + 32) = in8
+# asm 1: movl <in8=int32#3,32(<out=int32#6)
+# asm 2: movl <in8=%edx,32(<out=%edi)
+movl %edx,32(%edi)
+
+# qhasm: *(uint32 *) (out + 16) = in4
+# asm 1: movl <in4=int32#4,16(<out=int32#6)
+# asm 2: movl <in4=%ebx,16(<out=%edi)
+movl %ebx,16(%edi)
+
+# qhasm: in5 = diag0
+# asm 1: movd   <diag0=int6464#1,>in5=int32#1
+# asm 2: movd   <diag0=%xmm0,>in5=%eax
+movd   %xmm0,%eax
+
+# qhasm: in1 = diag1
+# asm 1: movd   <diag1=int6464#2,>in1=int32#2
+# asm 2: movd   <diag1=%xmm1,>in1=%ecx
+movd   %xmm1,%ecx
+
+# qhasm: in13 = diag2
+# asm 1: movd   <diag2=int6464#3,>in13=int32#3
+# asm 2: movd   <diag2=%xmm2,>in13=%edx
+movd   %xmm2,%edx
+
+# qhasm: in9 = diag3
+# asm 1: movd   <diag3=int6464#4,>in9=int32#4
+# asm 2: movd   <diag3=%xmm3,>in9=%ebx
+movd   %xmm3,%ebx
+
+# qhasm: diag0 <<<= 96
+# asm 1: pshufd $0x39,<diag0=int6464#1,<diag0=int6464#1
+# asm 2: pshufd $0x39,<diag0=%xmm0,<diag0=%xmm0
+pshufd $0x39,%xmm0,%xmm0
+
+# qhasm: diag1 <<<= 96
+# asm 1: pshufd $0x39,<diag1=int6464#2,<diag1=int6464#2
+# asm 2: pshufd $0x39,<diag1=%xmm1,<diag1=%xmm1
+pshufd $0x39,%xmm1,%xmm1
+
+# qhasm: diag2 <<<= 96
+# asm 1: pshufd $0x39,<diag2=int6464#3,<diag2=int6464#3
+# asm 2: pshufd $0x39,<diag2=%xmm2,<diag2=%xmm2
+pshufd $0x39,%xmm2,%xmm2
+
+# qhasm: diag3 <<<= 96
+# asm 1: pshufd $0x39,<diag3=int6464#4,<diag3=int6464#4
+# asm 2: pshufd $0x39,<diag3=%xmm3,<diag3=%xmm3
+pshufd $0x39,%xmm3,%xmm3
+
+# qhasm: in5 ^= *(uint32 *) (m + 20)
+# asm 1: xorl 20(<m=int32#5),<in5=int32#1
+# asm 2: xorl 20(<m=%esi),<in5=%eax
+xorl 20(%esi),%eax
+
+# qhasm: in1 ^= *(uint32 *) (m + 4)
+# asm 1: xorl 4(<m=int32#5),<in1=int32#2
+# asm 2: xorl 4(<m=%esi),<in1=%ecx
+xorl 4(%esi),%ecx
+
+# qhasm: in13 ^= *(uint32 *) (m + 52)
+# asm 1: xorl 52(<m=int32#5),<in13=int32#3
+# asm 2: xorl 52(<m=%esi),<in13=%edx
+xorl 52(%esi),%edx
+
+# qhasm: in9 ^= *(uint32 *) (m + 36)
+# asm 1: xorl 36(<m=int32#5),<in9=int32#4
+# asm 2: xorl 36(<m=%esi),<in9=%ebx
+xorl 36(%esi),%ebx
+
+# qhasm: *(uint32 *) (out + 20) = in5
+# asm 1: movl <in5=int32#1,20(<out=int32#6)
+# asm 2: movl <in5=%eax,20(<out=%edi)
+movl %eax,20(%edi)
+
+# qhasm: *(uint32 *) (out + 4) = in1
+# asm 1: movl <in1=int32#2,4(<out=int32#6)
+# asm 2: movl <in1=%ecx,4(<out=%edi)
+movl %ecx,4(%edi)
+
+# qhasm: *(uint32 *) (out + 52) = in13
+# asm 1: movl <in13=int32#3,52(<out=int32#6)
+# asm 2: movl <in13=%edx,52(<out=%edi)
+movl %edx,52(%edi)
+
+# qhasm: *(uint32 *) (out + 36) = in9
+# asm 1: movl <in9=int32#4,36(<out=int32#6)
+# asm 2: movl <in9=%ebx,36(<out=%edi)
+movl %ebx,36(%edi)
+
+# qhasm: in10 = diag0
+# asm 1: movd   <diag0=int6464#1,>in10=int32#1
+# asm 2: movd   <diag0=%xmm0,>in10=%eax
+movd   %xmm0,%eax
+
+# qhasm: in6 = diag1
+# asm 1: movd   <diag1=int6464#2,>in6=int32#2
+# asm 2: movd   <diag1=%xmm1,>in6=%ecx
+movd   %xmm1,%ecx
+
+# qhasm: in2 = diag2
+# asm 1: movd   <diag2=int6464#3,>in2=int32#3
+# asm 2: movd   <diag2=%xmm2,>in2=%edx
+movd   %xmm2,%edx
+
+# qhasm: in14 = diag3
+# asm 1: movd   <diag3=int6464#4,>in14=int32#4
+# asm 2: movd   <diag3=%xmm3,>in14=%ebx
+movd   %xmm3,%ebx
+
+# qhasm: diag0 <<<= 96
+# asm 1: pshufd $0x39,<diag0=int6464#1,<diag0=int6464#1
+# asm 2: pshufd $0x39,<diag0=%xmm0,<diag0=%xmm0
+pshufd $0x39,%xmm0,%xmm0
+
+# qhasm: diag1 <<<= 96
+# asm 1: pshufd $0x39,<diag1=int6464#2,<diag1=int6464#2
+# asm 2: pshufd $0x39,<diag1=%xmm1,<diag1=%xmm1
+pshufd $0x39,%xmm1,%xmm1
+
+# qhasm: diag2 <<<= 96
+# asm 1: pshufd $0x39,<diag2=int6464#3,<diag2=int6464#3
+# asm 2: pshufd $0x39,<diag2=%xmm2,<diag2=%xmm2
+pshufd $0x39,%xmm2,%xmm2
+
+# qhasm: diag3 <<<= 96
+# asm 1: pshufd $0x39,<diag3=int6464#4,<diag3=int6464#4
+# asm 2: pshufd $0x39,<diag3=%xmm3,<diag3=%xmm3
+pshufd $0x39,%xmm3,%xmm3
+
+# qhasm: in10 ^= *(uint32 *) (m + 40)
+# asm 1: xorl 40(<m=int32#5),<in10=int32#1
+# asm 2: xorl 40(<m=%esi),<in10=%eax
+xorl 40(%esi),%eax
+
+# qhasm: in6 ^= *(uint32 *) (m + 24)
+# asm 1: xorl 24(<m=int32#5),<in6=int32#2
+# asm 2: xorl 24(<m=%esi),<in6=%ecx
+xorl 24(%esi),%ecx
+
+# qhasm: in2 ^= *(uint32 *) (m + 8)
+# asm 1: xorl 8(<m=int32#5),<in2=int32#3
+# asm 2: xorl 8(<m=%esi),<in2=%edx
+xorl 8(%esi),%edx
+
+# qhasm: in14 ^= *(uint32 *) (m + 56)
+# asm 1: xorl 56(<m=int32#5),<in14=int32#4
+# asm 2: xorl 56(<m=%esi),<in14=%ebx
+xorl 56(%esi),%ebx
+
+# qhasm: *(uint32 *) (out + 40) = in10
+# asm 1: movl <in10=int32#1,40(<out=int32#6)
+# asm 2: movl <in10=%eax,40(<out=%edi)
+movl %eax,40(%edi)
+
+# qhasm: *(uint32 *) (out + 24) = in6
+# asm 1: movl <in6=int32#2,24(<out=int32#6)
+# asm 2: movl <in6=%ecx,24(<out=%edi)
+movl %ecx,24(%edi)
+
+# qhasm: *(uint32 *) (out + 8) = in2
+# asm 1: movl <in2=int32#3,8(<out=int32#6)
+# asm 2: movl <in2=%edx,8(<out=%edi)
+movl %edx,8(%edi)
+
+# qhasm: *(uint32 *) (out + 56) = in14
+# asm 1: movl <in14=int32#4,56(<out=int32#6)
+# asm 2: movl <in14=%ebx,56(<out=%edi)
+movl %ebx,56(%edi)
+
+# qhasm: in15 = diag0
+# asm 1: movd   <diag0=int6464#1,>in15=int32#1
+# asm 2: movd   <diag0=%xmm0,>in15=%eax
+movd   %xmm0,%eax
+
+# qhasm: in11 = diag1
+# asm 1: movd   <diag1=int6464#2,>in11=int32#2
+# asm 2: movd   <diag1=%xmm1,>in11=%ecx
+movd   %xmm1,%ecx
+
+# qhasm: in7 = diag2
+# asm 1: movd   <diag2=int6464#3,>in7=int32#3
+# asm 2: movd   <diag2=%xmm2,>in7=%edx
+movd   %xmm2,%edx
+
+# qhasm: in3 = diag3
+# asm 1: movd   <diag3=int6464#4,>in3=int32#4
+# asm 2: movd   <diag3=%xmm3,>in3=%ebx
+movd   %xmm3,%ebx
+
+# qhasm: in15 ^= *(uint32 *) (m + 60)
+# asm 1: xorl 60(<m=int32#5),<in15=int32#1
+# asm 2: xorl 60(<m=%esi),<in15=%eax
+xorl 60(%esi),%eax
+
+# qhasm: in11 ^= *(uint32 *) (m + 44)
+# asm 1: xorl 44(<m=int32#5),<in11=int32#2
+# asm 2: xorl 44(<m=%esi),<in11=%ecx
+xorl 44(%esi),%ecx
+
+# qhasm: in7 ^= *(uint32 *) (m + 28)
+# asm 1: xorl 28(<m=int32#5),<in7=int32#3
+# asm 2: xorl 28(<m=%esi),<in7=%edx
+xorl 28(%esi),%edx
+
+# qhasm: in3 ^= *(uint32 *) (m + 12)
+# asm 1: xorl 12(<m=int32#5),<in3=int32#4
+# asm 2: xorl 12(<m=%esi),<in3=%ebx
+xorl 12(%esi),%ebx
+
+# qhasm: *(uint32 *) (out + 60) = in15
+# asm 1: movl <in15=int32#1,60(<out=int32#6)
+# asm 2: movl <in15=%eax,60(<out=%edi)
+movl %eax,60(%edi)
+
+# qhasm: *(uint32 *) (out + 44) = in11
+# asm 1: movl <in11=int32#2,44(<out=int32#6)
+# asm 2: movl <in11=%ecx,44(<out=%edi)
+movl %ecx,44(%edi)
+
+# qhasm: *(uint32 *) (out + 28) = in7
+# asm 1: movl <in7=int32#3,28(<out=int32#6)
+# asm 2: movl <in7=%edx,28(<out=%edi)
+movl %edx,28(%edi)
+
+# qhasm: *(uint32 *) (out + 12) = in3
+# asm 1: movl <in3=int32#4,12(<out=int32#6)
+# asm 2: movl <in3=%ebx,12(<out=%edi)
+movl %ebx,12(%edi)
+
+# qhasm:   bytes = bytes_stack
+# asm 1: movl <bytes_stack=stack32#7,>bytes=int32#1
+# asm 2: movl <bytes_stack=24(%esp),>bytes=%eax
+movl 24(%esp),%eax
+
+# qhasm:   in8 = ((uint32 *)&x2)[0]
+# asm 1: movl <x2=stack128#4,>in8=int32#2
+# asm 2: movl <x2=80(%esp),>in8=%ecx
+movl 80(%esp),%ecx
+
+# qhasm:   in9 = ((uint32 *)&x3)[1]
+# asm 1: movl 4+<x3=stack128#1,>in9=int32#3
+# asm 2: movl 4+<x3=32(%esp),>in9=%edx
+movl 4+32(%esp),%edx
+
+# qhasm:   carry? in8 += 1
+# asm 1: add  $1,<in8=int32#2
+# asm 2: add  $1,<in8=%ecx
+add  $1,%ecx
+
+# qhasm:   in9 += 0 + carry
+# asm 1: adc $0,<in9=int32#3
+# asm 2: adc $0,<in9=%edx
+adc $0,%edx
+
+# qhasm:   ((uint32 *)&x2)[0] = in8
+# asm 1: movl <in8=int32#2,>x2=stack128#4
+# asm 2: movl <in8=%ecx,>x2=80(%esp)
+movl %ecx,80(%esp)
+
+# qhasm:   ((uint32 *)&x3)[1] = in9
+# asm 1: movl <in9=int32#3,4+<x3=stack128#1
+# asm 2: movl <in9=%edx,4+<x3=32(%esp)
+movl %edx,4+32(%esp)
+
+# qhasm:                          unsigned>? unsigned<? bytes - 64
+# asm 1: cmp  $64,<bytes=int32#1
+# asm 2: cmp  $64,<bytes=%eax
+cmp  $64,%eax
+# comment:fp stack unchanged by jump
+
+# qhasm:   goto bytesatleast65 if unsigned>
+ja ._bytesatleast65
+# comment:fp stack unchanged by jump
+
+# qhasm:     goto bytesatleast64 if !unsigned<
+jae ._bytesatleast64
+
+# qhasm:       m = out
+# asm 1: mov  <out=int32#6,>m=int32#5
+# asm 2: mov  <out=%edi,>m=%esi
+mov  %edi,%esi
+
+# qhasm:       out = ctarget
+# asm 1: movl <ctarget=stack32#6,>out=int32#6
+# asm 2: movl <ctarget=20(%esp),>out=%edi
+movl 20(%esp),%edi
+
+# qhasm:       i = bytes
+# asm 1: mov  <bytes=int32#1,>i=int32#2
+# asm 2: mov  <bytes=%eax,>i=%ecx
+mov  %eax,%ecx
+
+# qhasm:       while (i) { *out++ = *m++; --i }
+rep movsb
+# comment:fp stack unchanged by fallthrough
+
+# qhasm:     bytesatleast64:
+._bytesatleast64:
+# comment:fp stack unchanged by fallthrough
+
+# qhasm:     done:
+._done:
+
+# qhasm:     eax = eax_stack
+# asm 1: movl <eax_stack=stack32#1,>eax=int32#1
+# asm 2: movl <eax_stack=0(%esp),>eax=%eax
+movl 0(%esp),%eax
+
+# qhasm:     ebx = ebx_stack
+# asm 1: movl <ebx_stack=stack32#2,>ebx=int32#4
+# asm 2: movl <ebx_stack=4(%esp),>ebx=%ebx
+movl 4(%esp),%ebx
+
+# qhasm:     esi = esi_stack
+# asm 1: movl <esi_stack=stack32#3,>esi=int32#5
+# asm 2: movl <esi_stack=8(%esp),>esi=%esi
+movl 8(%esp),%esi
+
+# qhasm:     edi = edi_stack
+# asm 1: movl <edi_stack=stack32#4,>edi=int32#6
+# asm 2: movl <edi_stack=12(%esp),>edi=%edi
+movl 12(%esp),%edi
+
+# qhasm:     ebp = ebp_stack
+# asm 1: movl <ebp_stack=stack32#5,>ebp=int32#7
+# asm 2: movl <ebp_stack=16(%esp),>ebp=%ebp
+movl 16(%esp),%ebp
+
+# qhasm:     leave
+add %eax,%esp
+xor %eax,%eax
+ret
+
+# qhasm:   bytesatleast65:
+._bytesatleast65:
+
+# qhasm:   bytes -= 64
+# asm 1: sub  $64,<bytes=int32#1
+# asm 2: sub  $64,<bytes=%eax
+sub  $64,%eax
+
+# qhasm:   out += 64
+# asm 1: add  $64,<out=int32#6
+# asm 2: add  $64,<out=%edi
+add  $64,%edi
+
+# qhasm:   m += 64
+# asm 1: add  $64,<m=int32#5
+# asm 2: add  $64,<m=%esi
+add  $64,%esi
+# comment:fp stack unchanged by jump
+
+# qhasm: goto bytesbetween1and255
+jmp ._bytesbetween1and255
diff --git a/src/crypto/cipher/salsa20/xmm/salsa20_xmm.c b/src/crypto/cipher/salsa20/xmm/salsa20_xmm.c
new file mode 100644
index 0000000..3787a0e
--- /dev/null
+++ b/src/crypto/cipher/salsa20/xmm/salsa20_xmm.c
@@ -0,0 +1,84 @@
+/*
+  Copyright (c) 2012-2013, Matthias Schiffer <mschiffer@universe-factory.net>
+  All rights reserved.
+
+  Redistribution and use in source and binary forms, with or without
+  modification, are permitted provided that the following conditions are met:
+
+    1. Redistributions of source code must retain the above copyright notice,
+       this list of conditions and the following disclaimer.
+    2. Redistributions in binary form must reproduce the above copyright notice,
+       this list of conditions and the following disclaimer in the documentation
+       and/or other materials provided with the distribution.
+
+  THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS "AS IS"
+  AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
+  IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE ARE
+  DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT HOLDER OR CONTRIBUTORS BE LIABLE
+  FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL
+  DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR
+  SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER
+  CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY,
+  OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE
+  OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
+*/
+
+/*
+  The assembly implementations were written by D. J. Bernstein and are
+  Public Domain. For more information see http://cr.yp.to/snuffle.html
+*/
+
+#include "../../../../crypto.h"
+#include "../../../../cpuid.h"
+
+
+#define KEYBYTES 32
+
+
+#ifdef __x86_64__
+#define crypto_stream_salsa20_xor crypto_stream_salsa20_amd64_xmm6_xor
+#endif
+
+#ifdef __i386__
+#define crypto_stream_salsa20_xor crypto_stream_salsa20_x86_xmm5_xor
+#endif
+
+
+int crypto_stream_salsa20_xor(unsigned char *c, const unsigned char *m, unsigned long long mlen, const unsigned char *n, const unsigned char *k);
+
+
+struct fastd_cipher_state {
+	uint8_t key[KEYBYTES];
+};
+
+
+static bool salsa20_available(void) {
+	return fastd_cpuid() & CPUID_SSE2;
+}
+
+static fastd_cipher_state_t* salsa20_init(const uint8_t *key) {
+	fastd_cipher_state_t *state = malloc(sizeof(fastd_cipher_state_t));
+	memcpy(state->key, key, KEYBYTES);
+
+	return state;
+}
+
+static bool salsa20_crypt(const fastd_cipher_state_t *state, fastd_block128_t *out, const fastd_block128_t *in, size_t len, const uint8_t *iv) {
+	crypto_stream_salsa20_xor(out->b, in->b, len, iv, state->key);
+	return true;
+}
+
+static void salsa20_free(fastd_cipher_state_t *state) {
+	if (state) {
+		secure_memzero(state, sizeof(*state));
+		free(state);
+	}
+}
+
+const fastd_cipher_t fastd_cipher_salsa20_xmm = {
+	.available = salsa20_available,
+
+	.init = salsa20_init,
+	.crypt = salsa20_crypt,
+	.free = salsa20_free,
+};
diff --git a/src/crypto/cipher/salsa2012/CMakeLists.txt b/src/crypto/cipher/salsa2012/CMakeLists.txt
index eb853e0..4821f41 100644
--- a/src/crypto/cipher/salsa2012/CMakeLists.txt
+++ b/src/crypto/cipher/salsa2012/CMakeLists.txt
@@ -1,2 +1,3 @@
 fastd_cipher(salsa2012 salsa2012.c)
+add_subdirectory(xmm)
 add_subdirectory(nacl)
diff --git a/src/crypto/cipher/salsa2012/xmm/CMakeLists.txt b/src/crypto/cipher/salsa2012/xmm/CMakeLists.txt
new file mode 100644
index 0000000..9cb5222
--- /dev/null
+++ b/src/crypto/cipher/salsa2012/xmm/CMakeLists.txt
@@ -0,0 +1,13 @@
+if(ARCH_X86_64)
+  fastd_cipher_impl(salsa2012 xmm
+    salsa2012_xmm.c
+    salsa2012_amd64_xmm6.s
+  )
+endif(ARCH_X86_64)
+
+if(ARCH_X86)
+  fastd_cipher_impl(salsa2012 xmm
+    salsa2012_xmm.c
+    salsa2012_x86_xmm5.s
+  )
+endif(ARCH_X86)
diff --git a/src/crypto/cipher/salsa2012/xmm/salsa2012_amd64_xmm6.s b/src/crypto/cipher/salsa2012/xmm/salsa2012_amd64_xmm6.s
new file mode 100644
index 0000000..0e26dc9
--- /dev/null
+++ b/src/crypto/cipher/salsa2012/xmm/salsa2012_amd64_xmm6.s
@@ -0,0 +1,4823 @@
+
+# qhasm: int64 r11_caller
+
+# qhasm: int64 r12_caller
+
+# qhasm: int64 r13_caller
+
+# qhasm: int64 r14_caller
+
+# qhasm: int64 r15_caller
+
+# qhasm: int64 rbx_caller
+
+# qhasm: int64 rbp_caller
+
+# qhasm: caller r11_caller
+
+# qhasm: caller r12_caller
+
+# qhasm: caller r13_caller
+
+# qhasm: caller r14_caller
+
+# qhasm: caller r15_caller
+
+# qhasm: caller rbx_caller
+
+# qhasm: caller rbp_caller
+
+# qhasm: stack64 r11_stack
+
+# qhasm: stack64 r12_stack
+
+# qhasm: stack64 r13_stack
+
+# qhasm: stack64 r14_stack
+
+# qhasm: stack64 r15_stack
+
+# qhasm: stack64 rbx_stack
+
+# qhasm: stack64 rbp_stack
+
+# qhasm: int64 a
+
+# qhasm: int64 arg1
+
+# qhasm: int64 arg2
+
+# qhasm: int64 arg3
+
+# qhasm: int64 arg4
+
+# qhasm: int64 arg5
+
+# qhasm: input arg1
+
+# qhasm: input arg2
+
+# qhasm: input arg3
+
+# qhasm: input arg4
+
+# qhasm: input arg5
+
+# qhasm: int64 k
+
+# qhasm: int64 kbits
+
+# qhasm: int64 iv
+
+# qhasm: int64 i
+
+# qhasm: stack128 x0
+
+# qhasm: stack128 x1
+
+# qhasm: stack128 x2
+
+# qhasm: stack128 x3
+
+# qhasm: int64 m
+
+# qhasm: int64 out
+
+# qhasm: int64 bytes
+
+# qhasm: stack32 eax_stack
+
+# qhasm: stack32 ebx_stack
+
+# qhasm: stack32 esi_stack
+
+# qhasm: stack32 edi_stack
+
+# qhasm: stack32 ebp_stack
+
+# qhasm: int6464 diag0
+
+# qhasm: int6464 diag1
+
+# qhasm: int6464 diag2
+
+# qhasm: int6464 diag3
+
+# qhasm: int6464 a0
+
+# qhasm: int6464 a1
+
+# qhasm: int6464 a2
+
+# qhasm: int6464 a3
+
+# qhasm: int6464 a4
+
+# qhasm: int6464 a5
+
+# qhasm: int6464 a6
+
+# qhasm: int6464 a7
+
+# qhasm: int6464 b0
+
+# qhasm: int6464 b1
+
+# qhasm: int6464 b2
+
+# qhasm: int6464 b3
+
+# qhasm: int6464 b4
+
+# qhasm: int6464 b5
+
+# qhasm: int6464 b6
+
+# qhasm: int6464 b7
+
+# qhasm: int6464 z0
+
+# qhasm: int6464 z1
+
+# qhasm: int6464 z2
+
+# qhasm: int6464 z3
+
+# qhasm: int6464 z4
+
+# qhasm: int6464 z5
+
+# qhasm: int6464 z6
+
+# qhasm: int6464 z7
+
+# qhasm: int6464 z8
+
+# qhasm: int6464 z9
+
+# qhasm: int6464 z10
+
+# qhasm: int6464 z11
+
+# qhasm: int6464 z12
+
+# qhasm: int6464 z13
+
+# qhasm: int6464 z14
+
+# qhasm: int6464 z15
+
+# qhasm: stack128 z0_stack
+
+# qhasm: stack128 z1_stack
+
+# qhasm: stack128 z2_stack
+
+# qhasm: stack128 z3_stack
+
+# qhasm: stack128 z4_stack
+
+# qhasm: stack128 z5_stack
+
+# qhasm: stack128 z6_stack
+
+# qhasm: stack128 z7_stack
+
+# qhasm: stack128 z8_stack
+
+# qhasm: stack128 z9_stack
+
+# qhasm: stack128 z10_stack
+
+# qhasm: stack128 z11_stack
+
+# qhasm: stack128 z12_stack
+
+# qhasm: stack128 z13_stack
+
+# qhasm: stack128 z14_stack
+
+# qhasm: stack128 z15_stack
+
+# qhasm: int6464 y0
+
+# qhasm: int6464 y1
+
+# qhasm: int6464 y2
+
+# qhasm: int6464 y3
+
+# qhasm: int6464 y4
+
+# qhasm: int6464 y5
+
+# qhasm: int6464 y6
+
+# qhasm: int6464 y7
+
+# qhasm: int6464 y8
+
+# qhasm: int6464 y9
+
+# qhasm: int6464 y10
+
+# qhasm: int6464 y11
+
+# qhasm: int6464 y12
+
+# qhasm: int6464 y13
+
+# qhasm: int6464 y14
+
+# qhasm: int6464 y15
+
+# qhasm: int6464 r0
+
+# qhasm: int6464 r1
+
+# qhasm: int6464 r2
+
+# qhasm: int6464 r3
+
+# qhasm: int6464 r4
+
+# qhasm: int6464 r5
+
+# qhasm: int6464 r6
+
+# qhasm: int6464 r7
+
+# qhasm: int6464 r8
+
+# qhasm: int6464 r9
+
+# qhasm: int6464 r10
+
+# qhasm: int6464 r11
+
+# qhasm: int6464 r12
+
+# qhasm: int6464 r13
+
+# qhasm: int6464 r14
+
+# qhasm: int6464 r15
+
+# qhasm: stack128 orig0
+
+# qhasm: stack128 orig1
+
+# qhasm: stack128 orig2
+
+# qhasm: stack128 orig3
+
+# qhasm: stack128 orig4
+
+# qhasm: stack128 orig5
+
+# qhasm: stack128 orig6
+
+# qhasm: stack128 orig7
+
+# qhasm: stack128 orig8
+
+# qhasm: stack128 orig9
+
+# qhasm: stack128 orig10
+
+# qhasm: stack128 orig11
+
+# qhasm: stack128 orig12
+
+# qhasm: stack128 orig13
+
+# qhasm: stack128 orig14
+
+# qhasm: stack128 orig15
+
+# qhasm: int64 in0
+
+# qhasm: int64 in1
+
+# qhasm: int64 in2
+
+# qhasm: int64 in3
+
+# qhasm: int64 in4
+
+# qhasm: int64 in5
+
+# qhasm: int64 in6
+
+# qhasm: int64 in7
+
+# qhasm: int64 in8
+
+# qhasm: int64 in9
+
+# qhasm: int64 in10
+
+# qhasm: int64 in11
+
+# qhasm: int64 in12
+
+# qhasm: int64 in13
+
+# qhasm: int64 in14
+
+# qhasm: int64 in15
+
+# qhasm: stack512 tmp
+
+# qhasm: int64 ctarget
+
+# qhasm: stack64 bytes_backup
+
+# qhasm: enter crypto_stream_salsa2012_amd64_xmm6
+.text
+.p2align 5
+.globl _crypto_stream_salsa2012_amd64_xmm6
+.globl crypto_stream_salsa2012_amd64_xmm6
+_crypto_stream_salsa2012_amd64_xmm6:
+crypto_stream_salsa2012_amd64_xmm6:
+mov %rsp,%r11
+and $31,%r11
+add $480,%r11
+sub %r11,%rsp
+
+# qhasm: r11_stack = r11_caller
+# asm 1: movq <r11_caller=int64#9,>r11_stack=stack64#1
+# asm 2: movq <r11_caller=%r11,>r11_stack=352(%rsp)
+movq %r11,352(%rsp)
+
+# qhasm: r12_stack = r12_caller
+# asm 1: movq <r12_caller=int64#10,>r12_stack=stack64#2
+# asm 2: movq <r12_caller=%r12,>r12_stack=360(%rsp)
+movq %r12,360(%rsp)
+
+# qhasm: r13_stack = r13_caller
+# asm 1: movq <r13_caller=int64#11,>r13_stack=stack64#3
+# asm 2: movq <r13_caller=%r13,>r13_stack=368(%rsp)
+movq %r13,368(%rsp)
+
+# qhasm: r14_stack = r14_caller
+# asm 1: movq <r14_caller=int64#12,>r14_stack=stack64#4
+# asm 2: movq <r14_caller=%r14,>r14_stack=376(%rsp)
+movq %r14,376(%rsp)
+
+# qhasm: r15_stack = r15_caller
+# asm 1: movq <r15_caller=int64#13,>r15_stack=stack64#5
+# asm 2: movq <r15_caller=%r15,>r15_stack=384(%rsp)
+movq %r15,384(%rsp)
+
+# qhasm: rbx_stack = rbx_caller
+# asm 1: movq <rbx_caller=int64#14,>rbx_stack=stack64#6
+# asm 2: movq <rbx_caller=%rbx,>rbx_stack=392(%rsp)
+movq %rbx,392(%rsp)
+
+# qhasm: rbp_stack = rbp_caller
+# asm 1: movq <rbp_caller=int64#15,>rbp_stack=stack64#7
+# asm 2: movq <rbp_caller=%rbp,>rbp_stack=400(%rsp)
+movq %rbp,400(%rsp)
+
+# qhasm: bytes = arg2
+# asm 1: mov  <arg2=int64#2,>bytes=int64#6
+# asm 2: mov  <arg2=%rsi,>bytes=%r9
+mov  %rsi,%r9
+
+# qhasm: out = arg1
+# asm 1: mov  <arg1=int64#1,>out=int64#1
+# asm 2: mov  <arg1=%rdi,>out=%rdi
+mov  %rdi,%rdi
+
+# qhasm: m = out
+# asm 1: mov  <out=int64#1,>m=int64#2
+# asm 2: mov  <out=%rdi,>m=%rsi
+mov  %rdi,%rsi
+
+# qhasm: iv = arg3
+# asm 1: mov  <arg3=int64#3,>iv=int64#3
+# asm 2: mov  <arg3=%rdx,>iv=%rdx
+mov  %rdx,%rdx
+
+# qhasm: k = arg4
+# asm 1: mov  <arg4=int64#4,>k=int64#8
+# asm 2: mov  <arg4=%rcx,>k=%r10
+mov  %rcx,%r10
+
+# qhasm:               unsigned>? bytes - 0
+# asm 1: cmp  $0,<bytes=int64#6
+# asm 2: cmp  $0,<bytes=%r9
+cmp  $0,%r9
+# comment:fp stack unchanged by jump
+
+# qhasm: goto done if !unsigned>
+jbe ._done
+
+# qhasm: a = 0
+# asm 1: mov  $0,>a=int64#7
+# asm 2: mov  $0,>a=%rax
+mov  $0,%rax
+
+# qhasm: i = bytes
+# asm 1: mov  <bytes=int64#6,>i=int64#4
+# asm 2: mov  <bytes=%r9,>i=%rcx
+mov  %r9,%rcx
+
+# qhasm: while (i) { *out++ = a; --i }
+rep stosb
+
+# qhasm: out -= bytes
+# asm 1: sub  <bytes=int64#6,<out=int64#1
+# asm 2: sub  <bytes=%r9,<out=%rdi
+sub  %r9,%rdi
+# comment:fp stack unchanged by jump
+
+# qhasm: goto start
+jmp ._start
+
+# qhasm: enter crypto_stream_salsa2012_amd64_xmm6_xor
+.text
+.p2align 5
+.globl _crypto_stream_salsa2012_amd64_xmm6_xor
+.globl crypto_stream_salsa2012_amd64_xmm6_xor
+_crypto_stream_salsa2012_amd64_xmm6_xor:
+crypto_stream_salsa2012_amd64_xmm6_xor:
+mov %rsp,%r11
+and $31,%r11
+add $480,%r11
+sub %r11,%rsp
+
+# qhasm: r11_stack = r11_caller
+# asm 1: movq <r11_caller=int64#9,>r11_stack=stack64#1
+# asm 2: movq <r11_caller=%r11,>r11_stack=352(%rsp)
+movq %r11,352(%rsp)
+
+# qhasm: r12_stack = r12_caller
+# asm 1: movq <r12_caller=int64#10,>r12_stack=stack64#2
+# asm 2: movq <r12_caller=%r12,>r12_stack=360(%rsp)
+movq %r12,360(%rsp)
+
+# qhasm: r13_stack = r13_caller
+# asm 1: movq <r13_caller=int64#11,>r13_stack=stack64#3
+# asm 2: movq <r13_caller=%r13,>r13_stack=368(%rsp)
+movq %r13,368(%rsp)
+
+# qhasm: r14_stack = r14_caller
+# asm 1: movq <r14_caller=int64#12,>r14_stack=stack64#4
+# asm 2: movq <r14_caller=%r14,>r14_stack=376(%rsp)
+movq %r14,376(%rsp)
+
+# qhasm: r15_stack = r15_caller
+# asm 1: movq <r15_caller=int64#13,>r15_stack=stack64#5
+# asm 2: movq <r15_caller=%r15,>r15_stack=384(%rsp)
+movq %r15,384(%rsp)
+
+# qhasm: rbx_stack = rbx_caller
+# asm 1: movq <rbx_caller=int64#14,>rbx_stack=stack64#6
+# asm 2: movq <rbx_caller=%rbx,>rbx_stack=392(%rsp)
+movq %rbx,392(%rsp)
+
+# qhasm: rbp_stack = rbp_caller
+# asm 1: movq <rbp_caller=int64#15,>rbp_stack=stack64#7
+# asm 2: movq <rbp_caller=%rbp,>rbp_stack=400(%rsp)
+movq %rbp,400(%rsp)
+
+# qhasm: out = arg1
+# asm 1: mov  <arg1=int64#1,>out=int64#1
+# asm 2: mov  <arg1=%rdi,>out=%rdi
+mov  %rdi,%rdi
+
+# qhasm: m = arg2
+# asm 1: mov  <arg2=int64#2,>m=int64#2
+# asm 2: mov  <arg2=%rsi,>m=%rsi
+mov  %rsi,%rsi
+
+# qhasm: bytes = arg3
+# asm 1: mov  <arg3=int64#3,>bytes=int64#6
+# asm 2: mov  <arg3=%rdx,>bytes=%r9
+mov  %rdx,%r9
+
+# qhasm: iv = arg4
+# asm 1: mov  <arg4=int64#4,>iv=int64#3
+# asm 2: mov  <arg4=%rcx,>iv=%rdx
+mov  %rcx,%rdx
+
+# qhasm: k = arg5
+# asm 1: mov  <arg5=int64#5,>k=int64#8
+# asm 2: mov  <arg5=%r8,>k=%r10
+mov  %r8,%r10
+
+# qhasm:               unsigned>? bytes - 0
+# asm 1: cmp  $0,<bytes=int64#6
+# asm 2: cmp  $0,<bytes=%r9
+cmp  $0,%r9
+# comment:fp stack unchanged by jump
+
+# qhasm: goto done if !unsigned>
+jbe ._done
+# comment:fp stack unchanged by fallthrough
+
+# qhasm: start:
+._start:
+
+# qhasm:   in12 = *(uint32 *) (k + 20)
+# asm 1: movl   20(<k=int64#8),>in12=int64#4d
+# asm 2: movl   20(<k=%r10),>in12=%ecx
+movl   20(%r10),%ecx
+
+# qhasm:   in1 = *(uint32 *) (k + 0)
+# asm 1: movl   0(<k=int64#8),>in1=int64#5d
+# asm 2: movl   0(<k=%r10),>in1=%r8d
+movl   0(%r10),%r8d
+
+# qhasm:   in6 = *(uint32 *) (iv + 0)
+# asm 1: movl   0(<iv=int64#3),>in6=int64#7d
+# asm 2: movl   0(<iv=%rdx),>in6=%eax
+movl   0(%rdx),%eax
+
+# qhasm:   in11 = *(uint32 *) (k + 16)
+# asm 1: movl   16(<k=int64#8),>in11=int64#9d
+# asm 2: movl   16(<k=%r10),>in11=%r11d
+movl   16(%r10),%r11d
+
+# qhasm:   ((uint32 *)&x1)[0] = in12
+# asm 1: movl <in12=int64#4d,>x1=stack128#1
+# asm 2: movl <in12=%ecx,>x1=0(%rsp)
+movl %ecx,0(%rsp)
+
+# qhasm:   ((uint32 *)&x1)[1] = in1
+# asm 1: movl <in1=int64#5d,4+<x1=stack128#1
+# asm 2: movl <in1=%r8d,4+<x1=0(%rsp)
+movl %r8d,4+0(%rsp)
+
+# qhasm:   ((uint32 *)&x1)[2] = in6
+# asm 1: movl <in6=int64#7d,8+<x1=stack128#1
+# asm 2: movl <in6=%eax,8+<x1=0(%rsp)
+movl %eax,8+0(%rsp)
+
+# qhasm:   ((uint32 *)&x1)[3] = in11
+# asm 1: movl <in11=int64#9d,12+<x1=stack128#1
+# asm 2: movl <in11=%r11d,12+<x1=0(%rsp)
+movl %r11d,12+0(%rsp)
+
+# qhasm:   in8 = 0
+# asm 1: mov  $0,>in8=int64#4
+# asm 2: mov  $0,>in8=%rcx
+mov  $0,%rcx
+
+# qhasm:   in13 = *(uint32 *) (k + 24)
+# asm 1: movl   24(<k=int64#8),>in13=int64#5d
+# asm 2: movl   24(<k=%r10),>in13=%r8d
+movl   24(%r10),%r8d
+
+# qhasm:   in2 = *(uint32 *) (k + 4)
+# asm 1: movl   4(<k=int64#8),>in2=int64#7d
+# asm 2: movl   4(<k=%r10),>in2=%eax
+movl   4(%r10),%eax
+
+# qhasm:   in7 = *(uint32 *) (iv + 4)
+# asm 1: movl   4(<iv=int64#3),>in7=int64#3d
+# asm 2: movl   4(<iv=%rdx),>in7=%edx
+movl   4(%rdx),%edx
+
+# qhasm:   ((uint32 *)&x2)[0] = in8
+# asm 1: movl <in8=int64#4d,>x2=stack128#2
+# asm 2: movl <in8=%ecx,>x2=16(%rsp)
+movl %ecx,16(%rsp)
+
+# qhasm:   ((uint32 *)&x2)[1] = in13
+# asm 1: movl <in13=int64#5d,4+<x2=stack128#2
+# asm 2: movl <in13=%r8d,4+<x2=16(%rsp)
+movl %r8d,4+16(%rsp)
+
+# qhasm:   ((uint32 *)&x2)[2] = in2
+# asm 1: movl <in2=int64#7d,8+<x2=stack128#2
+# asm 2: movl <in2=%eax,8+<x2=16(%rsp)
+movl %eax,8+16(%rsp)
+
+# qhasm:   ((uint32 *)&x2)[3] = in7
+# asm 1: movl <in7=int64#3d,12+<x2=stack128#2
+# asm 2: movl <in7=%edx,12+<x2=16(%rsp)
+movl %edx,12+16(%rsp)
+
+# qhasm:   in4 = *(uint32 *) (k + 12)
+# asm 1: movl   12(<k=int64#8),>in4=int64#3d
+# asm 2: movl   12(<k=%r10),>in4=%edx
+movl   12(%r10),%edx
+
+# qhasm:   in9 = 0
+# asm 1: mov  $0,>in9=int64#4
+# asm 2: mov  $0,>in9=%rcx
+mov  $0,%rcx
+
+# qhasm:   in14 = *(uint32 *) (k + 28)
+# asm 1: movl   28(<k=int64#8),>in14=int64#5d
+# asm 2: movl   28(<k=%r10),>in14=%r8d
+movl   28(%r10),%r8d
+
+# qhasm:   in3 = *(uint32 *) (k + 8)
+# asm 1: movl   8(<k=int64#8),>in3=int64#7d
+# asm 2: movl   8(<k=%r10),>in3=%eax
+movl   8(%r10),%eax
+
+# qhasm:   ((uint32 *)&x3)[0] = in4
+# asm 1: movl <in4=int64#3d,>x3=stack128#3
+# asm 2: movl <in4=%edx,>x3=32(%rsp)
+movl %edx,32(%rsp)
+
+# qhasm:   ((uint32 *)&x3)[1] = in9
+# asm 1: movl <in9=int64#4d,4+<x3=stack128#3
+# asm 2: movl <in9=%ecx,4+<x3=32(%rsp)
+movl %ecx,4+32(%rsp)
+
+# qhasm:   ((uint32 *)&x3)[2] = in14
+# asm 1: movl <in14=int64#5d,8+<x3=stack128#3
+# asm 2: movl <in14=%r8d,8+<x3=32(%rsp)
+movl %r8d,8+32(%rsp)
+
+# qhasm:   ((uint32 *)&x3)[3] = in3
+# asm 1: movl <in3=int64#7d,12+<x3=stack128#3
+# asm 2: movl <in3=%eax,12+<x3=32(%rsp)
+movl %eax,12+32(%rsp)
+
+# qhasm:   in0 = 1634760805
+# asm 1: mov  $1634760805,>in0=int64#3
+# asm 2: mov  $1634760805,>in0=%rdx
+mov  $1634760805,%rdx
+
+# qhasm:   in5 = 857760878
+# asm 1: mov  $857760878,>in5=int64#4
+# asm 2: mov  $857760878,>in5=%rcx
+mov  $857760878,%rcx
+
+# qhasm:   in10 = 2036477234
+# asm 1: mov  $2036477234,>in10=int64#5
+# asm 2: mov  $2036477234,>in10=%r8
+mov  $2036477234,%r8
+
+# qhasm:   in15 = 1797285236
+# asm 1: mov  $1797285236,>in15=int64#7
+# asm 2: mov  $1797285236,>in15=%rax
+mov  $1797285236,%rax
+
+# qhasm:   ((uint32 *)&x0)[0] = in0
+# asm 1: movl <in0=int64#3d,>x0=stack128#4
+# asm 2: movl <in0=%edx,>x0=48(%rsp)
+movl %edx,48(%rsp)
+
+# qhasm:   ((uint32 *)&x0)[1] = in5
+# asm 1: movl <in5=int64#4d,4+<x0=stack128#4
+# asm 2: movl <in5=%ecx,4+<x0=48(%rsp)
+movl %ecx,4+48(%rsp)
+
+# qhasm:   ((uint32 *)&x0)[2] = in10
+# asm 1: movl <in10=int64#5d,8+<x0=stack128#4
+# asm 2: movl <in10=%r8d,8+<x0=48(%rsp)
+movl %r8d,8+48(%rsp)
+
+# qhasm:   ((uint32 *)&x0)[3] = in15
+# asm 1: movl <in15=int64#7d,12+<x0=stack128#4
+# asm 2: movl <in15=%eax,12+<x0=48(%rsp)
+movl %eax,12+48(%rsp)
+
+# qhasm:                               unsigned<? bytes - 256
+# asm 1: cmp  $256,<bytes=int64#6
+# asm 2: cmp  $256,<bytes=%r9
+cmp  $256,%r9
+# comment:fp stack unchanged by jump
+
+# qhasm:   goto bytesbetween1and255 if unsigned<
+jb ._bytesbetween1and255
+
+# qhasm:   z0 = x0
+# asm 1: movdqa <x0=stack128#4,>z0=int6464#1
+# asm 2: movdqa <x0=48(%rsp),>z0=%xmm0
+movdqa 48(%rsp),%xmm0
+
+# qhasm:   z5 = z0[1,1,1,1]
+# asm 1: pshufd $0x55,<z0=int6464#1,>z5=int6464#2
+# asm 2: pshufd $0x55,<z0=%xmm0,>z5=%xmm1
+pshufd $0x55,%xmm0,%xmm1
+
+# qhasm:   z10 = z0[2,2,2,2]
+# asm 1: pshufd $0xaa,<z0=int6464#1,>z10=int6464#3
+# asm 2: pshufd $0xaa,<z0=%xmm0,>z10=%xmm2
+pshufd $0xaa,%xmm0,%xmm2
+
+# qhasm:   z15 = z0[3,3,3,3]
+# asm 1: pshufd $0xff,<z0=int6464#1,>z15=int6464#4
+# asm 2: pshufd $0xff,<z0=%xmm0,>z15=%xmm3
+pshufd $0xff,%xmm0,%xmm3
+
+# qhasm:   z0 = z0[0,0,0,0]
+# asm 1: pshufd $0x00,<z0=int6464#1,>z0=int6464#1
+# asm 2: pshufd $0x00,<z0=%xmm0,>z0=%xmm0
+pshufd $0x00,%xmm0,%xmm0
+
+# qhasm:   orig5 = z5
+# asm 1: movdqa <z5=int6464#2,>orig5=stack128#5
+# asm 2: movdqa <z5=%xmm1,>orig5=64(%rsp)
+movdqa %xmm1,64(%rsp)
+
+# qhasm:   orig10 = z10
+# asm 1: movdqa <z10=int6464#3,>orig10=stack128#6
+# asm 2: movdqa <z10=%xmm2,>orig10=80(%rsp)
+movdqa %xmm2,80(%rsp)
+
+# qhasm:   orig15 = z15
+# asm 1: movdqa <z15=int6464#4,>orig15=stack128#7
+# asm 2: movdqa <z15=%xmm3,>orig15=96(%rsp)
+movdqa %xmm3,96(%rsp)
+
+# qhasm:   orig0 = z0
+# asm 1: movdqa <z0=int6464#1,>orig0=stack128#8
+# asm 2: movdqa <z0=%xmm0,>orig0=112(%rsp)
+movdqa %xmm0,112(%rsp)
+
+# qhasm:   z1 = x1
+# asm 1: movdqa <x1=stack128#1,>z1=int6464#1
+# asm 2: movdqa <x1=0(%rsp),>z1=%xmm0
+movdqa 0(%rsp),%xmm0
+
+# qhasm:   z6 = z1[2,2,2,2]
+# asm 1: pshufd $0xaa,<z1=int6464#1,>z6=int6464#2
+# asm 2: pshufd $0xaa,<z1=%xmm0,>z6=%xmm1
+pshufd $0xaa,%xmm0,%xmm1
+
+# qhasm:   z11 = z1[3,3,3,3]
+# asm 1: pshufd $0xff,<z1=int6464#1,>z11=int6464#3
+# asm 2: pshufd $0xff,<z1=%xmm0,>z11=%xmm2
+pshufd $0xff,%xmm0,%xmm2
+
+# qhasm:   z12 = z1[0,0,0,0]
+# asm 1: pshufd $0x00,<z1=int6464#1,>z12=int6464#4
+# asm 2: pshufd $0x00,<z1=%xmm0,>z12=%xmm3
+pshufd $0x00,%xmm0,%xmm3
+
+# qhasm:   z1 = z1[1,1,1,1]
+# asm 1: pshufd $0x55,<z1=int6464#1,>z1=int6464#1
+# asm 2: pshufd $0x55,<z1=%xmm0,>z1=%xmm0
+pshufd $0x55,%xmm0,%xmm0
+
+# qhasm:   orig6 = z6
+# asm 1: movdqa <z6=int6464#2,>orig6=stack128#9
+# asm 2: movdqa <z6=%xmm1,>orig6=128(%rsp)
+movdqa %xmm1,128(%rsp)
+
+# qhasm:   orig11 = z11
+# asm 1: movdqa <z11=int6464#3,>orig11=stack128#10
+# asm 2: movdqa <z11=%xmm2,>orig11=144(%rsp)
+movdqa %xmm2,144(%rsp)
+
+# qhasm:   orig12 = z12
+# asm 1: movdqa <z12=int6464#4,>orig12=stack128#11
+# asm 2: movdqa <z12=%xmm3,>orig12=160(%rsp)
+movdqa %xmm3,160(%rsp)
+
+# qhasm:   orig1 = z1
+# asm 1: movdqa <z1=int6464#1,>orig1=stack128#12
+# asm 2: movdqa <z1=%xmm0,>orig1=176(%rsp)
+movdqa %xmm0,176(%rsp)
+
+# qhasm:   z2 = x2
+# asm 1: movdqa <x2=stack128#2,>z2=int6464#1
+# asm 2: movdqa <x2=16(%rsp),>z2=%xmm0
+movdqa 16(%rsp),%xmm0
+
+# qhasm:   z7 = z2[3,3,3,3]
+# asm 1: pshufd $0xff,<z2=int6464#1,>z7=int6464#2
+# asm 2: pshufd $0xff,<z2=%xmm0,>z7=%xmm1
+pshufd $0xff,%xmm0,%xmm1
+
+# qhasm:   z13 = z2[1,1,1,1]
+# asm 1: pshufd $0x55,<z2=int6464#1,>z13=int6464#3
+# asm 2: pshufd $0x55,<z2=%xmm0,>z13=%xmm2
+pshufd $0x55,%xmm0,%xmm2
+
+# qhasm:   z2 = z2[2,2,2,2]
+# asm 1: pshufd $0xaa,<z2=int6464#1,>z2=int6464#1
+# asm 2: pshufd $0xaa,<z2=%xmm0,>z2=%xmm0
+pshufd $0xaa,%xmm0,%xmm0
+
+# qhasm:   orig7 = z7
+# asm 1: movdqa <z7=int6464#2,>orig7=stack128#13
+# asm 2: movdqa <z7=%xmm1,>orig7=192(%rsp)
+movdqa %xmm1,192(%rsp)
+
+# qhasm:   orig13 = z13
+# asm 1: movdqa <z13=int6464#3,>orig13=stack128#14
+# asm 2: movdqa <z13=%xmm2,>orig13=208(%rsp)
+movdqa %xmm2,208(%rsp)
+
+# qhasm:   orig2 = z2
+# asm 1: movdqa <z2=int6464#1,>orig2=stack128#15
+# asm 2: movdqa <z2=%xmm0,>orig2=224(%rsp)
+movdqa %xmm0,224(%rsp)
+
+# qhasm:   z3 = x3
+# asm 1: movdqa <x3=stack128#3,>z3=int6464#1
+# asm 2: movdqa <x3=32(%rsp),>z3=%xmm0
+movdqa 32(%rsp),%xmm0
+
+# qhasm:   z4 = z3[0,0,0,0]
+# asm 1: pshufd $0x00,<z3=int6464#1,>z4=int6464#2
+# asm 2: pshufd $0x00,<z3=%xmm0,>z4=%xmm1
+pshufd $0x00,%xmm0,%xmm1
+
+# qhasm:   z14 = z3[2,2,2,2]
+# asm 1: pshufd $0xaa,<z3=int6464#1,>z14=int6464#3
+# asm 2: pshufd $0xaa,<z3=%xmm0,>z14=%xmm2
+pshufd $0xaa,%xmm0,%xmm2
+
+# qhasm:   z3 = z3[3,3,3,3]
+# asm 1: pshufd $0xff,<z3=int6464#1,>z3=int6464#1
+# asm 2: pshufd $0xff,<z3=%xmm0,>z3=%xmm0
+pshufd $0xff,%xmm0,%xmm0
+
+# qhasm:   orig4 = z4
+# asm 1: movdqa <z4=int6464#2,>orig4=stack128#16
+# asm 2: movdqa <z4=%xmm1,>orig4=240(%rsp)
+movdqa %xmm1,240(%rsp)
+
+# qhasm:   orig14 = z14
+# asm 1: movdqa <z14=int6464#3,>orig14=stack128#17
+# asm 2: movdqa <z14=%xmm2,>orig14=256(%rsp)
+movdqa %xmm2,256(%rsp)
+
+# qhasm:   orig3 = z3
+# asm 1: movdqa <z3=int6464#1,>orig3=stack128#18
+# asm 2: movdqa <z3=%xmm0,>orig3=272(%rsp)
+movdqa %xmm0,272(%rsp)
+
+# qhasm: bytesatleast256:
+._bytesatleast256:
+
+# qhasm:   in8 = ((uint32 *)&x2)[0]
+# asm 1: movl <x2=stack128#2,>in8=int64#3d
+# asm 2: movl <x2=16(%rsp),>in8=%edx
+movl 16(%rsp),%edx
+
+# qhasm:   in9 = ((uint32 *)&x3)[1]
+# asm 1: movl 4+<x3=stack128#3,>in9=int64#4d
+# asm 2: movl 4+<x3=32(%rsp),>in9=%ecx
+movl 4+32(%rsp),%ecx
+
+# qhasm:   ((uint32 *) &orig8)[0] = in8
+# asm 1: movl <in8=int64#3d,>orig8=stack128#19
+# asm 2: movl <in8=%edx,>orig8=288(%rsp)
+movl %edx,288(%rsp)
+
+# qhasm:   ((uint32 *) &orig9)[0] = in9
+# asm 1: movl <in9=int64#4d,>orig9=stack128#20
+# asm 2: movl <in9=%ecx,>orig9=304(%rsp)
+movl %ecx,304(%rsp)
+
+# qhasm:   in8 += 1
+# asm 1: add  $1,<in8=int64#3
+# asm 2: add  $1,<in8=%rdx
+add  $1,%rdx
+
+# qhasm:   in9 <<= 32
+# asm 1: shl  $32,<in9=int64#4
+# asm 2: shl  $32,<in9=%rcx
+shl  $32,%rcx
+
+# qhasm:   in8 += in9
+# asm 1: add  <in9=int64#4,<in8=int64#3
+# asm 2: add  <in9=%rcx,<in8=%rdx
+add  %rcx,%rdx
+
+# qhasm:   in9 = in8
+# asm 1: mov  <in8=int64#3,>in9=int64#4
+# asm 2: mov  <in8=%rdx,>in9=%rcx
+mov  %rdx,%rcx
+
+# qhasm:   (uint64) in9 >>= 32
+# asm 1: shr  $32,<in9=int64#4
+# asm 2: shr  $32,<in9=%rcx
+shr  $32,%rcx
+
+# qhasm:   ((uint32 *) &orig8)[1] = in8
+# asm 1: movl <in8=int64#3d,4+<orig8=stack128#19
+# asm 2: movl <in8=%edx,4+<orig8=288(%rsp)
+movl %edx,4+288(%rsp)
+
+# qhasm:   ((uint32 *) &orig9)[1] = in9
+# asm 1: movl <in9=int64#4d,4+<orig9=stack128#20
+# asm 2: movl <in9=%ecx,4+<orig9=304(%rsp)
+movl %ecx,4+304(%rsp)
+
+# qhasm:   in8 += 1
+# asm 1: add  $1,<in8=int64#3
+# asm 2: add  $1,<in8=%rdx
+add  $1,%rdx
+
+# qhasm:   in9 <<= 32
+# asm 1: shl  $32,<in9=int64#4
+# asm 2: shl  $32,<in9=%rcx
+shl  $32,%rcx
+
+# qhasm:   in8 += in9
+# asm 1: add  <in9=int64#4,<in8=int64#3
+# asm 2: add  <in9=%rcx,<in8=%rdx
+add  %rcx,%rdx
+
+# qhasm:   in9 = in8
+# asm 1: mov  <in8=int64#3,>in9=int64#4
+# asm 2: mov  <in8=%rdx,>in9=%rcx
+mov  %rdx,%rcx
+
+# qhasm:   (uint64) in9 >>= 32
+# asm 1: shr  $32,<in9=int64#4
+# asm 2: shr  $32,<in9=%rcx
+shr  $32,%rcx
+
+# qhasm:   ((uint32 *) &orig8)[2] = in8
+# asm 1: movl <in8=int64#3d,8+<orig8=stack128#19
+# asm 2: movl <in8=%edx,8+<orig8=288(%rsp)
+movl %edx,8+288(%rsp)
+
+# qhasm:   ((uint32 *) &orig9)[2] = in9
+# asm 1: movl <in9=int64#4d,8+<orig9=stack128#20
+# asm 2: movl <in9=%ecx,8+<orig9=304(%rsp)
+movl %ecx,8+304(%rsp)
+
+# qhasm:   in8 += 1
+# asm 1: add  $1,<in8=int64#3
+# asm 2: add  $1,<in8=%rdx
+add  $1,%rdx
+
+# qhasm:   in9 <<= 32
+# asm 1: shl  $32,<in9=int64#4
+# asm 2: shl  $32,<in9=%rcx
+shl  $32,%rcx
+
+# qhasm:   in8 += in9
+# asm 1: add  <in9=int64#4,<in8=int64#3
+# asm 2: add  <in9=%rcx,<in8=%rdx
+add  %rcx,%rdx
+
+# qhasm:   in9 = in8
+# asm 1: mov  <in8=int64#3,>in9=int64#4
+# asm 2: mov  <in8=%rdx,>in9=%rcx
+mov  %rdx,%rcx
+
+# qhasm:   (uint64) in9 >>= 32
+# asm 1: shr  $32,<in9=int64#4
+# asm 2: shr  $32,<in9=%rcx
+shr  $32,%rcx
+
+# qhasm:   ((uint32 *) &orig8)[3] = in8
+# asm 1: movl <in8=int64#3d,12+<orig8=stack128#19
+# asm 2: movl <in8=%edx,12+<orig8=288(%rsp)
+movl %edx,12+288(%rsp)
+
+# qhasm:   ((uint32 *) &orig9)[3] = in9
+# asm 1: movl <in9=int64#4d,12+<orig9=stack128#20
+# asm 2: movl <in9=%ecx,12+<orig9=304(%rsp)
+movl %ecx,12+304(%rsp)
+
+# qhasm:   in8 += 1
+# asm 1: add  $1,<in8=int64#3
+# asm 2: add  $1,<in8=%rdx
+add  $1,%rdx
+
+# qhasm:   in9 <<= 32
+# asm 1: shl  $32,<in9=int64#4
+# asm 2: shl  $32,<in9=%rcx
+shl  $32,%rcx
+
+# qhasm:   in8 += in9
+# asm 1: add  <in9=int64#4,<in8=int64#3
+# asm 2: add  <in9=%rcx,<in8=%rdx
+add  %rcx,%rdx
+
+# qhasm:   in9 = in8
+# asm 1: mov  <in8=int64#3,>in9=int64#4
+# asm 2: mov  <in8=%rdx,>in9=%rcx
+mov  %rdx,%rcx
+
+# qhasm:   (uint64) in9 >>= 32
+# asm 1: shr  $32,<in9=int64#4
+# asm 2: shr  $32,<in9=%rcx
+shr  $32,%rcx
+
+# qhasm:   ((uint32 *)&x2)[0] = in8
+# asm 1: movl <in8=int64#3d,>x2=stack128#2
+# asm 2: movl <in8=%edx,>x2=16(%rsp)
+movl %edx,16(%rsp)
+
+# qhasm:   ((uint32 *)&x3)[1] = in9
+# asm 1: movl <in9=int64#4d,4+<x3=stack128#3
+# asm 2: movl <in9=%ecx,4+<x3=32(%rsp)
+movl %ecx,4+32(%rsp)
+
+# qhasm:   bytes_backup = bytes
+# asm 1: movq <bytes=int64#6,>bytes_backup=stack64#8
+# asm 2: movq <bytes=%r9,>bytes_backup=408(%rsp)
+movq %r9,408(%rsp)
+
+# qhasm: i = 12
+# asm 1: mov  $12,>i=int64#3
+# asm 2: mov  $12,>i=%rdx
+mov  $12,%rdx
+
+# qhasm:   z5 = orig5
+# asm 1: movdqa <orig5=stack128#5,>z5=int6464#1
+# asm 2: movdqa <orig5=64(%rsp),>z5=%xmm0
+movdqa 64(%rsp),%xmm0
+
+# qhasm:   z10 = orig10
+# asm 1: movdqa <orig10=stack128#6,>z10=int6464#2
+# asm 2: movdqa <orig10=80(%rsp),>z10=%xmm1
+movdqa 80(%rsp),%xmm1
+
+# qhasm:   z15 = orig15
+# asm 1: movdqa <orig15=stack128#7,>z15=int6464#3
+# asm 2: movdqa <orig15=96(%rsp),>z15=%xmm2
+movdqa 96(%rsp),%xmm2
+
+# qhasm:   z14 = orig14
+# asm 1: movdqa <orig14=stack128#17,>z14=int6464#4
+# asm 2: movdqa <orig14=256(%rsp),>z14=%xmm3
+movdqa 256(%rsp),%xmm3
+
+# qhasm:   z3 = orig3
+# asm 1: movdqa <orig3=stack128#18,>z3=int6464#5
+# asm 2: movdqa <orig3=272(%rsp),>z3=%xmm4
+movdqa 272(%rsp),%xmm4
+
+# qhasm:   z6 = orig6
+# asm 1: movdqa <orig6=stack128#9,>z6=int6464#6
+# asm 2: movdqa <orig6=128(%rsp),>z6=%xmm5
+movdqa 128(%rsp),%xmm5
+
+# qhasm:   z11 = orig11
+# asm 1: movdqa <orig11=stack128#10,>z11=int6464#7
+# asm 2: movdqa <orig11=144(%rsp),>z11=%xmm6
+movdqa 144(%rsp),%xmm6
+
+# qhasm:   z1 = orig1
+# asm 1: movdqa <orig1=stack128#12,>z1=int6464#8
+# asm 2: movdqa <orig1=176(%rsp),>z1=%xmm7
+movdqa 176(%rsp),%xmm7
+
+# qhasm:   z7 = orig7
+# asm 1: movdqa <orig7=stack128#13,>z7=int6464#9
+# asm 2: movdqa <orig7=192(%rsp),>z7=%xmm8
+movdqa 192(%rsp),%xmm8
+
+# qhasm:   z13 = orig13
+# asm 1: movdqa <orig13=stack128#14,>z13=int6464#10
+# asm 2: movdqa <orig13=208(%rsp),>z13=%xmm9
+movdqa 208(%rsp),%xmm9
+
+# qhasm:   z2 = orig2
+# asm 1: movdqa <orig2=stack128#15,>z2=int6464#11
+# asm 2: movdqa <orig2=224(%rsp),>z2=%xmm10
+movdqa 224(%rsp),%xmm10
+
+# qhasm:   z9 = orig9
+# asm 1: movdqa <orig9=stack128#20,>z9=int6464#12
+# asm 2: movdqa <orig9=304(%rsp),>z9=%xmm11
+movdqa 304(%rsp),%xmm11
+
+# qhasm:   z0 = orig0
+# asm 1: movdqa <orig0=stack128#8,>z0=int6464#13
+# asm 2: movdqa <orig0=112(%rsp),>z0=%xmm12
+movdqa 112(%rsp),%xmm12
+
+# qhasm:   z12 = orig12
+# asm 1: movdqa <orig12=stack128#11,>z12=int6464#14
+# asm 2: movdqa <orig12=160(%rsp),>z12=%xmm13
+movdqa 160(%rsp),%xmm13
+
+# qhasm:   z4 = orig4
+# asm 1: movdqa <orig4=stack128#16,>z4=int6464#15
+# asm 2: movdqa <orig4=240(%rsp),>z4=%xmm14
+movdqa 240(%rsp),%xmm14
+
+# qhasm:   z8 = orig8
+# asm 1: movdqa <orig8=stack128#19,>z8=int6464#16
+# asm 2: movdqa <orig8=288(%rsp),>z8=%xmm15
+movdqa 288(%rsp),%xmm15
+
+# qhasm: mainloop1:
+._mainloop1:
+
+# qhasm: 						z10_stack = z10
+# asm 1: movdqa <z10=int6464#2,>z10_stack=stack128#21
+# asm 2: movdqa <z10=%xmm1,>z10_stack=320(%rsp)
+movdqa %xmm1,320(%rsp)
+
+# qhasm: 								z15_stack = z15
+# asm 1: movdqa <z15=int6464#3,>z15_stack=stack128#22
+# asm 2: movdqa <z15=%xmm2,>z15_stack=336(%rsp)
+movdqa %xmm2,336(%rsp)
+
+# qhasm: 		y4 = z12
+# asm 1: movdqa <z12=int6464#14,>y4=int6464#2
+# asm 2: movdqa <z12=%xmm13,>y4=%xmm1
+movdqa %xmm13,%xmm1
+
+# qhasm: uint32323232	y4 += z0
+# asm 1: paddd <z0=int6464#13,<y4=int6464#2
+# asm 2: paddd <z0=%xmm12,<y4=%xmm1
+paddd %xmm12,%xmm1
+
+# qhasm: 		r4 = y4
+# asm 1: movdqa <y4=int6464#2,>r4=int6464#3
+# asm 2: movdqa <y4=%xmm1,>r4=%xmm2
+movdqa %xmm1,%xmm2
+
+# qhasm: uint32323232	y4 <<= 7
+# asm 1: pslld $7,<y4=int6464#2
+# asm 2: pslld $7,<y4=%xmm1
+pslld $7,%xmm1
+
+# qhasm: 		z4 ^= y4
+# asm 1: pxor  <y4=int6464#2,<z4=int6464#15
+# asm 2: pxor  <y4=%xmm1,<z4=%xmm14
+pxor  %xmm1,%xmm14
+
+# qhasm: uint32323232	r4 >>= 25
+# asm 1: psrld $25,<r4=int6464#3
+# asm 2: psrld $25,<r4=%xmm2
+psrld $25,%xmm2
+
+# qhasm: 		z4 ^= r4
+# asm 1: pxor  <r4=int6464#3,<z4=int6464#15
+# asm 2: pxor  <r4=%xmm2,<z4=%xmm14
+pxor  %xmm2,%xmm14
+
+# qhasm: 				y9 = z1
+# asm 1: movdqa <z1=int6464#8,>y9=int6464#2
+# asm 2: movdqa <z1=%xmm7,>y9=%xmm1
+movdqa %xmm7,%xmm1
+
+# qhasm: uint32323232			y9 += z5
+# asm 1: paddd <z5=int6464#1,<y9=int6464#2
+# asm 2: paddd <z5=%xmm0,<y9=%xmm1
+paddd %xmm0,%xmm1
+
+# qhasm: 				r9 = y9
+# asm 1: movdqa <y9=int6464#2,>r9=int6464#3
+# asm 2: movdqa <y9=%xmm1,>r9=%xmm2
+movdqa %xmm1,%xmm2
+
+# qhasm: uint32323232			y9 <<= 7
+# asm 1: pslld $7,<y9=int6464#2
+# asm 2: pslld $7,<y9=%xmm1
+pslld $7,%xmm1
+
+# qhasm: 				z9 ^= y9
+# asm 1: pxor  <y9=int6464#2,<z9=int6464#12
+# asm 2: pxor  <y9=%xmm1,<z9=%xmm11
+pxor  %xmm1,%xmm11
+
+# qhasm: uint32323232			r9 >>= 25
+# asm 1: psrld $25,<r9=int6464#3
+# asm 2: psrld $25,<r9=%xmm2
+psrld $25,%xmm2
+
+# qhasm: 				z9 ^= r9
+# asm 1: pxor  <r9=int6464#3,<z9=int6464#12
+# asm 2: pxor  <r9=%xmm2,<z9=%xmm11
+pxor  %xmm2,%xmm11
+
+# qhasm: 		y8 = z0
+# asm 1: movdqa <z0=int6464#13,>y8=int6464#2
+# asm 2: movdqa <z0=%xmm12,>y8=%xmm1
+movdqa %xmm12,%xmm1
+
+# qhasm: uint32323232	y8 += z4
+# asm 1: paddd <z4=int6464#15,<y8=int6464#2
+# asm 2: paddd <z4=%xmm14,<y8=%xmm1
+paddd %xmm14,%xmm1
+
+# qhasm: 		r8 = y8
+# asm 1: movdqa <y8=int6464#2,>r8=int6464#3
+# asm 2: movdqa <y8=%xmm1,>r8=%xmm2
+movdqa %xmm1,%xmm2
+
+# qhasm: uint32323232	y8 <<= 9
+# asm 1: pslld $9,<y8=int6464#2
+# asm 2: pslld $9,<y8=%xmm1
+pslld $9,%xmm1
+
+# qhasm: 		z8 ^= y8
+# asm 1: pxor  <y8=int6464#2,<z8=int6464#16
+# asm 2: pxor  <y8=%xmm1,<z8=%xmm15
+pxor  %xmm1,%xmm15
+
+# qhasm: uint32323232	r8 >>= 23
+# asm 1: psrld $23,<r8=int6464#3
+# asm 2: psrld $23,<r8=%xmm2
+psrld $23,%xmm2
+
+# qhasm: 		z8 ^= r8
+# asm 1: pxor  <r8=int6464#3,<z8=int6464#16
+# asm 2: pxor  <r8=%xmm2,<z8=%xmm15
+pxor  %xmm2,%xmm15
+
+# qhasm: 				y13 = z5
+# asm 1: movdqa <z5=int6464#1,>y13=int6464#2
+# asm 2: movdqa <z5=%xmm0,>y13=%xmm1
+movdqa %xmm0,%xmm1
+
+# qhasm: uint32323232			y13 += z9
+# asm 1: paddd <z9=int6464#12,<y13=int6464#2
+# asm 2: paddd <z9=%xmm11,<y13=%xmm1
+paddd %xmm11,%xmm1
+
+# qhasm: 				r13 = y13
+# asm 1: movdqa <y13=int6464#2,>r13=int6464#3
+# asm 2: movdqa <y13=%xmm1,>r13=%xmm2
+movdqa %xmm1,%xmm2
+
+# qhasm: uint32323232			y13 <<= 9
+# asm 1: pslld $9,<y13=int6464#2
+# asm 2: pslld $9,<y13=%xmm1
+pslld $9,%xmm1
+
+# qhasm: 				z13 ^= y13
+# asm 1: pxor  <y13=int6464#2,<z13=int6464#10
+# asm 2: pxor  <y13=%xmm1,<z13=%xmm9
+pxor  %xmm1,%xmm9
+
+# qhasm: uint32323232			r13 >>= 23
+# asm 1: psrld $23,<r13=int6464#3
+# asm 2: psrld $23,<r13=%xmm2
+psrld $23,%xmm2
+
+# qhasm: 				z13 ^= r13
+# asm 1: pxor  <r13=int6464#3,<z13=int6464#10
+# asm 2: pxor  <r13=%xmm2,<z13=%xmm9
+pxor  %xmm2,%xmm9
+
+# qhasm: 		y12 = z4
+# asm 1: movdqa <z4=int6464#15,>y12=int6464#2
+# asm 2: movdqa <z4=%xmm14,>y12=%xmm1
+movdqa %xmm14,%xmm1
+
+# qhasm: uint32323232	y12 += z8
+# asm 1: paddd <z8=int6464#16,<y12=int6464#2
+# asm 2: paddd <z8=%xmm15,<y12=%xmm1
+paddd %xmm15,%xmm1
+
+# qhasm: 		r12 = y12
+# asm 1: movdqa <y12=int6464#2,>r12=int6464#3
+# asm 2: movdqa <y12=%xmm1,>r12=%xmm2
+movdqa %xmm1,%xmm2
+
+# qhasm: uint32323232	y12 <<= 13
+# asm 1: pslld $13,<y12=int6464#2
+# asm 2: pslld $13,<y12=%xmm1
+pslld $13,%xmm1
+
+# qhasm: 		z12 ^= y12
+# asm 1: pxor  <y12=int6464#2,<z12=int6464#14
+# asm 2: pxor  <y12=%xmm1,<z12=%xmm13
+pxor  %xmm1,%xmm13
+
+# qhasm: uint32323232	r12 >>= 19
+# asm 1: psrld $19,<r12=int6464#3
+# asm 2: psrld $19,<r12=%xmm2
+psrld $19,%xmm2
+
+# qhasm: 		z12 ^= r12
+# asm 1: pxor  <r12=int6464#3,<z12=int6464#14
+# asm 2: pxor  <r12=%xmm2,<z12=%xmm13
+pxor  %xmm2,%xmm13
+
+# qhasm: 				y1 = z9
+# asm 1: movdqa <z9=int6464#12,>y1=int6464#2
+# asm 2: movdqa <z9=%xmm11,>y1=%xmm1
+movdqa %xmm11,%xmm1
+
+# qhasm: uint32323232			y1 += z13
+# asm 1: paddd <z13=int6464#10,<y1=int6464#2
+# asm 2: paddd <z13=%xmm9,<y1=%xmm1
+paddd %xmm9,%xmm1
+
+# qhasm: 				r1 = y1
+# asm 1: movdqa <y1=int6464#2,>r1=int6464#3
+# asm 2: movdqa <y1=%xmm1,>r1=%xmm2
+movdqa %xmm1,%xmm2
+
+# qhasm: uint32323232			y1 <<= 13
+# asm 1: pslld $13,<y1=int6464#2
+# asm 2: pslld $13,<y1=%xmm1
+pslld $13,%xmm1
+
+# qhasm: 				z1 ^= y1
+# asm 1: pxor  <y1=int6464#2,<z1=int6464#8
+# asm 2: pxor  <y1=%xmm1,<z1=%xmm7
+pxor  %xmm1,%xmm7
+
+# qhasm: uint32323232			r1 >>= 19
+# asm 1: psrld $19,<r1=int6464#3
+# asm 2: psrld $19,<r1=%xmm2
+psrld $19,%xmm2
+
+# qhasm: 				z1 ^= r1
+# asm 1: pxor  <r1=int6464#3,<z1=int6464#8
+# asm 2: pxor  <r1=%xmm2,<z1=%xmm7
+pxor  %xmm2,%xmm7
+
+# qhasm: 		y0 = z8
+# asm 1: movdqa <z8=int6464#16,>y0=int6464#2
+# asm 2: movdqa <z8=%xmm15,>y0=%xmm1
+movdqa %xmm15,%xmm1
+
+# qhasm: uint32323232	y0 += z12
+# asm 1: paddd <z12=int6464#14,<y0=int6464#2
+# asm 2: paddd <z12=%xmm13,<y0=%xmm1
+paddd %xmm13,%xmm1
+
+# qhasm: 		r0 = y0
+# asm 1: movdqa <y0=int6464#2,>r0=int6464#3
+# asm 2: movdqa <y0=%xmm1,>r0=%xmm2
+movdqa %xmm1,%xmm2
+
+# qhasm: uint32323232	y0 <<= 18
+# asm 1: pslld $18,<y0=int6464#2
+# asm 2: pslld $18,<y0=%xmm1
+pslld $18,%xmm1
+
+# qhasm: 		z0 ^= y0
+# asm 1: pxor  <y0=int6464#2,<z0=int6464#13
+# asm 2: pxor  <y0=%xmm1,<z0=%xmm12
+pxor  %xmm1,%xmm12
+
+# qhasm: uint32323232	r0 >>= 14
+# asm 1: psrld $14,<r0=int6464#3
+# asm 2: psrld $14,<r0=%xmm2
+psrld $14,%xmm2
+
+# qhasm: 		z0 ^= r0
+# asm 1: pxor  <r0=int6464#3,<z0=int6464#13
+# asm 2: pxor  <r0=%xmm2,<z0=%xmm12
+pxor  %xmm2,%xmm12
+
+# qhasm: 						z10 = z10_stack
+# asm 1: movdqa <z10_stack=stack128#21,>z10=int6464#2
+# asm 2: movdqa <z10_stack=320(%rsp),>z10=%xmm1
+movdqa 320(%rsp),%xmm1
+
+# qhasm: 		z0_stack = z0
+# asm 1: movdqa <z0=int6464#13,>z0_stack=stack128#21
+# asm 2: movdqa <z0=%xmm12,>z0_stack=320(%rsp)
+movdqa %xmm12,320(%rsp)
+
+# qhasm: 				y5 = z13
+# asm 1: movdqa <z13=int6464#10,>y5=int6464#3
+# asm 2: movdqa <z13=%xmm9,>y5=%xmm2
+movdqa %xmm9,%xmm2
+
+# qhasm: uint32323232			y5 += z1
+# asm 1: paddd <z1=int6464#8,<y5=int6464#3
+# asm 2: paddd <z1=%xmm7,<y5=%xmm2
+paddd %xmm7,%xmm2
+
+# qhasm: 				r5 = y5
+# asm 1: movdqa <y5=int6464#3,>r5=int6464#13
+# asm 2: movdqa <y5=%xmm2,>r5=%xmm12
+movdqa %xmm2,%xmm12
+
+# qhasm: uint32323232			y5 <<= 18
+# asm 1: pslld $18,<y5=int6464#3
+# asm 2: pslld $18,<y5=%xmm2
+pslld $18,%xmm2
+
+# qhasm: 				z5 ^= y5
+# asm 1: pxor  <y5=int6464#3,<z5=int6464#1
+# asm 2: pxor  <y5=%xmm2,<z5=%xmm0
+pxor  %xmm2,%xmm0
+
+# qhasm: uint32323232			r5 >>= 14
+# asm 1: psrld $14,<r5=int6464#13
+# asm 2: psrld $14,<r5=%xmm12
+psrld $14,%xmm12
+
+# qhasm: 				z5 ^= r5
+# asm 1: pxor  <r5=int6464#13,<z5=int6464#1
+# asm 2: pxor  <r5=%xmm12,<z5=%xmm0
+pxor  %xmm12,%xmm0
+
+# qhasm: 						y14 = z6
+# asm 1: movdqa <z6=int6464#6,>y14=int6464#3
+# asm 2: movdqa <z6=%xmm5,>y14=%xmm2
+movdqa %xmm5,%xmm2
+
+# qhasm: uint32323232					y14 += z10
+# asm 1: paddd <z10=int6464#2,<y14=int6464#3
+# asm 2: paddd <z10=%xmm1,<y14=%xmm2
+paddd %xmm1,%xmm2
+
+# qhasm: 						r14 = y14
+# asm 1: movdqa <y14=int6464#3,>r14=int6464#13
+# asm 2: movdqa <y14=%xmm2,>r14=%xmm12
+movdqa %xmm2,%xmm12
+
+# qhasm: uint32323232					y14 <<= 7
+# asm 1: pslld $7,<y14=int6464#3
+# asm 2: pslld $7,<y14=%xmm2
+pslld $7,%xmm2
+
+# qhasm: 						z14 ^= y14
+# asm 1: pxor  <y14=int6464#3,<z14=int6464#4
+# asm 2: pxor  <y14=%xmm2,<z14=%xmm3
+pxor  %xmm2,%xmm3
+
+# qhasm: uint32323232					r14 >>= 25
+# asm 1: psrld $25,<r14=int6464#13
+# asm 2: psrld $25,<r14=%xmm12
+psrld $25,%xmm12
+
+# qhasm: 						z14 ^= r14
+# asm 1: pxor  <r14=int6464#13,<z14=int6464#4
+# asm 2: pxor  <r14=%xmm12,<z14=%xmm3
+pxor  %xmm12,%xmm3
+
+# qhasm: 								z15 = z15_stack
+# asm 1: movdqa <z15_stack=stack128#22,>z15=int6464#3
+# asm 2: movdqa <z15_stack=336(%rsp),>z15=%xmm2
+movdqa 336(%rsp),%xmm2
+
+# qhasm: 				z5_stack = z5
+# asm 1: movdqa <z5=int6464#1,>z5_stack=stack128#22
+# asm 2: movdqa <z5=%xmm0,>z5_stack=336(%rsp)
+movdqa %xmm0,336(%rsp)
+
+# qhasm: 								y3 = z11
+# asm 1: movdqa <z11=int6464#7,>y3=int6464#1
+# asm 2: movdqa <z11=%xmm6,>y3=%xmm0
+movdqa %xmm6,%xmm0
+
+# qhasm: uint32323232							y3 += z15
+# asm 1: paddd <z15=int6464#3,<y3=int6464#1
+# asm 2: paddd <z15=%xmm2,<y3=%xmm0
+paddd %xmm2,%xmm0
+
+# qhasm: 								r3 = y3
+# asm 1: movdqa <y3=int6464#1,>r3=int6464#13
+# asm 2: movdqa <y3=%xmm0,>r3=%xmm12
+movdqa %xmm0,%xmm12
+
+# qhasm: uint32323232							y3 <<= 7
+# asm 1: pslld $7,<y3=int6464#1
+# asm 2: pslld $7,<y3=%xmm0
+pslld $7,%xmm0
+
+# qhasm: 								z3 ^= y3
+# asm 1: pxor  <y3=int6464#1,<z3=int6464#5
+# asm 2: pxor  <y3=%xmm0,<z3=%xmm4
+pxor  %xmm0,%xmm4
+
+# qhasm: uint32323232							r3 >>= 25
+# asm 1: psrld $25,<r3=int6464#13
+# asm 2: psrld $25,<r3=%xmm12
+psrld $25,%xmm12
+
+# qhasm: 								z3 ^= r3
+# asm 1: pxor  <r3=int6464#13,<z3=int6464#5
+# asm 2: pxor  <r3=%xmm12,<z3=%xmm4
+pxor  %xmm12,%xmm4
+
+# qhasm: 						y2 = z10
+# asm 1: movdqa <z10=int6464#2,>y2=int6464#1
+# asm 2: movdqa <z10=%xmm1,>y2=%xmm0
+movdqa %xmm1,%xmm0
+
+# qhasm: uint32323232					y2 += z14
+# asm 1: paddd <z14=int6464#4,<y2=int6464#1
+# asm 2: paddd <z14=%xmm3,<y2=%xmm0
+paddd %xmm3,%xmm0
+
+# qhasm: 						r2 = y2
+# asm 1: movdqa <y2=int6464#1,>r2=int6464#13
+# asm 2: movdqa <y2=%xmm0,>r2=%xmm12
+movdqa %xmm0,%xmm12
+
+# qhasm: uint32323232					y2 <<= 9
+# asm 1: pslld $9,<y2=int6464#1
+# asm 2: pslld $9,<y2=%xmm0
+pslld $9,%xmm0
+
+# qhasm: 						z2 ^= y2
+# asm 1: pxor  <y2=int6464#1,<z2=int6464#11
+# asm 2: pxor  <y2=%xmm0,<z2=%xmm10
+pxor  %xmm0,%xmm10
+
+# qhasm: uint32323232					r2 >>= 23
+# asm 1: psrld $23,<r2=int6464#13
+# asm 2: psrld $23,<r2=%xmm12
+psrld $23,%xmm12
+
+# qhasm: 						z2 ^= r2
+# asm 1: pxor  <r2=int6464#13,<z2=int6464#11
+# asm 2: pxor  <r2=%xmm12,<z2=%xmm10
+pxor  %xmm12,%xmm10
+
+# qhasm: 								y7 = z15
+# asm 1: movdqa <z15=int6464#3,>y7=int6464#1
+# asm 2: movdqa <z15=%xmm2,>y7=%xmm0
+movdqa %xmm2,%xmm0
+
+# qhasm: uint32323232							y7 += z3
+# asm 1: paddd <z3=int6464#5,<y7=int6464#1
+# asm 2: paddd <z3=%xmm4,<y7=%xmm0
+paddd %xmm4,%xmm0
+
+# qhasm: 								r7 = y7
+# asm 1: movdqa <y7=int6464#1,>r7=int6464#13
+# asm 2: movdqa <y7=%xmm0,>r7=%xmm12
+movdqa %xmm0,%xmm12
+
+# qhasm: uint32323232							y7 <<= 9
+# asm 1: pslld $9,<y7=int6464#1
+# asm 2: pslld $9,<y7=%xmm0
+pslld $9,%xmm0
+
+# qhasm: 								z7 ^= y7
+# asm 1: pxor  <y7=int6464#1,<z7=int6464#9
+# asm 2: pxor  <y7=%xmm0,<z7=%xmm8
+pxor  %xmm0,%xmm8
+
+# qhasm: uint32323232							r7 >>= 23
+# asm 1: psrld $23,<r7=int6464#13
+# asm 2: psrld $23,<r7=%xmm12
+psrld $23,%xmm12
+
+# qhasm: 								z7 ^= r7
+# asm 1: pxor  <r7=int6464#13,<z7=int6464#9
+# asm 2: pxor  <r7=%xmm12,<z7=%xmm8
+pxor  %xmm12,%xmm8
+
+# qhasm: 						y6 = z14
+# asm 1: movdqa <z14=int6464#4,>y6=int6464#1
+# asm 2: movdqa <z14=%xmm3,>y6=%xmm0
+movdqa %xmm3,%xmm0
+
+# qhasm: uint32323232					y6 += z2
+# asm 1: paddd <z2=int6464#11,<y6=int6464#1
+# asm 2: paddd <z2=%xmm10,<y6=%xmm0
+paddd %xmm10,%xmm0
+
+# qhasm: 						r6 = y6
+# asm 1: movdqa <y6=int6464#1,>r6=int6464#13
+# asm 2: movdqa <y6=%xmm0,>r6=%xmm12
+movdqa %xmm0,%xmm12
+
+# qhasm: uint32323232					y6 <<= 13
+# asm 1: pslld $13,<y6=int6464#1
+# asm 2: pslld $13,<y6=%xmm0
+pslld $13,%xmm0
+
+# qhasm: 						z6 ^= y6
+# asm 1: pxor  <y6=int6464#1,<z6=int6464#6
+# asm 2: pxor  <y6=%xmm0,<z6=%xmm5
+pxor  %xmm0,%xmm5
+
+# qhasm: uint32323232					r6 >>= 19
+# asm 1: psrld $19,<r6=int6464#13
+# asm 2: psrld $19,<r6=%xmm12
+psrld $19,%xmm12
+
+# qhasm: 						z6 ^= r6
+# asm 1: pxor  <r6=int6464#13,<z6=int6464#6
+# asm 2: pxor  <r6=%xmm12,<z6=%xmm5
+pxor  %xmm12,%xmm5
+
+# qhasm: 								y11 = z3
+# asm 1: movdqa <z3=int6464#5,>y11=int6464#1
+# asm 2: movdqa <z3=%xmm4,>y11=%xmm0
+movdqa %xmm4,%xmm0
+
+# qhasm: uint32323232							y11 += z7
+# asm 1: paddd <z7=int6464#9,<y11=int6464#1
+# asm 2: paddd <z7=%xmm8,<y11=%xmm0
+paddd %xmm8,%xmm0
+
+# qhasm: 								r11 = y11
+# asm 1: movdqa <y11=int6464#1,>r11=int6464#13
+# asm 2: movdqa <y11=%xmm0,>r11=%xmm12
+movdqa %xmm0,%xmm12
+
+# qhasm: uint32323232							y11 <<= 13
+# asm 1: pslld $13,<y11=int6464#1
+# asm 2: pslld $13,<y11=%xmm0
+pslld $13,%xmm0
+
+# qhasm: 								z11 ^= y11
+# asm 1: pxor  <y11=int6464#1,<z11=int6464#7
+# asm 2: pxor  <y11=%xmm0,<z11=%xmm6
+pxor  %xmm0,%xmm6
+
+# qhasm: uint32323232							r11 >>= 19
+# asm 1: psrld $19,<r11=int6464#13
+# asm 2: psrld $19,<r11=%xmm12
+psrld $19,%xmm12
+
+# qhasm: 								z11 ^= r11
+# asm 1: pxor  <r11=int6464#13,<z11=int6464#7
+# asm 2: pxor  <r11=%xmm12,<z11=%xmm6
+pxor  %xmm12,%xmm6
+
+# qhasm: 						y10 = z2
+# asm 1: movdqa <z2=int6464#11,>y10=int6464#1
+# asm 2: movdqa <z2=%xmm10,>y10=%xmm0
+movdqa %xmm10,%xmm0
+
+# qhasm: uint32323232					y10 += z6
+# asm 1: paddd <z6=int6464#6,<y10=int6464#1
+# asm 2: paddd <z6=%xmm5,<y10=%xmm0
+paddd %xmm5,%xmm0
+
+# qhasm: 						r10 = y10
+# asm 1: movdqa <y10=int6464#1,>r10=int6464#13
+# asm 2: movdqa <y10=%xmm0,>r10=%xmm12
+movdqa %xmm0,%xmm12
+
+# qhasm: uint32323232					y10 <<= 18
+# asm 1: pslld $18,<y10=int6464#1
+# asm 2: pslld $18,<y10=%xmm0
+pslld $18,%xmm0
+
+# qhasm: 						z10 ^= y10
+# asm 1: pxor  <y10=int6464#1,<z10=int6464#2
+# asm 2: pxor  <y10=%xmm0,<z10=%xmm1
+pxor  %xmm0,%xmm1
+
+# qhasm: uint32323232					r10 >>= 14
+# asm 1: psrld $14,<r10=int6464#13
+# asm 2: psrld $14,<r10=%xmm12
+psrld $14,%xmm12
+
+# qhasm: 						z10 ^= r10
+# asm 1: pxor  <r10=int6464#13,<z10=int6464#2
+# asm 2: pxor  <r10=%xmm12,<z10=%xmm1
+pxor  %xmm12,%xmm1
+
+# qhasm: 		z0 = z0_stack
+# asm 1: movdqa <z0_stack=stack128#21,>z0=int6464#1
+# asm 2: movdqa <z0_stack=320(%rsp),>z0=%xmm0
+movdqa 320(%rsp),%xmm0
+
+# qhasm: 						z10_stack = z10
+# asm 1: movdqa <z10=int6464#2,>z10_stack=stack128#21
+# asm 2: movdqa <z10=%xmm1,>z10_stack=320(%rsp)
+movdqa %xmm1,320(%rsp)
+
+# qhasm: 		y1 = z3
+# asm 1: movdqa <z3=int6464#5,>y1=int6464#2
+# asm 2: movdqa <z3=%xmm4,>y1=%xmm1
+movdqa %xmm4,%xmm1
+
+# qhasm: uint32323232	y1 += z0
+# asm 1: paddd <z0=int6464#1,<y1=int6464#2
+# asm 2: paddd <z0=%xmm0,<y1=%xmm1
+paddd %xmm0,%xmm1
+
+# qhasm: 		r1 = y1
+# asm 1: movdqa <y1=int6464#2,>r1=int6464#13
+# asm 2: movdqa <y1=%xmm1,>r1=%xmm12
+movdqa %xmm1,%xmm12
+
+# qhasm: uint32323232	y1 <<= 7
+# asm 1: pslld $7,<y1=int6464#2
+# asm 2: pslld $7,<y1=%xmm1
+pslld $7,%xmm1
+
+# qhasm: 		z1 ^= y1
+# asm 1: pxor  <y1=int6464#2,<z1=int6464#8
+# asm 2: pxor  <y1=%xmm1,<z1=%xmm7
+pxor  %xmm1,%xmm7
+
+# qhasm: uint32323232	r1 >>= 25
+# asm 1: psrld $25,<r1=int6464#13
+# asm 2: psrld $25,<r1=%xmm12
+psrld $25,%xmm12
+
+# qhasm: 		z1 ^= r1
+# asm 1: pxor  <r1=int6464#13,<z1=int6464#8
+# asm 2: pxor  <r1=%xmm12,<z1=%xmm7
+pxor  %xmm12,%xmm7
+
+# qhasm: 								y15 = z7
+# asm 1: movdqa <z7=int6464#9,>y15=int6464#2
+# asm 2: movdqa <z7=%xmm8,>y15=%xmm1
+movdqa %xmm8,%xmm1
+
+# qhasm: uint32323232							y15 += z11
+# asm 1: paddd <z11=int6464#7,<y15=int6464#2
+# asm 2: paddd <z11=%xmm6,<y15=%xmm1
+paddd %xmm6,%xmm1
+
+# qhasm: 								r15 = y15
+# asm 1: movdqa <y15=int6464#2,>r15=int6464#13
+# asm 2: movdqa <y15=%xmm1,>r15=%xmm12
+movdqa %xmm1,%xmm12
+
+# qhasm: uint32323232							y15 <<= 18
+# asm 1: pslld $18,<y15=int6464#2
+# asm 2: pslld $18,<y15=%xmm1
+pslld $18,%xmm1
+
+# qhasm: 								z15 ^= y15
+# asm 1: pxor  <y15=int6464#2,<z15=int6464#3
+# asm 2: pxor  <y15=%xmm1,<z15=%xmm2
+pxor  %xmm1,%xmm2
+
+# qhasm: uint32323232							r15 >>= 14
+# asm 1: psrld $14,<r15=int6464#13
+# asm 2: psrld $14,<r15=%xmm12
+psrld $14,%xmm12
+
+# qhasm: 								z15 ^= r15
+# asm 1: pxor  <r15=int6464#13,<z15=int6464#3
+# asm 2: pxor  <r15=%xmm12,<z15=%xmm2
+pxor  %xmm12,%xmm2
+
+# qhasm: 				z5 = z5_stack
+# asm 1: movdqa <z5_stack=stack128#22,>z5=int6464#13
+# asm 2: movdqa <z5_stack=336(%rsp),>z5=%xmm12
+movdqa 336(%rsp),%xmm12
+
+# qhasm: 								z15_stack = z15
+# asm 1: movdqa <z15=int6464#3,>z15_stack=stack128#22
+# asm 2: movdqa <z15=%xmm2,>z15_stack=336(%rsp)
+movdqa %xmm2,336(%rsp)
+
+# qhasm: 				y6 = z4
+# asm 1: movdqa <z4=int6464#15,>y6=int6464#2
+# asm 2: movdqa <z4=%xmm14,>y6=%xmm1
+movdqa %xmm14,%xmm1
+
+# qhasm: uint32323232			y6 += z5
+# asm 1: paddd <z5=int6464#13,<y6=int6464#2
+# asm 2: paddd <z5=%xmm12,<y6=%xmm1
+paddd %xmm12,%xmm1
+
+# qhasm: 				r6 = y6
+# asm 1: movdqa <y6=int6464#2,>r6=int6464#3
+# asm 2: movdqa <y6=%xmm1,>r6=%xmm2
+movdqa %xmm1,%xmm2
+
+# qhasm: uint32323232			y6 <<= 7
+# asm 1: pslld $7,<y6=int6464#2
+# asm 2: pslld $7,<y6=%xmm1
+pslld $7,%xmm1
+
+# qhasm: 				z6 ^= y6
+# asm 1: pxor  <y6=int6464#2,<z6=int6464#6
+# asm 2: pxor  <y6=%xmm1,<z6=%xmm5
+pxor  %xmm1,%xmm5
+
+# qhasm: uint32323232			r6 >>= 25
+# asm 1: psrld $25,<r6=int6464#3
+# asm 2: psrld $25,<r6=%xmm2
+psrld $25,%xmm2
+
+# qhasm: 				z6 ^= r6
+# asm 1: pxor  <r6=int6464#3,<z6=int6464#6
+# asm 2: pxor  <r6=%xmm2,<z6=%xmm5
+pxor  %xmm2,%xmm5
+
+# qhasm: 		y2 = z0
+# asm 1: movdqa <z0=int6464#1,>y2=int6464#2
+# asm 2: movdqa <z0=%xmm0,>y2=%xmm1
+movdqa %xmm0,%xmm1
+
+# qhasm: uint32323232	y2 += z1
+# asm 1: paddd <z1=int6464#8,<y2=int6464#2
+# asm 2: paddd <z1=%xmm7,<y2=%xmm1
+paddd %xmm7,%xmm1
+
+# qhasm: 		r2 = y2
+# asm 1: movdqa <y2=int6464#2,>r2=int6464#3
+# asm 2: movdqa <y2=%xmm1,>r2=%xmm2
+movdqa %xmm1,%xmm2
+
+# qhasm: uint32323232	y2 <<= 9
+# asm 1: pslld $9,<y2=int6464#2
+# asm 2: pslld $9,<y2=%xmm1
+pslld $9,%xmm1
+
+# qhasm: 		z2 ^= y2
+# asm 1: pxor  <y2=int6464#2,<z2=int6464#11
+# asm 2: pxor  <y2=%xmm1,<z2=%xmm10
+pxor  %xmm1,%xmm10
+
+# qhasm: uint32323232	r2 >>= 23
+# asm 1: psrld $23,<r2=int6464#3
+# asm 2: psrld $23,<r2=%xmm2
+psrld $23,%xmm2
+
+# qhasm: 		z2 ^= r2
+# asm 1: pxor  <r2=int6464#3,<z2=int6464#11
+# asm 2: pxor  <r2=%xmm2,<z2=%xmm10
+pxor  %xmm2,%xmm10
+
+# qhasm: 				y7 = z5
+# asm 1: movdqa <z5=int6464#13,>y7=int6464#2
+# asm 2: movdqa <z5=%xmm12,>y7=%xmm1
+movdqa %xmm12,%xmm1
+
+# qhasm: uint32323232			y7 += z6
+# asm 1: paddd <z6=int6464#6,<y7=int6464#2
+# asm 2: paddd <z6=%xmm5,<y7=%xmm1
+paddd %xmm5,%xmm1
+
+# qhasm: 				r7 = y7
+# asm 1: movdqa <y7=int6464#2,>r7=int6464#3
+# asm 2: movdqa <y7=%xmm1,>r7=%xmm2
+movdqa %xmm1,%xmm2
+
+# qhasm: uint32323232			y7 <<= 9
+# asm 1: pslld $9,<y7=int6464#2
+# asm 2: pslld $9,<y7=%xmm1
+pslld $9,%xmm1
+
+# qhasm: 				z7 ^= y7
+# asm 1: pxor  <y7=int6464#2,<z7=int6464#9
+# asm 2: pxor  <y7=%xmm1,<z7=%xmm8
+pxor  %xmm1,%xmm8
+
+# qhasm: uint32323232			r7 >>= 23
+# asm 1: psrld $23,<r7=int6464#3
+# asm 2: psrld $23,<r7=%xmm2
+psrld $23,%xmm2
+
+# qhasm: 				z7 ^= r7
+# asm 1: pxor  <r7=int6464#3,<z7=int6464#9
+# asm 2: pxor  <r7=%xmm2,<z7=%xmm8
+pxor  %xmm2,%xmm8
+
+# qhasm: 		y3 = z1
+# asm 1: movdqa <z1=int6464#8,>y3=int6464#2
+# asm 2: movdqa <z1=%xmm7,>y3=%xmm1
+movdqa %xmm7,%xmm1
+
+# qhasm: uint32323232	y3 += z2
+# asm 1: paddd <z2=int6464#11,<y3=int6464#2
+# asm 2: paddd <z2=%xmm10,<y3=%xmm1
+paddd %xmm10,%xmm1
+
+# qhasm: 		r3 = y3
+# asm 1: movdqa <y3=int6464#2,>r3=int6464#3
+# asm 2: movdqa <y3=%xmm1,>r3=%xmm2
+movdqa %xmm1,%xmm2
+
+# qhasm: uint32323232	y3 <<= 13
+# asm 1: pslld $13,<y3=int6464#2
+# asm 2: pslld $13,<y3=%xmm1
+pslld $13,%xmm1
+
+# qhasm: 		z3 ^= y3
+# asm 1: pxor  <y3=int6464#2,<z3=int6464#5
+# asm 2: pxor  <y3=%xmm1,<z3=%xmm4
+pxor  %xmm1,%xmm4
+
+# qhasm: uint32323232	r3 >>= 19
+# asm 1: psrld $19,<r3=int6464#3
+# asm 2: psrld $19,<r3=%xmm2
+psrld $19,%xmm2
+
+# qhasm: 		z3 ^= r3
+# asm 1: pxor  <r3=int6464#3,<z3=int6464#5
+# asm 2: pxor  <r3=%xmm2,<z3=%xmm4
+pxor  %xmm2,%xmm4
+
+# qhasm: 				y4 = z6
+# asm 1: movdqa <z6=int6464#6,>y4=int6464#2
+# asm 2: movdqa <z6=%xmm5,>y4=%xmm1
+movdqa %xmm5,%xmm1
+
+# qhasm: uint32323232			y4 += z7
+# asm 1: paddd <z7=int6464#9,<y4=int6464#2
+# asm 2: paddd <z7=%xmm8,<y4=%xmm1
+paddd %xmm8,%xmm1
+
+# qhasm: 				r4 = y4
+# asm 1: movdqa <y4=int6464#2,>r4=int6464#3
+# asm 2: movdqa <y4=%xmm1,>r4=%xmm2
+movdqa %xmm1,%xmm2
+
+# qhasm: uint32323232			y4 <<= 13
+# asm 1: pslld $13,<y4=int6464#2
+# asm 2: pslld $13,<y4=%xmm1
+pslld $13,%xmm1
+
+# qhasm: 				z4 ^= y4
+# asm 1: pxor  <y4=int6464#2,<z4=int6464#15
+# asm 2: pxor  <y4=%xmm1,<z4=%xmm14
+pxor  %xmm1,%xmm14
+
+# qhasm: uint32323232			r4 >>= 19
+# asm 1: psrld $19,<r4=int6464#3
+# asm 2: psrld $19,<r4=%xmm2
+psrld $19,%xmm2
+
+# qhasm: 				z4 ^= r4
+# asm 1: pxor  <r4=int6464#3,<z4=int6464#15
+# asm 2: pxor  <r4=%xmm2,<z4=%xmm14
+pxor  %xmm2,%xmm14
+
+# qhasm: 		y0 = z2
+# asm 1: movdqa <z2=int6464#11,>y0=int6464#2
+# asm 2: movdqa <z2=%xmm10,>y0=%xmm1
+movdqa %xmm10,%xmm1
+
+# qhasm: uint32323232	y0 += z3
+# asm 1: paddd <z3=int6464#5,<y0=int6464#2
+# asm 2: paddd <z3=%xmm4,<y0=%xmm1
+paddd %xmm4,%xmm1
+
+# qhasm: 		r0 = y0
+# asm 1: movdqa <y0=int6464#2,>r0=int6464#3
+# asm 2: movdqa <y0=%xmm1,>r0=%xmm2
+movdqa %xmm1,%xmm2
+
+# qhasm: uint32323232	y0 <<= 18
+# asm 1: pslld $18,<y0=int6464#2
+# asm 2: pslld $18,<y0=%xmm1
+pslld $18,%xmm1
+
+# qhasm: 		z0 ^= y0
+# asm 1: pxor  <y0=int6464#2,<z0=int6464#1
+# asm 2: pxor  <y0=%xmm1,<z0=%xmm0
+pxor  %xmm1,%xmm0
+
+# qhasm: uint32323232	r0 >>= 14
+# asm 1: psrld $14,<r0=int6464#3
+# asm 2: psrld $14,<r0=%xmm2
+psrld $14,%xmm2
+
+# qhasm: 		z0 ^= r0
+# asm 1: pxor  <r0=int6464#3,<z0=int6464#1
+# asm 2: pxor  <r0=%xmm2,<z0=%xmm0
+pxor  %xmm2,%xmm0
+
+# qhasm: 						z10 = z10_stack
+# asm 1: movdqa <z10_stack=stack128#21,>z10=int6464#2
+# asm 2: movdqa <z10_stack=320(%rsp),>z10=%xmm1
+movdqa 320(%rsp),%xmm1
+
+# qhasm: 		z0_stack = z0
+# asm 1: movdqa <z0=int6464#1,>z0_stack=stack128#21
+# asm 2: movdqa <z0=%xmm0,>z0_stack=320(%rsp)
+movdqa %xmm0,320(%rsp)
+
+# qhasm: 				y5 = z7
+# asm 1: movdqa <z7=int6464#9,>y5=int6464#1
+# asm 2: movdqa <z7=%xmm8,>y5=%xmm0
+movdqa %xmm8,%xmm0
+
+# qhasm: uint32323232			y5 += z4
+# asm 1: paddd <z4=int6464#15,<y5=int6464#1
+# asm 2: paddd <z4=%xmm14,<y5=%xmm0
+paddd %xmm14,%xmm0
+
+# qhasm: 				r5 = y5
+# asm 1: movdqa <y5=int6464#1,>r5=int6464#3
+# asm 2: movdqa <y5=%xmm0,>r5=%xmm2
+movdqa %xmm0,%xmm2
+
+# qhasm: uint32323232			y5 <<= 18
+# asm 1: pslld $18,<y5=int6464#1
+# asm 2: pslld $18,<y5=%xmm0
+pslld $18,%xmm0
+
+# qhasm: 				z5 ^= y5
+# asm 1: pxor  <y5=int6464#1,<z5=int6464#13
+# asm 2: pxor  <y5=%xmm0,<z5=%xmm12
+pxor  %xmm0,%xmm12
+
+# qhasm: uint32323232			r5 >>= 14
+# asm 1: psrld $14,<r5=int6464#3
+# asm 2: psrld $14,<r5=%xmm2
+psrld $14,%xmm2
+
+# qhasm: 				z5 ^= r5
+# asm 1: pxor  <r5=int6464#3,<z5=int6464#13
+# asm 2: pxor  <r5=%xmm2,<z5=%xmm12
+pxor  %xmm2,%xmm12
+
+# qhasm: 						y11 = z9
+# asm 1: movdqa <z9=int6464#12,>y11=int6464#1
+# asm 2: movdqa <z9=%xmm11,>y11=%xmm0
+movdqa %xmm11,%xmm0
+
+# qhasm: uint32323232					y11 += z10
+# asm 1: paddd <z10=int6464#2,<y11=int6464#1
+# asm 2: paddd <z10=%xmm1,<y11=%xmm0
+paddd %xmm1,%xmm0
+
+# qhasm: 						r11 = y11
+# asm 1: movdqa <y11=int6464#1,>r11=int6464#3
+# asm 2: movdqa <y11=%xmm0,>r11=%xmm2
+movdqa %xmm0,%xmm2
+
+# qhasm: uint32323232					y11 <<= 7
+# asm 1: pslld $7,<y11=int6464#1
+# asm 2: pslld $7,<y11=%xmm0
+pslld $7,%xmm0
+
+# qhasm: 						z11 ^= y11
+# asm 1: pxor  <y11=int6464#1,<z11=int6464#7
+# asm 2: pxor  <y11=%xmm0,<z11=%xmm6
+pxor  %xmm0,%xmm6
+
+# qhasm: uint32323232					r11 >>= 25
+# asm 1: psrld $25,<r11=int6464#3
+# asm 2: psrld $25,<r11=%xmm2
+psrld $25,%xmm2
+
+# qhasm: 						z11 ^= r11
+# asm 1: pxor  <r11=int6464#3,<z11=int6464#7
+# asm 2: pxor  <r11=%xmm2,<z11=%xmm6
+pxor  %xmm2,%xmm6
+
+# qhasm: 								z15 = z15_stack
+# asm 1: movdqa <z15_stack=stack128#22,>z15=int6464#3
+# asm 2: movdqa <z15_stack=336(%rsp),>z15=%xmm2
+movdqa 336(%rsp),%xmm2
+
+# qhasm: 				z5_stack = z5
+# asm 1: movdqa <z5=int6464#13,>z5_stack=stack128#22
+# asm 2: movdqa <z5=%xmm12,>z5_stack=336(%rsp)
+movdqa %xmm12,336(%rsp)
+
+# qhasm: 								y12 = z14
+# asm 1: movdqa <z14=int6464#4,>y12=int6464#1
+# asm 2: movdqa <z14=%xmm3,>y12=%xmm0
+movdqa %xmm3,%xmm0
+
+# qhasm: uint32323232							y12 += z15
+# asm 1: paddd <z15=int6464#3,<y12=int6464#1
+# asm 2: paddd <z15=%xmm2,<y12=%xmm0
+paddd %xmm2,%xmm0
+
+# qhasm: 								r12 = y12
+# asm 1: movdqa <y12=int6464#1,>r12=int6464#13
+# asm 2: movdqa <y12=%xmm0,>r12=%xmm12
+movdqa %xmm0,%xmm12
+
+# qhasm: uint32323232							y12 <<= 7
+# asm 1: pslld $7,<y12=int6464#1
+# asm 2: pslld $7,<y12=%xmm0
+pslld $7,%xmm0
+
+# qhasm: 								z12 ^= y12
+# asm 1: pxor  <y12=int6464#1,<z12=int6464#14
+# asm 2: pxor  <y12=%xmm0,<z12=%xmm13
+pxor  %xmm0,%xmm13
+
+# qhasm: uint32323232							r12 >>= 25
+# asm 1: psrld $25,<r12=int6464#13
+# asm 2: psrld $25,<r12=%xmm12
+psrld $25,%xmm12
+
+# qhasm: 								z12 ^= r12
+# asm 1: pxor  <r12=int6464#13,<z12=int6464#14
+# asm 2: pxor  <r12=%xmm12,<z12=%xmm13
+pxor  %xmm12,%xmm13
+
+# qhasm: 						y8 = z10
+# asm 1: movdqa <z10=int6464#2,>y8=int6464#1
+# asm 2: movdqa <z10=%xmm1,>y8=%xmm0
+movdqa %xmm1,%xmm0
+
+# qhasm: uint32323232					y8 += z11
+# asm 1: paddd <z11=int6464#7,<y8=int6464#1
+# asm 2: paddd <z11=%xmm6,<y8=%xmm0
+paddd %xmm6,%xmm0
+
+# qhasm: 						r8 = y8
+# asm 1: movdqa <y8=int6464#1,>r8=int6464#13
+# asm 2: movdqa <y8=%xmm0,>r8=%xmm12
+movdqa %xmm0,%xmm12
+
+# qhasm: uint32323232					y8 <<= 9
+# asm 1: pslld $9,<y8=int6464#1
+# asm 2: pslld $9,<y8=%xmm0
+pslld $9,%xmm0
+
+# qhasm: 						z8 ^= y8
+# asm 1: pxor  <y8=int6464#1,<z8=int6464#16
+# asm 2: pxor  <y8=%xmm0,<z8=%xmm15
+pxor  %xmm0,%xmm15
+
+# qhasm: uint32323232					r8 >>= 23
+# asm 1: psrld $23,<r8=int6464#13
+# asm 2: psrld $23,<r8=%xmm12
+psrld $23,%xmm12
+
+# qhasm: 						z8 ^= r8
+# asm 1: pxor  <r8=int6464#13,<z8=int6464#16
+# asm 2: pxor  <r8=%xmm12,<z8=%xmm15
+pxor  %xmm12,%xmm15
+
+# qhasm: 								y13 = z15
+# asm 1: movdqa <z15=int6464#3,>y13=int6464#1
+# asm 2: movdqa <z15=%xmm2,>y13=%xmm0
+movdqa %xmm2,%xmm0
+
+# qhasm: uint32323232							y13 += z12
+# asm 1: paddd <z12=int6464#14,<y13=int6464#1
+# asm 2: paddd <z12=%xmm13,<y13=%xmm0
+paddd %xmm13,%xmm0
+
+# qhasm: 								r13 = y13
+# asm 1: movdqa <y13=int6464#1,>r13=int6464#13
+# asm 2: movdqa <y13=%xmm0,>r13=%xmm12
+movdqa %xmm0,%xmm12
+
+# qhasm: uint32323232							y13 <<= 9
+# asm 1: pslld $9,<y13=int6464#1
+# asm 2: pslld $9,<y13=%xmm0
+pslld $9,%xmm0
+
+# qhasm: 								z13 ^= y13
+# asm 1: pxor  <y13=int6464#1,<z13=int6464#10
+# asm 2: pxor  <y13=%xmm0,<z13=%xmm9
+pxor  %xmm0,%xmm9
+
+# qhasm: uint32323232							r13 >>= 23
+# asm 1: psrld $23,<r13=int6464#13
+# asm 2: psrld $23,<r13=%xmm12
+psrld $23,%xmm12
+
+# qhasm: 								z13 ^= r13
+# asm 1: pxor  <r13=int6464#13,<z13=int6464#10
+# asm 2: pxor  <r13=%xmm12,<z13=%xmm9
+pxor  %xmm12,%xmm9
+
+# qhasm: 						y9 = z11
+# asm 1: movdqa <z11=int6464#7,>y9=int6464#1
+# asm 2: movdqa <z11=%xmm6,>y9=%xmm0
+movdqa %xmm6,%xmm0
+
+# qhasm: uint32323232					y9 += z8
+# asm 1: paddd <z8=int6464#16,<y9=int6464#1
+# asm 2: paddd <z8=%xmm15,<y9=%xmm0
+paddd %xmm15,%xmm0
+
+# qhasm: 						r9 = y9
+# asm 1: movdqa <y9=int6464#1,>r9=int6464#13
+# asm 2: movdqa <y9=%xmm0,>r9=%xmm12
+movdqa %xmm0,%xmm12
+
+# qhasm: uint32323232					y9 <<= 13
+# asm 1: pslld $13,<y9=int6464#1
+# asm 2: pslld $13,<y9=%xmm0
+pslld $13,%xmm0
+
+# qhasm: 						z9 ^= y9
+# asm 1: pxor  <y9=int6464#1,<z9=int6464#12
+# asm 2: pxor  <y9=%xmm0,<z9=%xmm11
+pxor  %xmm0,%xmm11
+
+# qhasm: uint32323232					r9 >>= 19
+# asm 1: psrld $19,<r9=int6464#13
+# asm 2: psrld $19,<r9=%xmm12
+psrld $19,%xmm12
+
+# qhasm: 						z9 ^= r9
+# asm 1: pxor  <r9=int6464#13,<z9=int6464#12
+# asm 2: pxor  <r9=%xmm12,<z9=%xmm11
+pxor  %xmm12,%xmm11
+
+# qhasm: 								y14 = z12
+# asm 1: movdqa <z12=int6464#14,>y14=int6464#1
+# asm 2: movdqa <z12=%xmm13,>y14=%xmm0
+movdqa %xmm13,%xmm0
+
+# qhasm: uint32323232							y14 += z13
+# asm 1: paddd <z13=int6464#10,<y14=int6464#1
+# asm 2: paddd <z13=%xmm9,<y14=%xmm0
+paddd %xmm9,%xmm0
+
+# qhasm: 								r14 = y14
+# asm 1: movdqa <y14=int6464#1,>r14=int6464#13
+# asm 2: movdqa <y14=%xmm0,>r14=%xmm12
+movdqa %xmm0,%xmm12
+
+# qhasm: uint32323232							y14 <<= 13
+# asm 1: pslld $13,<y14=int6464#1
+# asm 2: pslld $13,<y14=%xmm0
+pslld $13,%xmm0
+
+# qhasm: 								z14 ^= y14
+# asm 1: pxor  <y14=int6464#1,<z14=int6464#4
+# asm 2: pxor  <y14=%xmm0,<z14=%xmm3
+pxor  %xmm0,%xmm3
+
+# qhasm: uint32323232							r14 >>= 19
+# asm 1: psrld $19,<r14=int6464#13
+# asm 2: psrld $19,<r14=%xmm12
+psrld $19,%xmm12
+
+# qhasm: 								z14 ^= r14
+# asm 1: pxor  <r14=int6464#13,<z14=int6464#4
+# asm 2: pxor  <r14=%xmm12,<z14=%xmm3
+pxor  %xmm12,%xmm3
+
+# qhasm: 						y10 = z8
+# asm 1: movdqa <z8=int6464#16,>y10=int6464#1
+# asm 2: movdqa <z8=%xmm15,>y10=%xmm0
+movdqa %xmm15,%xmm0
+
+# qhasm: uint32323232					y10 += z9
+# asm 1: paddd <z9=int6464#12,<y10=int6464#1
+# asm 2: paddd <z9=%xmm11,<y10=%xmm0
+paddd %xmm11,%xmm0
+
+# qhasm: 						r10 = y10
+# asm 1: movdqa <y10=int6464#1,>r10=int6464#13
+# asm 2: movdqa <y10=%xmm0,>r10=%xmm12
+movdqa %xmm0,%xmm12
+
+# qhasm: uint32323232					y10 <<= 18
+# asm 1: pslld $18,<y10=int6464#1
+# asm 2: pslld $18,<y10=%xmm0
+pslld $18,%xmm0
+
+# qhasm: 						z10 ^= y10
+# asm 1: pxor  <y10=int6464#1,<z10=int6464#2
+# asm 2: pxor  <y10=%xmm0,<z10=%xmm1
+pxor  %xmm0,%xmm1
+
+# qhasm: uint32323232					r10 >>= 14
+# asm 1: psrld $14,<r10=int6464#13
+# asm 2: psrld $14,<r10=%xmm12
+psrld $14,%xmm12
+
+# qhasm: 						z10 ^= r10
+# asm 1: pxor  <r10=int6464#13,<z10=int6464#2
+# asm 2: pxor  <r10=%xmm12,<z10=%xmm1
+pxor  %xmm12,%xmm1
+
+# qhasm: 								y15 = z13
+# asm 1: movdqa <z13=int6464#10,>y15=int6464#1
+# asm 2: movdqa <z13=%xmm9,>y15=%xmm0
+movdqa %xmm9,%xmm0
+
+# qhasm: uint32323232							y15 += z14
+# asm 1: paddd <z14=int6464#4,<y15=int6464#1
+# asm 2: paddd <z14=%xmm3,<y15=%xmm0
+paddd %xmm3,%xmm0
+
+# qhasm: 								r15 = y15
+# asm 1: movdqa <y15=int6464#1,>r15=int6464#13
+# asm 2: movdqa <y15=%xmm0,>r15=%xmm12
+movdqa %xmm0,%xmm12
+
+# qhasm: uint32323232							y15 <<= 18
+# asm 1: pslld $18,<y15=int6464#1
+# asm 2: pslld $18,<y15=%xmm0
+pslld $18,%xmm0
+
+# qhasm: 								z15 ^= y15
+# asm 1: pxor  <y15=int6464#1,<z15=int6464#3
+# asm 2: pxor  <y15=%xmm0,<z15=%xmm2
+pxor  %xmm0,%xmm2
+
+# qhasm: uint32323232							r15 >>= 14
+# asm 1: psrld $14,<r15=int6464#13
+# asm 2: psrld $14,<r15=%xmm12
+psrld $14,%xmm12
+
+# qhasm: 								z15 ^= r15
+# asm 1: pxor  <r15=int6464#13,<z15=int6464#3
+# asm 2: pxor  <r15=%xmm12,<z15=%xmm2
+pxor  %xmm12,%xmm2
+
+# qhasm: 		z0 = z0_stack
+# asm 1: movdqa <z0_stack=stack128#21,>z0=int6464#13
+# asm 2: movdqa <z0_stack=320(%rsp),>z0=%xmm12
+movdqa 320(%rsp),%xmm12
+
+# qhasm: 				z5 = z5_stack
+# asm 1: movdqa <z5_stack=stack128#22,>z5=int6464#1
+# asm 2: movdqa <z5_stack=336(%rsp),>z5=%xmm0
+movdqa 336(%rsp),%xmm0
+
+# qhasm:                   unsigned>? i -= 2
+# asm 1: sub  $2,<i=int64#3
+# asm 2: sub  $2,<i=%rdx
+sub  $2,%rdx
+# comment:fp stack unchanged by jump
+
+# qhasm: goto mainloop1 if unsigned>
+ja ._mainloop1
+
+# qhasm:   uint32323232 z0 += orig0
+# asm 1: paddd <orig0=stack128#8,<z0=int6464#13
+# asm 2: paddd <orig0=112(%rsp),<z0=%xmm12
+paddd 112(%rsp),%xmm12
+
+# qhasm:   uint32323232 z1 += orig1
+# asm 1: paddd <orig1=stack128#12,<z1=int6464#8
+# asm 2: paddd <orig1=176(%rsp),<z1=%xmm7
+paddd 176(%rsp),%xmm7
+
+# qhasm:   uint32323232 z2 += orig2
+# asm 1: paddd <orig2=stack128#15,<z2=int6464#11
+# asm 2: paddd <orig2=224(%rsp),<z2=%xmm10
+paddd 224(%rsp),%xmm10
+
+# qhasm:   uint32323232 z3 += orig3
+# asm 1: paddd <orig3=stack128#18,<z3=int6464#5
+# asm 2: paddd <orig3=272(%rsp),<z3=%xmm4
+paddd 272(%rsp),%xmm4
+
+# qhasm:   in0 = z0
+# asm 1: movd   <z0=int6464#13,>in0=int64#3
+# asm 2: movd   <z0=%xmm12,>in0=%rdx
+movd   %xmm12,%rdx
+
+# qhasm:   in1 = z1
+# asm 1: movd   <z1=int6464#8,>in1=int64#4
+# asm 2: movd   <z1=%xmm7,>in1=%rcx
+movd   %xmm7,%rcx
+
+# qhasm:   in2 = z2
+# asm 1: movd   <z2=int6464#11,>in2=int64#5
+# asm 2: movd   <z2=%xmm10,>in2=%r8
+movd   %xmm10,%r8
+
+# qhasm:   in3 = z3
+# asm 1: movd   <z3=int6464#5,>in3=int64#6
+# asm 2: movd   <z3=%xmm4,>in3=%r9
+movd   %xmm4,%r9
+
+# qhasm:   z0 <<<= 96
+# asm 1: pshufd $0x39,<z0=int6464#13,<z0=int6464#13
+# asm 2: pshufd $0x39,<z0=%xmm12,<z0=%xmm12
+pshufd $0x39,%xmm12,%xmm12
+
+# qhasm:   z1 <<<= 96
+# asm 1: pshufd $0x39,<z1=int6464#8,<z1=int6464#8
+# asm 2: pshufd $0x39,<z1=%xmm7,<z1=%xmm7
+pshufd $0x39,%xmm7,%xmm7
+
+# qhasm:   z2 <<<= 96
+# asm 1: pshufd $0x39,<z2=int6464#11,<z2=int6464#11
+# asm 2: pshufd $0x39,<z2=%xmm10,<z2=%xmm10
+pshufd $0x39,%xmm10,%xmm10
+
+# qhasm:   z3 <<<= 96
+# asm 1: pshufd $0x39,<z3=int6464#5,<z3=int6464#5
+# asm 2: pshufd $0x39,<z3=%xmm4,<z3=%xmm4
+pshufd $0x39,%xmm4,%xmm4
+
+# qhasm:   (uint32) in0 ^= *(uint32 *) (m + 0)
+# asm 1: xorl 0(<m=int64#2),<in0=int64#3d
+# asm 2: xorl 0(<m=%rsi),<in0=%edx
+xorl 0(%rsi),%edx
+
+# qhasm:   (uint32) in1 ^= *(uint32 *) (m + 4)
+# asm 1: xorl 4(<m=int64#2),<in1=int64#4d
+# asm 2: xorl 4(<m=%rsi),<in1=%ecx
+xorl 4(%rsi),%ecx
+
+# qhasm:   (uint32) in2 ^= *(uint32 *) (m + 8)
+# asm 1: xorl 8(<m=int64#2),<in2=int64#5d
+# asm 2: xorl 8(<m=%rsi),<in2=%r8d
+xorl 8(%rsi),%r8d
+
+# qhasm:   (uint32) in3 ^= *(uint32 *) (m + 12)
+# asm 1: xorl 12(<m=int64#2),<in3=int64#6d
+# asm 2: xorl 12(<m=%rsi),<in3=%r9d
+xorl 12(%rsi),%r9d
+
+# qhasm:   *(uint32 *) (out + 0) = in0
+# asm 1: movl   <in0=int64#3d,0(<out=int64#1)
+# asm 2: movl   <in0=%edx,0(<out=%rdi)
+movl   %edx,0(%rdi)
+
+# qhasm:   *(uint32 *) (out + 4) = in1
+# asm 1: movl   <in1=int64#4d,4(<out=int64#1)
+# asm 2: movl   <in1=%ecx,4(<out=%rdi)
+movl   %ecx,4(%rdi)
+
+# qhasm:   *(uint32 *) (out + 8) = in2
+# asm 1: movl   <in2=int64#5d,8(<out=int64#1)
+# asm 2: movl   <in2=%r8d,8(<out=%rdi)
+movl   %r8d,8(%rdi)
+
+# qhasm:   *(uint32 *) (out + 12) = in3
+# asm 1: movl   <in3=int64#6d,12(<out=int64#1)
+# asm 2: movl   <in3=%r9d,12(<out=%rdi)
+movl   %r9d,12(%rdi)
+
+# qhasm:   in0 = z0
+# asm 1: movd   <z0=int6464#13,>in0=int64#3
+# asm 2: movd   <z0=%xmm12,>in0=%rdx
+movd   %xmm12,%rdx
+
+# qhasm:   in1 = z1
+# asm 1: movd   <z1=int6464#8,>in1=int64#4
+# asm 2: movd   <z1=%xmm7,>in1=%rcx
+movd   %xmm7,%rcx
+
+# qhasm:   in2 = z2
+# asm 1: movd   <z2=int6464#11,>in2=int64#5
+# asm 2: movd   <z2=%xmm10,>in2=%r8
+movd   %xmm10,%r8
+
+# qhasm:   in3 = z3
+# asm 1: movd   <z3=int6464#5,>in3=int64#6
+# asm 2: movd   <z3=%xmm4,>in3=%r9
+movd   %xmm4,%r9
+
+# qhasm:   z0 <<<= 96
+# asm 1: pshufd $0x39,<z0=int6464#13,<z0=int6464#13
+# asm 2: pshufd $0x39,<z0=%xmm12,<z0=%xmm12
+pshufd $0x39,%xmm12,%xmm12
+
+# qhasm:   z1 <<<= 96
+# asm 1: pshufd $0x39,<z1=int6464#8,<z1=int6464#8
+# asm 2: pshufd $0x39,<z1=%xmm7,<z1=%xmm7
+pshufd $0x39,%xmm7,%xmm7
+
+# qhasm:   z2 <<<= 96
+# asm 1: pshufd $0x39,<z2=int6464#11,<z2=int6464#11
+# asm 2: pshufd $0x39,<z2=%xmm10,<z2=%xmm10
+pshufd $0x39,%xmm10,%xmm10
+
+# qhasm:   z3 <<<= 96
+# asm 1: pshufd $0x39,<z3=int6464#5,<z3=int6464#5
+# asm 2: pshufd $0x39,<z3=%xmm4,<z3=%xmm4
+pshufd $0x39,%xmm4,%xmm4
+
+# qhasm:   (uint32) in0 ^= *(uint32 *) (m + 64)
+# asm 1: xorl 64(<m=int64#2),<in0=int64#3d
+# asm 2: xorl 64(<m=%rsi),<in0=%edx
+xorl 64(%rsi),%edx
+
+# qhasm:   (uint32) in1 ^= *(uint32 *) (m + 68)
+# asm 1: xorl 68(<m=int64#2),<in1=int64#4d
+# asm 2: xorl 68(<m=%rsi),<in1=%ecx
+xorl 68(%rsi),%ecx
+
+# qhasm:   (uint32) in2 ^= *(uint32 *) (m + 72)
+# asm 1: xorl 72(<m=int64#2),<in2=int64#5d
+# asm 2: xorl 72(<m=%rsi),<in2=%r8d
+xorl 72(%rsi),%r8d
+
+# qhasm:   (uint32) in3 ^= *(uint32 *) (m + 76)
+# asm 1: xorl 76(<m=int64#2),<in3=int64#6d
+# asm 2: xorl 76(<m=%rsi),<in3=%r9d
+xorl 76(%rsi),%r9d
+
+# qhasm:   *(uint32 *) (out + 64) = in0
+# asm 1: movl   <in0=int64#3d,64(<out=int64#1)
+# asm 2: movl   <in0=%edx,64(<out=%rdi)
+movl   %edx,64(%rdi)
+
+# qhasm:   *(uint32 *) (out + 68) = in1
+# asm 1: movl   <in1=int64#4d,68(<out=int64#1)
+# asm 2: movl   <in1=%ecx,68(<out=%rdi)
+movl   %ecx,68(%rdi)
+
+# qhasm:   *(uint32 *) (out + 72) = in2
+# asm 1: movl   <in2=int64#5d,72(<out=int64#1)
+# asm 2: movl   <in2=%r8d,72(<out=%rdi)
+movl   %r8d,72(%rdi)
+
+# qhasm:   *(uint32 *) (out + 76) = in3
+# asm 1: movl   <in3=int64#6d,76(<out=int64#1)
+# asm 2: movl   <in3=%r9d,76(<out=%rdi)
+movl   %r9d,76(%rdi)
+
+# qhasm:   in0 = z0
+# asm 1: movd   <z0=int6464#13,>in0=int64#3
+# asm 2: movd   <z0=%xmm12,>in0=%rdx
+movd   %xmm12,%rdx
+
+# qhasm:   in1 = z1
+# asm 1: movd   <z1=int6464#8,>in1=int64#4
+# asm 2: movd   <z1=%xmm7,>in1=%rcx
+movd   %xmm7,%rcx
+
+# qhasm:   in2 = z2
+# asm 1: movd   <z2=int6464#11,>in2=int64#5
+# asm 2: movd   <z2=%xmm10,>in2=%r8
+movd   %xmm10,%r8
+
+# qhasm:   in3 = z3
+# asm 1: movd   <z3=int6464#5,>in3=int64#6
+# asm 2: movd   <z3=%xmm4,>in3=%r9
+movd   %xmm4,%r9
+
+# qhasm:   z0 <<<= 96
+# asm 1: pshufd $0x39,<z0=int6464#13,<z0=int6464#13
+# asm 2: pshufd $0x39,<z0=%xmm12,<z0=%xmm12
+pshufd $0x39,%xmm12,%xmm12
+
+# qhasm:   z1 <<<= 96
+# asm 1: pshufd $0x39,<z1=int6464#8,<z1=int6464#8
+# asm 2: pshufd $0x39,<z1=%xmm7,<z1=%xmm7
+pshufd $0x39,%xmm7,%xmm7
+
+# qhasm:   z2 <<<= 96
+# asm 1: pshufd $0x39,<z2=int6464#11,<z2=int6464#11
+# asm 2: pshufd $0x39,<z2=%xmm10,<z2=%xmm10
+pshufd $0x39,%xmm10,%xmm10
+
+# qhasm:   z3 <<<= 96
+# asm 1: pshufd $0x39,<z3=int6464#5,<z3=int6464#5
+# asm 2: pshufd $0x39,<z3=%xmm4,<z3=%xmm4
+pshufd $0x39,%xmm4,%xmm4
+
+# qhasm:   (uint32) in0 ^= *(uint32 *) (m + 128)
+# asm 1: xorl 128(<m=int64#2),<in0=int64#3d
+# asm 2: xorl 128(<m=%rsi),<in0=%edx
+xorl 128(%rsi),%edx
+
+# qhasm:   (uint32) in1 ^= *(uint32 *) (m + 132)
+# asm 1: xorl 132(<m=int64#2),<in1=int64#4d
+# asm 2: xorl 132(<m=%rsi),<in1=%ecx
+xorl 132(%rsi),%ecx
+
+# qhasm:   (uint32) in2 ^= *(uint32 *) (m + 136)
+# asm 1: xorl 136(<m=int64#2),<in2=int64#5d
+# asm 2: xorl 136(<m=%rsi),<in2=%r8d
+xorl 136(%rsi),%r8d
+
+# qhasm:   (uint32) in3 ^= *(uint32 *) (m + 140)
+# asm 1: xorl 140(<m=int64#2),<in3=int64#6d
+# asm 2: xorl 140(<m=%rsi),<in3=%r9d
+xorl 140(%rsi),%r9d
+
+# qhasm:   *(uint32 *) (out + 128) = in0
+# asm 1: movl   <in0=int64#3d,128(<out=int64#1)
+# asm 2: movl   <in0=%edx,128(<out=%rdi)
+movl   %edx,128(%rdi)
+
+# qhasm:   *(uint32 *) (out + 132) = in1
+# asm 1: movl   <in1=int64#4d,132(<out=int64#1)
+# asm 2: movl   <in1=%ecx,132(<out=%rdi)
+movl   %ecx,132(%rdi)
+
+# qhasm:   *(uint32 *) (out + 136) = in2
+# asm 1: movl   <in2=int64#5d,136(<out=int64#1)
+# asm 2: movl   <in2=%r8d,136(<out=%rdi)
+movl   %r8d,136(%rdi)
+
+# qhasm:   *(uint32 *) (out + 140) = in3
+# asm 1: movl   <in3=int64#6d,140(<out=int64#1)
+# asm 2: movl   <in3=%r9d,140(<out=%rdi)
+movl   %r9d,140(%rdi)
+
+# qhasm:   in0 = z0
+# asm 1: movd   <z0=int6464#13,>in0=int64#3
+# asm 2: movd   <z0=%xmm12,>in0=%rdx
+movd   %xmm12,%rdx
+
+# qhasm:   in1 = z1
+# asm 1: movd   <z1=int6464#8,>in1=int64#4
+# asm 2: movd   <z1=%xmm7,>in1=%rcx
+movd   %xmm7,%rcx
+
+# qhasm:   in2 = z2
+# asm 1: movd   <z2=int6464#11,>in2=int64#5
+# asm 2: movd   <z2=%xmm10,>in2=%r8
+movd   %xmm10,%r8
+
+# qhasm:   in3 = z3
+# asm 1: movd   <z3=int6464#5,>in3=int64#6
+# asm 2: movd   <z3=%xmm4,>in3=%r9
+movd   %xmm4,%r9
+
+# qhasm:   (uint32) in0 ^= *(uint32 *) (m + 192)
+# asm 1: xorl 192(<m=int64#2),<in0=int64#3d
+# asm 2: xorl 192(<m=%rsi),<in0=%edx
+xorl 192(%rsi),%edx
+
+# qhasm:   (uint32) in1 ^= *(uint32 *) (m + 196)
+# asm 1: xorl 196(<m=int64#2),<in1=int64#4d
+# asm 2: xorl 196(<m=%rsi),<in1=%ecx
+xorl 196(%rsi),%ecx
+
+# qhasm:   (uint32) in2 ^= *(uint32 *) (m + 200)
+# asm 1: xorl 200(<m=int64#2),<in2=int64#5d
+# asm 2: xorl 200(<m=%rsi),<in2=%r8d
+xorl 200(%rsi),%r8d
+
+# qhasm:   (uint32) in3 ^= *(uint32 *) (m + 204)
+# asm 1: xorl 204(<m=int64#2),<in3=int64#6d
+# asm 2: xorl 204(<m=%rsi),<in3=%r9d
+xorl 204(%rsi),%r9d
+
+# qhasm:   *(uint32 *) (out + 192) = in0
+# asm 1: movl   <in0=int64#3d,192(<out=int64#1)
+# asm 2: movl   <in0=%edx,192(<out=%rdi)
+movl   %edx,192(%rdi)
+
+# qhasm:   *(uint32 *) (out + 196) = in1
+# asm 1: movl   <in1=int64#4d,196(<out=int64#1)
+# asm 2: movl   <in1=%ecx,196(<out=%rdi)
+movl   %ecx,196(%rdi)
+
+# qhasm:   *(uint32 *) (out + 200) = in2
+# asm 1: movl   <in2=int64#5d,200(<out=int64#1)
+# asm 2: movl   <in2=%r8d,200(<out=%rdi)
+movl   %r8d,200(%rdi)
+
+# qhasm:   *(uint32 *) (out + 204) = in3
+# asm 1: movl   <in3=int64#6d,204(<out=int64#1)
+# asm 2: movl   <in3=%r9d,204(<out=%rdi)
+movl   %r9d,204(%rdi)
+
+# qhasm:   uint32323232 z4 += orig4
+# asm 1: paddd <orig4=stack128#16,<z4=int6464#15
+# asm 2: paddd <orig4=240(%rsp),<z4=%xmm14
+paddd 240(%rsp),%xmm14
+
+# qhasm:   uint32323232 z5 += orig5
+# asm 1: paddd <orig5=stack128#5,<z5=int6464#1
+# asm 2: paddd <orig5=64(%rsp),<z5=%xmm0
+paddd 64(%rsp),%xmm0
+
+# qhasm:   uint32323232 z6 += orig6
+# asm 1: paddd <orig6=stack128#9,<z6=int6464#6
+# asm 2: paddd <orig6=128(%rsp),<z6=%xmm5
+paddd 128(%rsp),%xmm5
+
+# qhasm:   uint32323232 z7 += orig7
+# asm 1: paddd <orig7=stack128#13,<z7=int6464#9
+# asm 2: paddd <orig7=192(%rsp),<z7=%xmm8
+paddd 192(%rsp),%xmm8
+
+# qhasm:   in4 = z4
+# asm 1: movd   <z4=int6464#15,>in4=int64#3
+# asm 2: movd   <z4=%xmm14,>in4=%rdx
+movd   %xmm14,%rdx
+
+# qhasm:   in5 = z5
+# asm 1: movd   <z5=int6464#1,>in5=int64#4
+# asm 2: movd   <z5=%xmm0,>in5=%rcx
+movd   %xmm0,%rcx
+
+# qhasm:   in6 = z6
+# asm 1: movd   <z6=int6464#6,>in6=int64#5
+# asm 2: movd   <z6=%xmm5,>in6=%r8
+movd   %xmm5,%r8
+
+# qhasm:   in7 = z7
+# asm 1: movd   <z7=int6464#9,>in7=int64#6
+# asm 2: movd   <z7=%xmm8,>in7=%r9
+movd   %xmm8,%r9
+
+# qhasm:   z4 <<<= 96
+# asm 1: pshufd $0x39,<z4=int6464#15,<z4=int6464#15
+# asm 2: pshufd $0x39,<z4=%xmm14,<z4=%xmm14
+pshufd $0x39,%xmm14,%xmm14
+
+# qhasm:   z5 <<<= 96
+# asm 1: pshufd $0x39,<z5=int6464#1,<z5=int6464#1
+# asm 2: pshufd $0x39,<z5=%xmm0,<z5=%xmm0
+pshufd $0x39,%xmm0,%xmm0
+
+# qhasm:   z6 <<<= 96
+# asm 1: pshufd $0x39,<z6=int6464#6,<z6=int6464#6
+# asm 2: pshufd $0x39,<z6=%xmm5,<z6=%xmm5
+pshufd $0x39,%xmm5,%xmm5
+
+# qhasm:   z7 <<<= 96
+# asm 1: pshufd $0x39,<z7=int6464#9,<z7=int6464#9
+# asm 2: pshufd $0x39,<z7=%xmm8,<z7=%xmm8
+pshufd $0x39,%xmm8,%xmm8
+
+# qhasm:   (uint32) in4 ^= *(uint32 *) (m + 16)
+# asm 1: xorl 16(<m=int64#2),<in4=int64#3d
+# asm 2: xorl 16(<m=%rsi),<in4=%edx
+xorl 16(%rsi),%edx
+
+# qhasm:   (uint32) in5 ^= *(uint32 *) (m + 20)
+# asm 1: xorl 20(<m=int64#2),<in5=int64#4d
+# asm 2: xorl 20(<m=%rsi),<in5=%ecx
+xorl 20(%rsi),%ecx
+
+# qhasm:   (uint32) in6 ^= *(uint32 *) (m + 24)
+# asm 1: xorl 24(<m=int64#2),<in6=int64#5d
+# asm 2: xorl 24(<m=%rsi),<in6=%r8d
+xorl 24(%rsi),%r8d
+
+# qhasm:   (uint32) in7 ^= *(uint32 *) (m + 28)
+# asm 1: xorl 28(<m=int64#2),<in7=int64#6d
+# asm 2: xorl 28(<m=%rsi),<in7=%r9d
+xorl 28(%rsi),%r9d
+
+# qhasm:   *(uint32 *) (out + 16) = in4
+# asm 1: movl   <in4=int64#3d,16(<out=int64#1)
+# asm 2: movl   <in4=%edx,16(<out=%rdi)
+movl   %edx,16(%rdi)
+
+# qhasm:   *(uint32 *) (out + 20) = in5
+# asm 1: movl   <in5=int64#4d,20(<out=int64#1)
+# asm 2: movl   <in5=%ecx,20(<out=%rdi)
+movl   %ecx,20(%rdi)
+
+# qhasm:   *(uint32 *) (out + 24) = in6
+# asm 1: movl   <in6=int64#5d,24(<out=int64#1)
+# asm 2: movl   <in6=%r8d,24(<out=%rdi)
+movl   %r8d,24(%rdi)
+
+# qhasm:   *(uint32 *) (out + 28) = in7
+# asm 1: movl   <in7=int64#6d,28(<out=int64#1)
+# asm 2: movl   <in7=%r9d,28(<out=%rdi)
+movl   %r9d,28(%rdi)
+
+# qhasm:   in4 = z4
+# asm 1: movd   <z4=int6464#15,>in4=int64#3
+# asm 2: movd   <z4=%xmm14,>in4=%rdx
+movd   %xmm14,%rdx
+
+# qhasm:   in5 = z5
+# asm 1: movd   <z5=int6464#1,>in5=int64#4
+# asm 2: movd   <z5=%xmm0,>in5=%rcx
+movd   %xmm0,%rcx
+
+# qhasm:   in6 = z6
+# asm 1: movd   <z6=int6464#6,>in6=int64#5
+# asm 2: movd   <z6=%xmm5,>in6=%r8
+movd   %xmm5,%r8
+
+# qhasm:   in7 = z7
+# asm 1: movd   <z7=int6464#9,>in7=int64#6
+# asm 2: movd   <z7=%xmm8,>in7=%r9
+movd   %xmm8,%r9
+
+# qhasm:   z4 <<<= 96
+# asm 1: pshufd $0x39,<z4=int6464#15,<z4=int6464#15
+# asm 2: pshufd $0x39,<z4=%xmm14,<z4=%xmm14
+pshufd $0x39,%xmm14,%xmm14
+
+# qhasm:   z5 <<<= 96
+# asm 1: pshufd $0x39,<z5=int6464#1,<z5=int6464#1
+# asm 2: pshufd $0x39,<z5=%xmm0,<z5=%xmm0
+pshufd $0x39,%xmm0,%xmm0
+
+# qhasm:   z6 <<<= 96
+# asm 1: pshufd $0x39,<z6=int6464#6,<z6=int6464#6
+# asm 2: pshufd $0x39,<z6=%xmm5,<z6=%xmm5
+pshufd $0x39,%xmm5,%xmm5
+
+# qhasm:   z7 <<<= 96
+# asm 1: pshufd $0x39,<z7=int6464#9,<z7=int6464#9
+# asm 2: pshufd $0x39,<z7=%xmm8,<z7=%xmm8
+pshufd $0x39,%xmm8,%xmm8
+
+# qhasm:   (uint32) in4 ^= *(uint32 *) (m + 80)
+# asm 1: xorl 80(<m=int64#2),<in4=int64#3d
+# asm 2: xorl 80(<m=%rsi),<in4=%edx
+xorl 80(%rsi),%edx
+
+# qhasm:   (uint32) in5 ^= *(uint32 *) (m + 84)
+# asm 1: xorl 84(<m=int64#2),<in5=int64#4d
+# asm 2: xorl 84(<m=%rsi),<in5=%ecx
+xorl 84(%rsi),%ecx
+
+# qhasm:   (uint32) in6 ^= *(uint32 *) (m + 88)
+# asm 1: xorl 88(<m=int64#2),<in6=int64#5d
+# asm 2: xorl 88(<m=%rsi),<in6=%r8d
+xorl 88(%rsi),%r8d
+
+# qhasm:   (uint32) in7 ^= *(uint32 *) (m + 92)
+# asm 1: xorl 92(<m=int64#2),<in7=int64#6d
+# asm 2: xorl 92(<m=%rsi),<in7=%r9d
+xorl 92(%rsi),%r9d
+
+# qhasm:   *(uint32 *) (out + 80) = in4
+# asm 1: movl   <in4=int64#3d,80(<out=int64#1)
+# asm 2: movl   <in4=%edx,80(<out=%rdi)
+movl   %edx,80(%rdi)
+
+# qhasm:   *(uint32 *) (out + 84) = in5
+# asm 1: movl   <in5=int64#4d,84(<out=int64#1)
+# asm 2: movl   <in5=%ecx,84(<out=%rdi)
+movl   %ecx,84(%rdi)
+
+# qhasm:   *(uint32 *) (out + 88) = in6
+# asm 1: movl   <in6=int64#5d,88(<out=int64#1)
+# asm 2: movl   <in6=%r8d,88(<out=%rdi)
+movl   %r8d,88(%rdi)
+
+# qhasm:   *(uint32 *) (out + 92) = in7
+# asm 1: movl   <in7=int64#6d,92(<out=int64#1)
+# asm 2: movl   <in7=%r9d,92(<out=%rdi)
+movl   %r9d,92(%rdi)
+
+# qhasm:   in4 = z4
+# asm 1: movd   <z4=int6464#15,>in4=int64#3
+# asm 2: movd   <z4=%xmm14,>in4=%rdx
+movd   %xmm14,%rdx
+
+# qhasm:   in5 = z5
+# asm 1: movd   <z5=int6464#1,>in5=int64#4
+# asm 2: movd   <z5=%xmm0,>in5=%rcx
+movd   %xmm0,%rcx
+
+# qhasm:   in6 = z6
+# asm 1: movd   <z6=int6464#6,>in6=int64#5
+# asm 2: movd   <z6=%xmm5,>in6=%r8
+movd   %xmm5,%r8
+
+# qhasm:   in7 = z7
+# asm 1: movd   <z7=int6464#9,>in7=int64#6
+# asm 2: movd   <z7=%xmm8,>in7=%r9
+movd   %xmm8,%r9
+
+# qhasm:   z4 <<<= 96
+# asm 1: pshufd $0x39,<z4=int6464#15,<z4=int6464#15
+# asm 2: pshufd $0x39,<z4=%xmm14,<z4=%xmm14
+pshufd $0x39,%xmm14,%xmm14
+
+# qhasm:   z5 <<<= 96
+# asm 1: pshufd $0x39,<z5=int6464#1,<z5=int6464#1
+# asm 2: pshufd $0x39,<z5=%xmm0,<z5=%xmm0
+pshufd $0x39,%xmm0,%xmm0
+
+# qhasm:   z6 <<<= 96
+# asm 1: pshufd $0x39,<z6=int6464#6,<z6=int6464#6
+# asm 2: pshufd $0x39,<z6=%xmm5,<z6=%xmm5
+pshufd $0x39,%xmm5,%xmm5
+
+# qhasm:   z7 <<<= 96
+# asm 1: pshufd $0x39,<z7=int6464#9,<z7=int6464#9
+# asm 2: pshufd $0x39,<z7=%xmm8,<z7=%xmm8
+pshufd $0x39,%xmm8,%xmm8
+
+# qhasm:   (uint32) in4 ^= *(uint32 *) (m + 144)
+# asm 1: xorl 144(<m=int64#2),<in4=int64#3d
+# asm 2: xorl 144(<m=%rsi),<in4=%edx
+xorl 144(%rsi),%edx
+
+# qhasm:   (uint32) in5 ^= *(uint32 *) (m + 148)
+# asm 1: xorl 148(<m=int64#2),<in5=int64#4d
+# asm 2: xorl 148(<m=%rsi),<in5=%ecx
+xorl 148(%rsi),%ecx
+
+# qhasm:   (uint32) in6 ^= *(uint32 *) (m + 152)
+# asm 1: xorl 152(<m=int64#2),<in6=int64#5d
+# asm 2: xorl 152(<m=%rsi),<in6=%r8d
+xorl 152(%rsi),%r8d
+
+# qhasm:   (uint32) in7 ^= *(uint32 *) (m + 156)
+# asm 1: xorl 156(<m=int64#2),<in7=int64#6d
+# asm 2: xorl 156(<m=%rsi),<in7=%r9d
+xorl 156(%rsi),%r9d
+
+# qhasm:   *(uint32 *) (out + 144) = in4
+# asm 1: movl   <in4=int64#3d,144(<out=int64#1)
+# asm 2: movl   <in4=%edx,144(<out=%rdi)
+movl   %edx,144(%rdi)
+
+# qhasm:   *(uint32 *) (out + 148) = in5
+# asm 1: movl   <in5=int64#4d,148(<out=int64#1)
+# asm 2: movl   <in5=%ecx,148(<out=%rdi)
+movl   %ecx,148(%rdi)
+
+# qhasm:   *(uint32 *) (out + 152) = in6
+# asm 1: movl   <in6=int64#5d,152(<out=int64#1)
+# asm 2: movl   <in6=%r8d,152(<out=%rdi)
+movl   %r8d,152(%rdi)
+
+# qhasm:   *(uint32 *) (out + 156) = in7
+# asm 1: movl   <in7=int64#6d,156(<out=int64#1)
+# asm 2: movl   <in7=%r9d,156(<out=%rdi)
+movl   %r9d,156(%rdi)
+
+# qhasm:   in4 = z4
+# asm 1: movd   <z4=int6464#15,>in4=int64#3
+# asm 2: movd   <z4=%xmm14,>in4=%rdx
+movd   %xmm14,%rdx
+
+# qhasm:   in5 = z5
+# asm 1: movd   <z5=int6464#1,>in5=int64#4
+# asm 2: movd   <z5=%xmm0,>in5=%rcx
+movd   %xmm0,%rcx
+
+# qhasm:   in6 = z6
+# asm 1: movd   <z6=int6464#6,>in6=int64#5
+# asm 2: movd   <z6=%xmm5,>in6=%r8
+movd   %xmm5,%r8
+
+# qhasm:   in7 = z7
+# asm 1: movd   <z7=int6464#9,>in7=int64#6
+# asm 2: movd   <z7=%xmm8,>in7=%r9
+movd   %xmm8,%r9
+
+# qhasm:   (uint32) in4 ^= *(uint32 *) (m + 208)
+# asm 1: xorl 208(<m=int64#2),<in4=int64#3d
+# asm 2: xorl 208(<m=%rsi),<in4=%edx
+xorl 208(%rsi),%edx
+
+# qhasm:   (uint32) in5 ^= *(uint32 *) (m + 212)
+# asm 1: xorl 212(<m=int64#2),<in5=int64#4d
+# asm 2: xorl 212(<m=%rsi),<in5=%ecx
+xorl 212(%rsi),%ecx
+
+# qhasm:   (uint32) in6 ^= *(uint32 *) (m + 216)
+# asm 1: xorl 216(<m=int64#2),<in6=int64#5d
+# asm 2: xorl 216(<m=%rsi),<in6=%r8d
+xorl 216(%rsi),%r8d
+
+# qhasm:   (uint32) in7 ^= *(uint32 *) (m + 220)
+# asm 1: xorl 220(<m=int64#2),<in7=int64#6d
+# asm 2: xorl 220(<m=%rsi),<in7=%r9d
+xorl 220(%rsi),%r9d
+
+# qhasm:   *(uint32 *) (out + 208) = in4
+# asm 1: movl   <in4=int64#3d,208(<out=int64#1)
+# asm 2: movl   <in4=%edx,208(<out=%rdi)
+movl   %edx,208(%rdi)
+
+# qhasm:   *(uint32 *) (out + 212) = in5
+# asm 1: movl   <in5=int64#4d,212(<out=int64#1)
+# asm 2: movl   <in5=%ecx,212(<out=%rdi)
+movl   %ecx,212(%rdi)
+
+# qhasm:   *(uint32 *) (out + 216) = in6
+# asm 1: movl   <in6=int64#5d,216(<out=int64#1)
+# asm 2: movl   <in6=%r8d,216(<out=%rdi)
+movl   %r8d,216(%rdi)
+
+# qhasm:   *(uint32 *) (out + 220) = in7
+# asm 1: movl   <in7=int64#6d,220(<out=int64#1)
+# asm 2: movl   <in7=%r9d,220(<out=%rdi)
+movl   %r9d,220(%rdi)
+
+# qhasm:   uint32323232 z8 += orig8
+# asm 1: paddd <orig8=stack128#19,<z8=int6464#16
+# asm 2: paddd <orig8=288(%rsp),<z8=%xmm15
+paddd 288(%rsp),%xmm15
+
+# qhasm:   uint32323232 z9 += orig9
+# asm 1: paddd <orig9=stack128#20,<z9=int6464#12
+# asm 2: paddd <orig9=304(%rsp),<z9=%xmm11
+paddd 304(%rsp),%xmm11
+
+# qhasm:   uint32323232 z10 += orig10
+# asm 1: paddd <orig10=stack128#6,<z10=int6464#2
+# asm 2: paddd <orig10=80(%rsp),<z10=%xmm1
+paddd 80(%rsp),%xmm1
+
+# qhasm:   uint32323232 z11 += orig11
+# asm 1: paddd <orig11=stack128#10,<z11=int6464#7
+# asm 2: paddd <orig11=144(%rsp),<z11=%xmm6
+paddd 144(%rsp),%xmm6
+
+# qhasm:   in8 = z8
+# asm 1: movd   <z8=int6464#16,>in8=int64#3
+# asm 2: movd   <z8=%xmm15,>in8=%rdx
+movd   %xmm15,%rdx
+
+# qhasm:   in9 = z9
+# asm 1: movd   <z9=int6464#12,>in9=int64#4
+# asm 2: movd   <z9=%xmm11,>in9=%rcx
+movd   %xmm11,%rcx
+
+# qhasm:   in10 = z10
+# asm 1: movd   <z10=int6464#2,>in10=int64#5
+# asm 2: movd   <z10=%xmm1,>in10=%r8
+movd   %xmm1,%r8
+
+# qhasm:   in11 = z11
+# asm 1: movd   <z11=int6464#7,>in11=int64#6
+# asm 2: movd   <z11=%xmm6,>in11=%r9
+movd   %xmm6,%r9
+
+# qhasm:   z8 <<<= 96
+# asm 1: pshufd $0x39,<z8=int6464#16,<z8=int6464#16
+# asm 2: pshufd $0x39,<z8=%xmm15,<z8=%xmm15
+pshufd $0x39,%xmm15,%xmm15
+
+# qhasm:   z9 <<<= 96
+# asm 1: pshufd $0x39,<z9=int6464#12,<z9=int6464#12
+# asm 2: pshufd $0x39,<z9=%xmm11,<z9=%xmm11
+pshufd $0x39,%xmm11,%xmm11
+
+# qhasm:   z10 <<<= 96
+# asm 1: pshufd $0x39,<z10=int6464#2,<z10=int6464#2
+# asm 2: pshufd $0x39,<z10=%xmm1,<z10=%xmm1
+pshufd $0x39,%xmm1,%xmm1
+
+# qhasm:   z11 <<<= 96
+# asm 1: pshufd $0x39,<z11=int6464#7,<z11=int6464#7
+# asm 2: pshufd $0x39,<z11=%xmm6,<z11=%xmm6
+pshufd $0x39,%xmm6,%xmm6
+
+# qhasm:   (uint32) in8 ^= *(uint32 *) (m + 32)
+# asm 1: xorl 32(<m=int64#2),<in8=int64#3d
+# asm 2: xorl 32(<m=%rsi),<in8=%edx
+xorl 32(%rsi),%edx
+
+# qhasm:   (uint32) in9 ^= *(uint32 *) (m + 36)
+# asm 1: xorl 36(<m=int64#2),<in9=int64#4d
+# asm 2: xorl 36(<m=%rsi),<in9=%ecx
+xorl 36(%rsi),%ecx
+
+# qhasm:   (uint32) in10 ^= *(uint32 *) (m + 40)
+# asm 1: xorl 40(<m=int64#2),<in10=int64#5d
+# asm 2: xorl 40(<m=%rsi),<in10=%r8d
+xorl 40(%rsi),%r8d
+
+# qhasm:   (uint32) in11 ^= *(uint32 *) (m + 44)
+# asm 1: xorl 44(<m=int64#2),<in11=int64#6d
+# asm 2: xorl 44(<m=%rsi),<in11=%r9d
+xorl 44(%rsi),%r9d
+
+# qhasm:   *(uint32 *) (out + 32) = in8
+# asm 1: movl   <in8=int64#3d,32(<out=int64#1)
+# asm 2: movl   <in8=%edx,32(<out=%rdi)
+movl   %edx,32(%rdi)
+
+# qhasm:   *(uint32 *) (out + 36) = in9
+# asm 1: movl   <in9=int64#4d,36(<out=int64#1)
+# asm 2: movl   <in9=%ecx,36(<out=%rdi)
+movl   %ecx,36(%rdi)
+
+# qhasm:   *(uint32 *) (out + 40) = in10
+# asm 1: movl   <in10=int64#5d,40(<out=int64#1)
+# asm 2: movl   <in10=%r8d,40(<out=%rdi)
+movl   %r8d,40(%rdi)
+
+# qhasm:   *(uint32 *) (out + 44) = in11
+# asm 1: movl   <in11=int64#6d,44(<out=int64#1)
+# asm 2: movl   <in11=%r9d,44(<out=%rdi)
+movl   %r9d,44(%rdi)
+
+# qhasm:   in8 = z8
+# asm 1: movd   <z8=int6464#16,>in8=int64#3
+# asm 2: movd   <z8=%xmm15,>in8=%rdx
+movd   %xmm15,%rdx
+
+# qhasm:   in9 = z9
+# asm 1: movd   <z9=int6464#12,>in9=int64#4
+# asm 2: movd   <z9=%xmm11,>in9=%rcx
+movd   %xmm11,%rcx
+
+# qhasm:   in10 = z10
+# asm 1: movd   <z10=int6464#2,>in10=int64#5
+# asm 2: movd   <z10=%xmm1,>in10=%r8
+movd   %xmm1,%r8
+
+# qhasm:   in11 = z11
+# asm 1: movd   <z11=int6464#7,>in11=int64#6
+# asm 2: movd   <z11=%xmm6,>in11=%r9
+movd   %xmm6,%r9
+
+# qhasm:   z8 <<<= 96
+# asm 1: pshufd $0x39,<z8=int6464#16,<z8=int6464#16
+# asm 2: pshufd $0x39,<z8=%xmm15,<z8=%xmm15
+pshufd $0x39,%xmm15,%xmm15
+
+# qhasm:   z9 <<<= 96
+# asm 1: pshufd $0x39,<z9=int6464#12,<z9=int6464#12
+# asm 2: pshufd $0x39,<z9=%xmm11,<z9=%xmm11
+pshufd $0x39,%xmm11,%xmm11
+
+# qhasm:   z10 <<<= 96
+# asm 1: pshufd $0x39,<z10=int6464#2,<z10=int6464#2
+# asm 2: pshufd $0x39,<z10=%xmm1,<z10=%xmm1
+pshufd $0x39,%xmm1,%xmm1
+
+# qhasm:   z11 <<<= 96
+# asm 1: pshufd $0x39,<z11=int6464#7,<z11=int6464#7
+# asm 2: pshufd $0x39,<z11=%xmm6,<z11=%xmm6
+pshufd $0x39,%xmm6,%xmm6
+
+# qhasm:   (uint32) in8 ^= *(uint32 *) (m + 96)
+# asm 1: xorl 96(<m=int64#2),<in8=int64#3d
+# asm 2: xorl 96(<m=%rsi),<in8=%edx
+xorl 96(%rsi),%edx
+
+# qhasm:   (uint32) in9 ^= *(uint32 *) (m + 100)
+# asm 1: xorl 100(<m=int64#2),<in9=int64#4d
+# asm 2: xorl 100(<m=%rsi),<in9=%ecx
+xorl 100(%rsi),%ecx
+
+# qhasm:   (uint32) in10 ^= *(uint32 *) (m + 104)
+# asm 1: xorl 104(<m=int64#2),<in10=int64#5d
+# asm 2: xorl 104(<m=%rsi),<in10=%r8d
+xorl 104(%rsi),%r8d
+
+# qhasm:   (uint32) in11 ^= *(uint32 *) (m + 108)
+# asm 1: xorl 108(<m=int64#2),<in11=int64#6d
+# asm 2: xorl 108(<m=%rsi),<in11=%r9d
+xorl 108(%rsi),%r9d
+
+# qhasm:   *(uint32 *) (out + 96) = in8
+# asm 1: movl   <in8=int64#3d,96(<out=int64#1)
+# asm 2: movl   <in8=%edx,96(<out=%rdi)
+movl   %edx,96(%rdi)
+
+# qhasm:   *(uint32 *) (out + 100) = in9
+# asm 1: movl   <in9=int64#4d,100(<out=int64#1)
+# asm 2: movl   <in9=%ecx,100(<out=%rdi)
+movl   %ecx,100(%rdi)
+
+# qhasm:   *(uint32 *) (out + 104) = in10
+# asm 1: movl   <in10=int64#5d,104(<out=int64#1)
+# asm 2: movl   <in10=%r8d,104(<out=%rdi)
+movl   %r8d,104(%rdi)
+
+# qhasm:   *(uint32 *) (out + 108) = in11
+# asm 1: movl   <in11=int64#6d,108(<out=int64#1)
+# asm 2: movl   <in11=%r9d,108(<out=%rdi)
+movl   %r9d,108(%rdi)
+
+# qhasm:   in8 = z8
+# asm 1: movd   <z8=int6464#16,>in8=int64#3
+# asm 2: movd   <z8=%xmm15,>in8=%rdx
+movd   %xmm15,%rdx
+
+# qhasm:   in9 = z9
+# asm 1: movd   <z9=int6464#12,>in9=int64#4
+# asm 2: movd   <z9=%xmm11,>in9=%rcx
+movd   %xmm11,%rcx
+
+# qhasm:   in10 = z10
+# asm 1: movd   <z10=int6464#2,>in10=int64#5
+# asm 2: movd   <z10=%xmm1,>in10=%r8
+movd   %xmm1,%r8
+
+# qhasm:   in11 = z11
+# asm 1: movd   <z11=int6464#7,>in11=int64#6
+# asm 2: movd   <z11=%xmm6,>in11=%r9
+movd   %xmm6,%r9
+
+# qhasm:   z8 <<<= 96
+# asm 1: pshufd $0x39,<z8=int6464#16,<z8=int6464#16
+# asm 2: pshufd $0x39,<z8=%xmm15,<z8=%xmm15
+pshufd $0x39,%xmm15,%xmm15
+
+# qhasm:   z9 <<<= 96
+# asm 1: pshufd $0x39,<z9=int6464#12,<z9=int6464#12
+# asm 2: pshufd $0x39,<z9=%xmm11,<z9=%xmm11
+pshufd $0x39,%xmm11,%xmm11
+
+# qhasm:   z10 <<<= 96
+# asm 1: pshufd $0x39,<z10=int6464#2,<z10=int6464#2
+# asm 2: pshufd $0x39,<z10=%xmm1,<z10=%xmm1
+pshufd $0x39,%xmm1,%xmm1
+
+# qhasm:   z11 <<<= 96
+# asm 1: pshufd $0x39,<z11=int6464#7,<z11=int6464#7
+# asm 2: pshufd $0x39,<z11=%xmm6,<z11=%xmm6
+pshufd $0x39,%xmm6,%xmm6
+
+# qhasm:   (uint32) in8 ^= *(uint32 *) (m + 160)
+# asm 1: xorl 160(<m=int64#2),<in8=int64#3d
+# asm 2: xorl 160(<m=%rsi),<in8=%edx
+xorl 160(%rsi),%edx
+
+# qhasm:   (uint32) in9 ^= *(uint32 *) (m + 164)
+# asm 1: xorl 164(<m=int64#2),<in9=int64#4d
+# asm 2: xorl 164(<m=%rsi),<in9=%ecx
+xorl 164(%rsi),%ecx
+
+# qhasm:   (uint32) in10 ^= *(uint32 *) (m + 168)
+# asm 1: xorl 168(<m=int64#2),<in10=int64#5d
+# asm 2: xorl 168(<m=%rsi),<in10=%r8d
+xorl 168(%rsi),%r8d
+
+# qhasm:   (uint32) in11 ^= *(uint32 *) (m + 172)
+# asm 1: xorl 172(<m=int64#2),<in11=int64#6d
+# asm 2: xorl 172(<m=%rsi),<in11=%r9d
+xorl 172(%rsi),%r9d
+
+# qhasm:   *(uint32 *) (out + 160) = in8
+# asm 1: movl   <in8=int64#3d,160(<out=int64#1)
+# asm 2: movl   <in8=%edx,160(<out=%rdi)
+movl   %edx,160(%rdi)
+
+# qhasm:   *(uint32 *) (out + 164) = in9
+# asm 1: movl   <in9=int64#4d,164(<out=int64#1)
+# asm 2: movl   <in9=%ecx,164(<out=%rdi)
+movl   %ecx,164(%rdi)
+
+# qhasm:   *(uint32 *) (out + 168) = in10
+# asm 1: movl   <in10=int64#5d,168(<out=int64#1)
+# asm 2: movl   <in10=%r8d,168(<out=%rdi)
+movl   %r8d,168(%rdi)
+
+# qhasm:   *(uint32 *) (out + 172) = in11
+# asm 1: movl   <in11=int64#6d,172(<out=int64#1)
+# asm 2: movl   <in11=%r9d,172(<out=%rdi)
+movl   %r9d,172(%rdi)
+
+# qhasm:   in8 = z8
+# asm 1: movd   <z8=int6464#16,>in8=int64#3
+# asm 2: movd   <z8=%xmm15,>in8=%rdx
+movd   %xmm15,%rdx
+
+# qhasm:   in9 = z9
+# asm 1: movd   <z9=int6464#12,>in9=int64#4
+# asm 2: movd   <z9=%xmm11,>in9=%rcx
+movd   %xmm11,%rcx
+
+# qhasm:   in10 = z10
+# asm 1: movd   <z10=int6464#2,>in10=int64#5
+# asm 2: movd   <z10=%xmm1,>in10=%r8
+movd   %xmm1,%r8
+
+# qhasm:   in11 = z11
+# asm 1: movd   <z11=int6464#7,>in11=int64#6
+# asm 2: movd   <z11=%xmm6,>in11=%r9
+movd   %xmm6,%r9
+
+# qhasm:   (uint32) in8 ^= *(uint32 *) (m + 224)
+# asm 1: xorl 224(<m=int64#2),<in8=int64#3d
+# asm 2: xorl 224(<m=%rsi),<in8=%edx
+xorl 224(%rsi),%edx
+
+# qhasm:   (uint32) in9 ^= *(uint32 *) (m + 228)
+# asm 1: xorl 228(<m=int64#2),<in9=int64#4d
+# asm 2: xorl 228(<m=%rsi),<in9=%ecx
+xorl 228(%rsi),%ecx
+
+# qhasm:   (uint32) in10 ^= *(uint32 *) (m + 232)
+# asm 1: xorl 232(<m=int64#2),<in10=int64#5d
+# asm 2: xorl 232(<m=%rsi),<in10=%r8d
+xorl 232(%rsi),%r8d
+
+# qhasm:   (uint32) in11 ^= *(uint32 *) (m + 236)
+# asm 1: xorl 236(<m=int64#2),<in11=int64#6d
+# asm 2: xorl 236(<m=%rsi),<in11=%r9d
+xorl 236(%rsi),%r9d
+
+# qhasm:   *(uint32 *) (out + 224) = in8
+# asm 1: movl   <in8=int64#3d,224(<out=int64#1)
+# asm 2: movl   <in8=%edx,224(<out=%rdi)
+movl   %edx,224(%rdi)
+
+# qhasm:   *(uint32 *) (out + 228) = in9
+# asm 1: movl   <in9=int64#4d,228(<out=int64#1)
+# asm 2: movl   <in9=%ecx,228(<out=%rdi)
+movl   %ecx,228(%rdi)
+
+# qhasm:   *(uint32 *) (out + 232) = in10
+# asm 1: movl   <in10=int64#5d,232(<out=int64#1)
+# asm 2: movl   <in10=%r8d,232(<out=%rdi)
+movl   %r8d,232(%rdi)
+
+# qhasm:   *(uint32 *) (out + 236) = in11
+# asm 1: movl   <in11=int64#6d,236(<out=int64#1)
+# asm 2: movl   <in11=%r9d,236(<out=%rdi)
+movl   %r9d,236(%rdi)
+
+# qhasm:   uint32323232 z12 += orig12
+# asm 1: paddd <orig12=stack128#11,<z12=int6464#14
+# asm 2: paddd <orig12=160(%rsp),<z12=%xmm13
+paddd 160(%rsp),%xmm13
+
+# qhasm:   uint32323232 z13 += orig13
+# asm 1: paddd <orig13=stack128#14,<z13=int6464#10
+# asm 2: paddd <orig13=208(%rsp),<z13=%xmm9
+paddd 208(%rsp),%xmm9
+
+# qhasm:   uint32323232 z14 += orig14
+# asm 1: paddd <orig14=stack128#17,<z14=int6464#4
+# asm 2: paddd <orig14=256(%rsp),<z14=%xmm3
+paddd 256(%rsp),%xmm3
+
+# qhasm:   uint32323232 z15 += orig15
+# asm 1: paddd <orig15=stack128#7,<z15=int6464#3
+# asm 2: paddd <orig15=96(%rsp),<z15=%xmm2
+paddd 96(%rsp),%xmm2
+
+# qhasm:   in12 = z12
+# asm 1: movd   <z12=int6464#14,>in12=int64#3
+# asm 2: movd   <z12=%xmm13,>in12=%rdx
+movd   %xmm13,%rdx
+
+# qhasm:   in13 = z13
+# asm 1: movd   <z13=int6464#10,>in13=int64#4
+# asm 2: movd   <z13=%xmm9,>in13=%rcx
+movd   %xmm9,%rcx
+
+# qhasm:   in14 = z14
+# asm 1: movd   <z14=int6464#4,>in14=int64#5
+# asm 2: movd   <z14=%xmm3,>in14=%r8
+movd   %xmm3,%r8
+
+# qhasm:   in15 = z15
+# asm 1: movd   <z15=int6464#3,>in15=int64#6
+# asm 2: movd   <z15=%xmm2,>in15=%r9
+movd   %xmm2,%r9
+
+# qhasm:   z12 <<<= 96
+# asm 1: pshufd $0x39,<z12=int6464#14,<z12=int6464#14
+# asm 2: pshufd $0x39,<z12=%xmm13,<z12=%xmm13
+pshufd $0x39,%xmm13,%xmm13
+
+# qhasm:   z13 <<<= 96
+# asm 1: pshufd $0x39,<z13=int6464#10,<z13=int6464#10
+# asm 2: pshufd $0x39,<z13=%xmm9,<z13=%xmm9
+pshufd $0x39,%xmm9,%xmm9
+
+# qhasm:   z14 <<<= 96
+# asm 1: pshufd $0x39,<z14=int6464#4,<z14=int6464#4
+# asm 2: pshufd $0x39,<z14=%xmm3,<z14=%xmm3
+pshufd $0x39,%xmm3,%xmm3
+
+# qhasm:   z15 <<<= 96
+# asm 1: pshufd $0x39,<z15=int6464#3,<z15=int6464#3
+# asm 2: pshufd $0x39,<z15=%xmm2,<z15=%xmm2
+pshufd $0x39,%xmm2,%xmm2
+
+# qhasm:   (uint32) in12 ^= *(uint32 *) (m + 48)
+# asm 1: xorl 48(<m=int64#2),<in12=int64#3d
+# asm 2: xorl 48(<m=%rsi),<in12=%edx
+xorl 48(%rsi),%edx
+
+# qhasm:   (uint32) in13 ^= *(uint32 *) (m + 52)
+# asm 1: xorl 52(<m=int64#2),<in13=int64#4d
+# asm 2: xorl 52(<m=%rsi),<in13=%ecx
+xorl 52(%rsi),%ecx
+
+# qhasm:   (uint32) in14 ^= *(uint32 *) (m + 56)
+# asm 1: xorl 56(<m=int64#2),<in14=int64#5d
+# asm 2: xorl 56(<m=%rsi),<in14=%r8d
+xorl 56(%rsi),%r8d
+
+# qhasm:   (uint32) in15 ^= *(uint32 *) (m + 60)
+# asm 1: xorl 60(<m=int64#2),<in15=int64#6d
+# asm 2: xorl 60(<m=%rsi),<in15=%r9d
+xorl 60(%rsi),%r9d
+
+# qhasm:   *(uint32 *) (out + 48) = in12
+# asm 1: movl   <in12=int64#3d,48(<out=int64#1)
+# asm 2: movl   <in12=%edx,48(<out=%rdi)
+movl   %edx,48(%rdi)
+
+# qhasm:   *(uint32 *) (out + 52) = in13
+# asm 1: movl   <in13=int64#4d,52(<out=int64#1)
+# asm 2: movl   <in13=%ecx,52(<out=%rdi)
+movl   %ecx,52(%rdi)
+
+# qhasm:   *(uint32 *) (out + 56) = in14
+# asm 1: movl   <in14=int64#5d,56(<out=int64#1)
+# asm 2: movl   <in14=%r8d,56(<out=%rdi)
+movl   %r8d,56(%rdi)
+
+# qhasm:   *(uint32 *) (out + 60) = in15
+# asm 1: movl   <in15=int64#6d,60(<out=int64#1)
+# asm 2: movl   <in15=%r9d,60(<out=%rdi)
+movl   %r9d,60(%rdi)
+
+# qhasm:   in12 = z12
+# asm 1: movd   <z12=int6464#14,>in12=int64#3
+# asm 2: movd   <z12=%xmm13,>in12=%rdx
+movd   %xmm13,%rdx
+
+# qhasm:   in13 = z13
+# asm 1: movd   <z13=int6464#10,>in13=int64#4
+# asm 2: movd   <z13=%xmm9,>in13=%rcx
+movd   %xmm9,%rcx
+
+# qhasm:   in14 = z14
+# asm 1: movd   <z14=int6464#4,>in14=int64#5
+# asm 2: movd   <z14=%xmm3,>in14=%r8
+movd   %xmm3,%r8
+
+# qhasm:   in15 = z15
+# asm 1: movd   <z15=int6464#3,>in15=int64#6
+# asm 2: movd   <z15=%xmm2,>in15=%r9
+movd   %xmm2,%r9
+
+# qhasm:   z12 <<<= 96
+# asm 1: pshufd $0x39,<z12=int6464#14,<z12=int6464#14
+# asm 2: pshufd $0x39,<z12=%xmm13,<z12=%xmm13
+pshufd $0x39,%xmm13,%xmm13
+
+# qhasm:   z13 <<<= 96
+# asm 1: pshufd $0x39,<z13=int6464#10,<z13=int6464#10
+# asm 2: pshufd $0x39,<z13=%xmm9,<z13=%xmm9
+pshufd $0x39,%xmm9,%xmm9
+
+# qhasm:   z14 <<<= 96
+# asm 1: pshufd $0x39,<z14=int6464#4,<z14=int6464#4
+# asm 2: pshufd $0x39,<z14=%xmm3,<z14=%xmm3
+pshufd $0x39,%xmm3,%xmm3
+
+# qhasm:   z15 <<<= 96
+# asm 1: pshufd $0x39,<z15=int6464#3,<z15=int6464#3
+# asm 2: pshufd $0x39,<z15=%xmm2,<z15=%xmm2
+pshufd $0x39,%xmm2,%xmm2
+
+# qhasm:   (uint32) in12 ^= *(uint32 *) (m + 112)
+# asm 1: xorl 112(<m=int64#2),<in12=int64#3d
+# asm 2: xorl 112(<m=%rsi),<in12=%edx
+xorl 112(%rsi),%edx
+
+# qhasm:   (uint32) in13 ^= *(uint32 *) (m + 116)
+# asm 1: xorl 116(<m=int64#2),<in13=int64#4d
+# asm 2: xorl 116(<m=%rsi),<in13=%ecx
+xorl 116(%rsi),%ecx
+
+# qhasm:   (uint32) in14 ^= *(uint32 *) (m + 120)
+# asm 1: xorl 120(<m=int64#2),<in14=int64#5d
+# asm 2: xorl 120(<m=%rsi),<in14=%r8d
+xorl 120(%rsi),%r8d
+
+# qhasm:   (uint32) in15 ^= *(uint32 *) (m + 124)
+# asm 1: xorl 124(<m=int64#2),<in15=int64#6d
+# asm 2: xorl 124(<m=%rsi),<in15=%r9d
+xorl 124(%rsi),%r9d
+
+# qhasm:   *(uint32 *) (out + 112) = in12
+# asm 1: movl   <in12=int64#3d,112(<out=int64#1)
+# asm 2: movl   <in12=%edx,112(<out=%rdi)
+movl   %edx,112(%rdi)
+
+# qhasm:   *(uint32 *) (out + 116) = in13
+# asm 1: movl   <in13=int64#4d,116(<out=int64#1)
+# asm 2: movl   <in13=%ecx,116(<out=%rdi)
+movl   %ecx,116(%rdi)
+
+# qhasm:   *(uint32 *) (out + 120) = in14
+# asm 1: movl   <in14=int64#5d,120(<out=int64#1)
+# asm 2: movl   <in14=%r8d,120(<out=%rdi)
+movl   %r8d,120(%rdi)
+
+# qhasm:   *(uint32 *) (out + 124) = in15
+# asm 1: movl   <in15=int64#6d,124(<out=int64#1)
+# asm 2: movl   <in15=%r9d,124(<out=%rdi)
+movl   %r9d,124(%rdi)
+
+# qhasm:   in12 = z12
+# asm 1: movd   <z12=int6464#14,>in12=int64#3
+# asm 2: movd   <z12=%xmm13,>in12=%rdx
+movd   %xmm13,%rdx
+
+# qhasm:   in13 = z13
+# asm 1: movd   <z13=int6464#10,>in13=int64#4
+# asm 2: movd   <z13=%xmm9,>in13=%rcx
+movd   %xmm9,%rcx
+
+# qhasm:   in14 = z14
+# asm 1: movd   <z14=int6464#4,>in14=int64#5
+# asm 2: movd   <z14=%xmm3,>in14=%r8
+movd   %xmm3,%r8
+
+# qhasm:   in15 = z15
+# asm 1: movd   <z15=int6464#3,>in15=int64#6
+# asm 2: movd   <z15=%xmm2,>in15=%r9
+movd   %xmm2,%r9
+
+# qhasm:   z12 <<<= 96
+# asm 1: pshufd $0x39,<z12=int6464#14,<z12=int6464#14
+# asm 2: pshufd $0x39,<z12=%xmm13,<z12=%xmm13
+pshufd $0x39,%xmm13,%xmm13
+
+# qhasm:   z13 <<<= 96
+# asm 1: pshufd $0x39,<z13=int6464#10,<z13=int6464#10
+# asm 2: pshufd $0x39,<z13=%xmm9,<z13=%xmm9
+pshufd $0x39,%xmm9,%xmm9
+
+# qhasm:   z14 <<<= 96
+# asm 1: pshufd $0x39,<z14=int6464#4,<z14=int6464#4
+# asm 2: pshufd $0x39,<z14=%xmm3,<z14=%xmm3
+pshufd $0x39,%xmm3,%xmm3
+
+# qhasm:   z15 <<<= 96
+# asm 1: pshufd $0x39,<z15=int6464#3,<z15=int6464#3
+# asm 2: pshufd $0x39,<z15=%xmm2,<z15=%xmm2
+pshufd $0x39,%xmm2,%xmm2
+
+# qhasm:   (uint32) in12 ^= *(uint32 *) (m + 176)
+# asm 1: xorl 176(<m=int64#2),<in12=int64#3d
+# asm 2: xorl 176(<m=%rsi),<in12=%edx
+xorl 176(%rsi),%edx
+
+# qhasm:   (uint32) in13 ^= *(uint32 *) (m + 180)
+# asm 1: xorl 180(<m=int64#2),<in13=int64#4d
+# asm 2: xorl 180(<m=%rsi),<in13=%ecx
+xorl 180(%rsi),%ecx
+
+# qhasm:   (uint32) in14 ^= *(uint32 *) (m + 184)
+# asm 1: xorl 184(<m=int64#2),<in14=int64#5d
+# asm 2: xorl 184(<m=%rsi),<in14=%r8d
+xorl 184(%rsi),%r8d
+
+# qhasm:   (uint32) in15 ^= *(uint32 *) (m + 188)
+# asm 1: xorl 188(<m=int64#2),<in15=int64#6d
+# asm 2: xorl 188(<m=%rsi),<in15=%r9d
+xorl 188(%rsi),%r9d
+
+# qhasm:   *(uint32 *) (out + 176) = in12
+# asm 1: movl   <in12=int64#3d,176(<out=int64#1)
+# asm 2: movl   <in12=%edx,176(<out=%rdi)
+movl   %edx,176(%rdi)
+
+# qhasm:   *(uint32 *) (out + 180) = in13
+# asm 1: movl   <in13=int64#4d,180(<out=int64#1)
+# asm 2: movl   <in13=%ecx,180(<out=%rdi)
+movl   %ecx,180(%rdi)
+
+# qhasm:   *(uint32 *) (out + 184) = in14
+# asm 1: movl   <in14=int64#5d,184(<out=int64#1)
+# asm 2: movl   <in14=%r8d,184(<out=%rdi)
+movl   %r8d,184(%rdi)
+
+# qhasm:   *(uint32 *) (out + 188) = in15
+# asm 1: movl   <in15=int64#6d,188(<out=int64#1)
+# asm 2: movl   <in15=%r9d,188(<out=%rdi)
+movl   %r9d,188(%rdi)
+
+# qhasm:   in12 = z12
+# asm 1: movd   <z12=int6464#14,>in12=int64#3
+# asm 2: movd   <z12=%xmm13,>in12=%rdx
+movd   %xmm13,%rdx
+
+# qhasm:   in13 = z13
+# asm 1: movd   <z13=int6464#10,>in13=int64#4
+# asm 2: movd   <z13=%xmm9,>in13=%rcx
+movd   %xmm9,%rcx
+
+# qhasm:   in14 = z14
+# asm 1: movd   <z14=int6464#4,>in14=int64#5
+# asm 2: movd   <z14=%xmm3,>in14=%r8
+movd   %xmm3,%r8
+
+# qhasm:   in15 = z15
+# asm 1: movd   <z15=int6464#3,>in15=int64#6
+# asm 2: movd   <z15=%xmm2,>in15=%r9
+movd   %xmm2,%r9
+
+# qhasm:   (uint32) in12 ^= *(uint32 *) (m + 240)
+# asm 1: xorl 240(<m=int64#2),<in12=int64#3d
+# asm 2: xorl 240(<m=%rsi),<in12=%edx
+xorl 240(%rsi),%edx
+
+# qhasm:   (uint32) in13 ^= *(uint32 *) (m + 244)
+# asm 1: xorl 244(<m=int64#2),<in13=int64#4d
+# asm 2: xorl 244(<m=%rsi),<in13=%ecx
+xorl 244(%rsi),%ecx
+
+# qhasm:   (uint32) in14 ^= *(uint32 *) (m + 248)
+# asm 1: xorl 248(<m=int64#2),<in14=int64#5d
+# asm 2: xorl 248(<m=%rsi),<in14=%r8d
+xorl 248(%rsi),%r8d
+
+# qhasm:   (uint32) in15 ^= *(uint32 *) (m + 252)
+# asm 1: xorl 252(<m=int64#2),<in15=int64#6d
+# asm 2: xorl 252(<m=%rsi),<in15=%r9d
+xorl 252(%rsi),%r9d
+
+# qhasm:   *(uint32 *) (out + 240) = in12
+# asm 1: movl   <in12=int64#3d,240(<out=int64#1)
+# asm 2: movl   <in12=%edx,240(<out=%rdi)
+movl   %edx,240(%rdi)
+
+# qhasm:   *(uint32 *) (out + 244) = in13
+# asm 1: movl   <in13=int64#4d,244(<out=int64#1)
+# asm 2: movl   <in13=%ecx,244(<out=%rdi)
+movl   %ecx,244(%rdi)
+
+# qhasm:   *(uint32 *) (out + 248) = in14
+# asm 1: movl   <in14=int64#5d,248(<out=int64#1)
+# asm 2: movl   <in14=%r8d,248(<out=%rdi)
+movl   %r8d,248(%rdi)
+
+# qhasm:   *(uint32 *) (out + 252) = in15
+# asm 1: movl   <in15=int64#6d,252(<out=int64#1)
+# asm 2: movl   <in15=%r9d,252(<out=%rdi)
+movl   %r9d,252(%rdi)
+
+# qhasm:   bytes = bytes_backup
+# asm 1: movq <bytes_backup=stack64#8,>bytes=int64#6
+# asm 2: movq <bytes_backup=408(%rsp),>bytes=%r9
+movq 408(%rsp),%r9
+
+# qhasm:   bytes -= 256
+# asm 1: sub  $256,<bytes=int64#6
+# asm 2: sub  $256,<bytes=%r9
+sub  $256,%r9
+
+# qhasm:   m += 256
+# asm 1: add  $256,<m=int64#2
+# asm 2: add  $256,<m=%rsi
+add  $256,%rsi
+
+# qhasm:   out += 256
+# asm 1: add  $256,<out=int64#1
+# asm 2: add  $256,<out=%rdi
+add  $256,%rdi
+
+# qhasm:                            unsigned<? bytes - 256
+# asm 1: cmp  $256,<bytes=int64#6
+# asm 2: cmp  $256,<bytes=%r9
+cmp  $256,%r9
+# comment:fp stack unchanged by jump
+
+# qhasm:   goto bytesatleast256 if !unsigned<
+jae ._bytesatleast256
+
+# qhasm:                 unsigned>? bytes - 0
+# asm 1: cmp  $0,<bytes=int64#6
+# asm 2: cmp  $0,<bytes=%r9
+cmp  $0,%r9
+# comment:fp stack unchanged by jump
+
+# qhasm:   goto done if !unsigned>
+jbe ._done
+# comment:fp stack unchanged by fallthrough
+
+# qhasm: bytesbetween1and255:
+._bytesbetween1and255:
+
+# qhasm:                   unsigned<? bytes - 64
+# asm 1: cmp  $64,<bytes=int64#6
+# asm 2: cmp  $64,<bytes=%r9
+cmp  $64,%r9
+# comment:fp stack unchanged by jump
+
+# qhasm:   goto nocopy if !unsigned<
+jae ._nocopy
+
+# qhasm:     ctarget = out
+# asm 1: mov  <out=int64#1,>ctarget=int64#3
+# asm 2: mov  <out=%rdi,>ctarget=%rdx
+mov  %rdi,%rdx
+
+# qhasm:     out = &tmp
+# asm 1: leaq <tmp=stack512#1,>out=int64#1
+# asm 2: leaq <tmp=416(%rsp),>out=%rdi
+leaq 416(%rsp),%rdi
+
+# qhasm:     i = bytes
+# asm 1: mov  <bytes=int64#6,>i=int64#4
+# asm 2: mov  <bytes=%r9,>i=%rcx
+mov  %r9,%rcx
+
+# qhasm:     while (i) { *out++ = *m++; --i }
+rep movsb
+
+# qhasm:     out = &tmp
+# asm 1: leaq <tmp=stack512#1,>out=int64#1
+# asm 2: leaq <tmp=416(%rsp),>out=%rdi
+leaq 416(%rsp),%rdi
+
+# qhasm:     m = &tmp
+# asm 1: leaq <tmp=stack512#1,>m=int64#2
+# asm 2: leaq <tmp=416(%rsp),>m=%rsi
+leaq 416(%rsp),%rsi
+# comment:fp stack unchanged by fallthrough
+
+# qhasm:   nocopy:
+._nocopy:
+
+# qhasm:   bytes_backup = bytes
+# asm 1: movq <bytes=int64#6,>bytes_backup=stack64#8
+# asm 2: movq <bytes=%r9,>bytes_backup=408(%rsp)
+movq %r9,408(%rsp)
+
+# qhasm: diag0 = x0
+# asm 1: movdqa <x0=stack128#4,>diag0=int6464#1
+# asm 2: movdqa <x0=48(%rsp),>diag0=%xmm0
+movdqa 48(%rsp),%xmm0
+
+# qhasm: diag1 = x1
+# asm 1: movdqa <x1=stack128#1,>diag1=int6464#2
+# asm 2: movdqa <x1=0(%rsp),>diag1=%xmm1
+movdqa 0(%rsp),%xmm1
+
+# qhasm: diag2 = x2
+# asm 1: movdqa <x2=stack128#2,>diag2=int6464#3
+# asm 2: movdqa <x2=16(%rsp),>diag2=%xmm2
+movdqa 16(%rsp),%xmm2
+
+# qhasm: diag3 = x3
+# asm 1: movdqa <x3=stack128#3,>diag3=int6464#4
+# asm 2: movdqa <x3=32(%rsp),>diag3=%xmm3
+movdqa 32(%rsp),%xmm3
+
+# qhasm:                     a0 = diag1
+# asm 1: movdqa <diag1=int6464#2,>a0=int6464#5
+# asm 2: movdqa <diag1=%xmm1,>a0=%xmm4
+movdqa %xmm1,%xmm4
+
+# qhasm: i = 12
+# asm 1: mov  $12,>i=int64#4
+# asm 2: mov  $12,>i=%rcx
+mov  $12,%rcx
+
+# qhasm: mainloop2:
+._mainloop2:
+
+# qhasm: uint32323232        a0 += diag0
+# asm 1: paddd <diag0=int6464#1,<a0=int6464#5
+# asm 2: paddd <diag0=%xmm0,<a0=%xmm4
+paddd %xmm0,%xmm4
+
+# qhasm:                                 a1 = diag0
+# asm 1: movdqa <diag0=int6464#1,>a1=int6464#6
+# asm 2: movdqa <diag0=%xmm0,>a1=%xmm5
+movdqa %xmm0,%xmm5
+
+# qhasm:                     b0 = a0
+# asm 1: movdqa <a0=int6464#5,>b0=int6464#7
+# asm 2: movdqa <a0=%xmm4,>b0=%xmm6
+movdqa %xmm4,%xmm6
+
+# qhasm: uint32323232        a0 <<= 7
+# asm 1: pslld $7,<a0=int6464#5
+# asm 2: pslld $7,<a0=%xmm4
+pslld $7,%xmm4
+
+# qhasm: uint32323232        b0 >>= 25
+# asm 1: psrld $25,<b0=int6464#7
+# asm 2: psrld $25,<b0=%xmm6
+psrld $25,%xmm6
+
+# qhasm:                 diag3 ^= a0
+# asm 1: pxor  <a0=int6464#5,<diag3=int6464#4
+# asm 2: pxor  <a0=%xmm4,<diag3=%xmm3
+pxor  %xmm4,%xmm3
+
+# qhasm:                 diag3 ^= b0
+# asm 1: pxor  <b0=int6464#7,<diag3=int6464#4
+# asm 2: pxor  <b0=%xmm6,<diag3=%xmm3
+pxor  %xmm6,%xmm3
+
+# qhasm: uint32323232                        a1 += diag3
+# asm 1: paddd <diag3=int6464#4,<a1=int6464#6
+# asm 2: paddd <diag3=%xmm3,<a1=%xmm5
+paddd %xmm3,%xmm5
+
+# qhasm:                                                 a2 = diag3
+# asm 1: movdqa <diag3=int6464#4,>a2=int6464#5
+# asm 2: movdqa <diag3=%xmm3,>a2=%xmm4
+movdqa %xmm3,%xmm4
+
+# qhasm:                                     b1 = a1
+# asm 1: movdqa <a1=int6464#6,>b1=int6464#7
+# asm 2: movdqa <a1=%xmm5,>b1=%xmm6
+movdqa %xmm5,%xmm6
+
+# qhasm: uint32323232                        a1 <<= 9
+# asm 1: pslld $9,<a1=int6464#6
+# asm 2: pslld $9,<a1=%xmm5
+pslld $9,%xmm5
+
+# qhasm: uint32323232                        b1 >>= 23
+# asm 1: psrld $23,<b1=int6464#7
+# asm 2: psrld $23,<b1=%xmm6
+psrld $23,%xmm6
+
+# qhasm:                                 diag2 ^= a1
+# asm 1: pxor  <a1=int6464#6,<diag2=int6464#3
+# asm 2: pxor  <a1=%xmm5,<diag2=%xmm2
+pxor  %xmm5,%xmm2
+
+# qhasm:                 diag3 <<<= 32
+# asm 1: pshufd $0x93,<diag3=int6464#4,<diag3=int6464#4
+# asm 2: pshufd $0x93,<diag3=%xmm3,<diag3=%xmm3
+pshufd $0x93,%xmm3,%xmm3
+
+# qhasm:                                 diag2 ^= b1
+# asm 1: pxor  <b1=int6464#7,<diag2=int6464#3
+# asm 2: pxor  <b1=%xmm6,<diag2=%xmm2
+pxor  %xmm6,%xmm2
+
+# qhasm: uint32323232                                        a2 += diag2
+# asm 1: paddd <diag2=int6464#3,<a2=int6464#5
+# asm 2: paddd <diag2=%xmm2,<a2=%xmm4
+paddd %xmm2,%xmm4
+
+# qhasm:                                                                 a3 = diag2
+# asm 1: movdqa <diag2=int6464#3,>a3=int6464#6
+# asm 2: movdqa <diag2=%xmm2,>a3=%xmm5
+movdqa %xmm2,%xmm5
+
+# qhasm:                                                     b2 = a2
+# asm 1: movdqa <a2=int6464#5,>b2=int6464#7
+# asm 2: movdqa <a2=%xmm4,>b2=%xmm6
+movdqa %xmm4,%xmm6
+
+# qhasm: uint32323232                                        a2 <<= 13
+# asm 1: pslld $13,<a2=int6464#5
+# asm 2: pslld $13,<a2=%xmm4
+pslld $13,%xmm4
+
+# qhasm: uint32323232                                        b2 >>= 19
+# asm 1: psrld $19,<b2=int6464#7
+# asm 2: psrld $19,<b2=%xmm6
+psrld $19,%xmm6
+
+# qhasm:                                                 diag1 ^= a2
+# asm 1: pxor  <a2=int6464#5,<diag1=int6464#2
+# asm 2: pxor  <a2=%xmm4,<diag1=%xmm1
+pxor  %xmm4,%xmm1
+
+# qhasm:                                 diag2 <<<= 64
+# asm 1: pshufd $0x4e,<diag2=int6464#3,<diag2=int6464#3
+# asm 2: pshufd $0x4e,<diag2=%xmm2,<diag2=%xmm2
+pshufd $0x4e,%xmm2,%xmm2
+
+# qhasm:                                                 diag1 ^= b2
+# asm 1: pxor  <b2=int6464#7,<diag1=int6464#2
+# asm 2: pxor  <b2=%xmm6,<diag1=%xmm1
+pxor  %xmm6,%xmm1
+
+# qhasm: uint32323232                                                        a3 += diag1
+# asm 1: paddd <diag1=int6464#2,<a3=int6464#6
+# asm 2: paddd <diag1=%xmm1,<a3=%xmm5
+paddd %xmm1,%xmm5
+
+# qhasm:                 a4 = diag3
+# asm 1: movdqa <diag3=int6464#4,>a4=int6464#5
+# asm 2: movdqa <diag3=%xmm3,>a4=%xmm4
+movdqa %xmm3,%xmm4
+
+# qhasm:                                                                     b3 = a3
+# asm 1: movdqa <a3=int6464#6,>b3=int6464#7
+# asm 2: movdqa <a3=%xmm5,>b3=%xmm6
+movdqa %xmm5,%xmm6
+
+# qhasm: uint32323232                                                        a3 <<= 18
+# asm 1: pslld $18,<a3=int6464#6
+# asm 2: pslld $18,<a3=%xmm5
+pslld $18,%xmm5
+
+# qhasm: uint32323232                                                        b3 >>= 14
+# asm 1: psrld $14,<b3=int6464#7
+# asm 2: psrld $14,<b3=%xmm6
+psrld $14,%xmm6
+
+# qhasm:                                                                 diag0 ^= a3
+# asm 1: pxor  <a3=int6464#6,<diag0=int6464#1
+# asm 2: pxor  <a3=%xmm5,<diag0=%xmm0
+pxor  %xmm5,%xmm0
+
+# qhasm:                                                 diag1 <<<= 96
+# asm 1: pshufd $0x39,<diag1=int6464#2,<diag1=int6464#2
+# asm 2: pshufd $0x39,<diag1=%xmm1,<diag1=%xmm1
+pshufd $0x39,%xmm1,%xmm1
+
+# qhasm:                                                                 diag0 ^= b3
+# asm 1: pxor  <b3=int6464#7,<diag0=int6464#1
+# asm 2: pxor  <b3=%xmm6,<diag0=%xmm0
+pxor  %xmm6,%xmm0
+
+# qhasm: uint32323232        a4 += diag0
+# asm 1: paddd <diag0=int6464#1,<a4=int6464#5
+# asm 2: paddd <diag0=%xmm0,<a4=%xmm4
+paddd %xmm0,%xmm4
+
+# qhasm:                                 a5 = diag0
+# asm 1: movdqa <diag0=int6464#1,>a5=int6464#6
+# asm 2: movdqa <diag0=%xmm0,>a5=%xmm5
+movdqa %xmm0,%xmm5
+
+# qhasm:                     b4 = a4
+# asm 1: movdqa <a4=int6464#5,>b4=int6464#7
+# asm 2: movdqa <a4=%xmm4,>b4=%xmm6
+movdqa %xmm4,%xmm6
+
+# qhasm: uint32323232        a4 <<= 7
+# asm 1: pslld $7,<a4=int6464#5
+# asm 2: pslld $7,<a4=%xmm4
+pslld $7,%xmm4
+
+# qhasm: uint32323232        b4 >>= 25
+# asm 1: psrld $25,<b4=int6464#7
+# asm 2: psrld $25,<b4=%xmm6
+psrld $25,%xmm6
+
+# qhasm:                 diag1 ^= a4
+# asm 1: pxor  <a4=int6464#5,<diag1=int6464#2
+# asm 2: pxor  <a4=%xmm4,<diag1=%xmm1
+pxor  %xmm4,%xmm1
+
+# qhasm:                 diag1 ^= b4
+# asm 1: pxor  <b4=int6464#7,<diag1=int6464#2
+# asm 2: pxor  <b4=%xmm6,<diag1=%xmm1
+pxor  %xmm6,%xmm1
+
+# qhasm: uint32323232                        a5 += diag1
+# asm 1: paddd <diag1=int6464#2,<a5=int6464#6
+# asm 2: paddd <diag1=%xmm1,<a5=%xmm5
+paddd %xmm1,%xmm5
+
+# qhasm:                                                 a6 = diag1
+# asm 1: movdqa <diag1=int6464#2,>a6=int6464#5
+# asm 2: movdqa <diag1=%xmm1,>a6=%xmm4
+movdqa %xmm1,%xmm4
+
+# qhasm:                                     b5 = a5
+# asm 1: movdqa <a5=int6464#6,>b5=int6464#7
+# asm 2: movdqa <a5=%xmm5,>b5=%xmm6
+movdqa %xmm5,%xmm6
+
+# qhasm: uint32323232                        a5 <<= 9
+# asm 1: pslld $9,<a5=int6464#6
+# asm 2: pslld $9,<a5=%xmm5
+pslld $9,%xmm5
+
+# qhasm: uint32323232                        b5 >>= 23
+# asm 1: psrld $23,<b5=int6464#7
+# asm 2: psrld $23,<b5=%xmm6
+psrld $23,%xmm6
+
+# qhasm:                                 diag2 ^= a5
+# asm 1: pxor  <a5=int6464#6,<diag2=int6464#3
+# asm 2: pxor  <a5=%xmm5,<diag2=%xmm2
+pxor  %xmm5,%xmm2
+
+# qhasm:                 diag1 <<<= 32
+# asm 1: pshufd $0x93,<diag1=int6464#2,<diag1=int6464#2
+# asm 2: pshufd $0x93,<diag1=%xmm1,<diag1=%xmm1
+pshufd $0x93,%xmm1,%xmm1
+
+# qhasm:                                 diag2 ^= b5
+# asm 1: pxor  <b5=int6464#7,<diag2=int6464#3
+# asm 2: pxor  <b5=%xmm6,<diag2=%xmm2
+pxor  %xmm6,%xmm2
+
+# qhasm: uint32323232                                        a6 += diag2
+# asm 1: paddd <diag2=int6464#3,<a6=int6464#5
+# asm 2: paddd <diag2=%xmm2,<a6=%xmm4
+paddd %xmm2,%xmm4
+
+# qhasm:                                                                 a7 = diag2
+# asm 1: movdqa <diag2=int6464#3,>a7=int6464#6
+# asm 2: movdqa <diag2=%xmm2,>a7=%xmm5
+movdqa %xmm2,%xmm5
+
+# qhasm:                                                     b6 = a6
+# asm 1: movdqa <a6=int6464#5,>b6=int6464#7
+# asm 2: movdqa <a6=%xmm4,>b6=%xmm6
+movdqa %xmm4,%xmm6
+
+# qhasm: uint32323232                                        a6 <<= 13
+# asm 1: pslld $13,<a6=int6464#5
+# asm 2: pslld $13,<a6=%xmm4
+pslld $13,%xmm4
+
+# qhasm: uint32323232                                        b6 >>= 19
+# asm 1: psrld $19,<b6=int6464#7
+# asm 2: psrld $19,<b6=%xmm6
+psrld $19,%xmm6
+
+# qhasm:                                                 diag3 ^= a6
+# asm 1: pxor  <a6=int6464#5,<diag3=int6464#4
+# asm 2: pxor  <a6=%xmm4,<diag3=%xmm3
+pxor  %xmm4,%xmm3
+
+# qhasm:                                 diag2 <<<= 64
+# asm 1: pshufd $0x4e,<diag2=int6464#3,<diag2=int6464#3
+# asm 2: pshufd $0x4e,<diag2=%xmm2,<diag2=%xmm2
+pshufd $0x4e,%xmm2,%xmm2
+
+# qhasm:                                                 diag3 ^= b6
+# asm 1: pxor  <b6=int6464#7,<diag3=int6464#4
+# asm 2: pxor  <b6=%xmm6,<diag3=%xmm3
+pxor  %xmm6,%xmm3
+
+# qhasm: uint32323232                                                        a7 += diag3
+# asm 1: paddd <diag3=int6464#4,<a7=int6464#6
+# asm 2: paddd <diag3=%xmm3,<a7=%xmm5
+paddd %xmm3,%xmm5
+
+# qhasm:                 a0 = diag1
+# asm 1: movdqa <diag1=int6464#2,>a0=int6464#5
+# asm 2: movdqa <diag1=%xmm1,>a0=%xmm4
+movdqa %xmm1,%xmm4
+
+# qhasm:                                                                     b7 = a7
+# asm 1: movdqa <a7=int6464#6,>b7=int6464#7
+# asm 2: movdqa <a7=%xmm5,>b7=%xmm6
+movdqa %xmm5,%xmm6
+
+# qhasm: uint32323232                                                        a7 <<= 18
+# asm 1: pslld $18,<a7=int6464#6
+# asm 2: pslld $18,<a7=%xmm5
+pslld $18,%xmm5
+
+# qhasm: uint32323232                                                        b7 >>= 14
+# asm 1: psrld $14,<b7=int6464#7
+# asm 2: psrld $14,<b7=%xmm6
+psrld $14,%xmm6
+
+# qhasm:                                                                 diag0 ^= a7
+# asm 1: pxor  <a7=int6464#6,<diag0=int6464#1
+# asm 2: pxor  <a7=%xmm5,<diag0=%xmm0
+pxor  %xmm5,%xmm0
+
+# qhasm:                                                 diag3 <<<= 96
+# asm 1: pshufd $0x39,<diag3=int6464#4,<diag3=int6464#4
+# asm 2: pshufd $0x39,<diag3=%xmm3,<diag3=%xmm3
+pshufd $0x39,%xmm3,%xmm3
+
+# qhasm:                                                                 diag0 ^= b7
+# asm 1: pxor  <b7=int6464#7,<diag0=int6464#1
+# asm 2: pxor  <b7=%xmm6,<diag0=%xmm0
+pxor  %xmm6,%xmm0
+
+# qhasm: uint32323232        a0 += diag0
+# asm 1: paddd <diag0=int6464#1,<a0=int6464#5
+# asm 2: paddd <diag0=%xmm0,<a0=%xmm4
+paddd %xmm0,%xmm4
+
+# qhasm:                                 a1 = diag0
+# asm 1: movdqa <diag0=int6464#1,>a1=int6464#6
+# asm 2: movdqa <diag0=%xmm0,>a1=%xmm5
+movdqa %xmm0,%xmm5
+
+# qhasm:                     b0 = a0
+# asm 1: movdqa <a0=int6464#5,>b0=int6464#7
+# asm 2: movdqa <a0=%xmm4,>b0=%xmm6
+movdqa %xmm4,%xmm6
+
+# qhasm: uint32323232        a0 <<= 7
+# asm 1: pslld $7,<a0=int6464#5
+# asm 2: pslld $7,<a0=%xmm4
+pslld $7,%xmm4
+
+# qhasm: uint32323232        b0 >>= 25
+# asm 1: psrld $25,<b0=int6464#7
+# asm 2: psrld $25,<b0=%xmm6
+psrld $25,%xmm6
+
+# qhasm:                 diag3 ^= a0
+# asm 1: pxor  <a0=int6464#5,<diag3=int6464#4
+# asm 2: pxor  <a0=%xmm4,<diag3=%xmm3
+pxor  %xmm4,%xmm3
+
+# qhasm:                 diag3 ^= b0
+# asm 1: pxor  <b0=int6464#7,<diag3=int6464#4
+# asm 2: pxor  <b0=%xmm6,<diag3=%xmm3
+pxor  %xmm6,%xmm3
+
+# qhasm: uint32323232                        a1 += diag3
+# asm 1: paddd <diag3=int6464#4,<a1=int6464#6
+# asm 2: paddd <diag3=%xmm3,<a1=%xmm5
+paddd %xmm3,%xmm5
+
+# qhasm:                                                 a2 = diag3
+# asm 1: movdqa <diag3=int6464#4,>a2=int6464#5
+# asm 2: movdqa <diag3=%xmm3,>a2=%xmm4
+movdqa %xmm3,%xmm4
+
+# qhasm:                                     b1 = a1
+# asm 1: movdqa <a1=int6464#6,>b1=int6464#7
+# asm 2: movdqa <a1=%xmm5,>b1=%xmm6
+movdqa %xmm5,%xmm6
+
+# qhasm: uint32323232                        a1 <<= 9
+# asm 1: pslld $9,<a1=int6464#6
+# asm 2: pslld $9,<a1=%xmm5
+pslld $9,%xmm5
+
+# qhasm: uint32323232                        b1 >>= 23
+# asm 1: psrld $23,<b1=int6464#7
+# asm 2: psrld $23,<b1=%xmm6
+psrld $23,%xmm6
+
+# qhasm:                                 diag2 ^= a1
+# asm 1: pxor  <a1=int6464#6,<diag2=int6464#3
+# asm 2: pxor  <a1=%xmm5,<diag2=%xmm2
+pxor  %xmm5,%xmm2
+
+# qhasm:                 diag3 <<<= 32
+# asm 1: pshufd $0x93,<diag3=int6464#4,<diag3=int6464#4
+# asm 2: pshufd $0x93,<diag3=%xmm3,<diag3=%xmm3
+pshufd $0x93,%xmm3,%xmm3
+
+# qhasm:                                 diag2 ^= b1
+# asm 1: pxor  <b1=int6464#7,<diag2=int6464#3
+# asm 2: pxor  <b1=%xmm6,<diag2=%xmm2
+pxor  %xmm6,%xmm2
+
+# qhasm: uint32323232                                        a2 += diag2
+# asm 1: paddd <diag2=int6464#3,<a2=int6464#5
+# asm 2: paddd <diag2=%xmm2,<a2=%xmm4
+paddd %xmm2,%xmm4
+
+# qhasm:                                                                 a3 = diag2
+# asm 1: movdqa <diag2=int6464#3,>a3=int6464#6
+# asm 2: movdqa <diag2=%xmm2,>a3=%xmm5
+movdqa %xmm2,%xmm5
+
+# qhasm:                                                     b2 = a2
+# asm 1: movdqa <a2=int6464#5,>b2=int6464#7
+# asm 2: movdqa <a2=%xmm4,>b2=%xmm6
+movdqa %xmm4,%xmm6
+
+# qhasm: uint32323232                                        a2 <<= 13
+# asm 1: pslld $13,<a2=int6464#5
+# asm 2: pslld $13,<a2=%xmm4
+pslld $13,%xmm4
+
+# qhasm: uint32323232                                        b2 >>= 19
+# asm 1: psrld $19,<b2=int6464#7
+# asm 2: psrld $19,<b2=%xmm6
+psrld $19,%xmm6
+
+# qhasm:                                                 diag1 ^= a2
+# asm 1: pxor  <a2=int6464#5,<diag1=int6464#2
+# asm 2: pxor  <a2=%xmm4,<diag1=%xmm1
+pxor  %xmm4,%xmm1
+
+# qhasm:                                 diag2 <<<= 64
+# asm 1: pshufd $0x4e,<diag2=int6464#3,<diag2=int6464#3
+# asm 2: pshufd $0x4e,<diag2=%xmm2,<diag2=%xmm2
+pshufd $0x4e,%xmm2,%xmm2
+
+# qhasm:                                                 diag1 ^= b2
+# asm 1: pxor  <b2=int6464#7,<diag1=int6464#2
+# asm 2: pxor  <b2=%xmm6,<diag1=%xmm1
+pxor  %xmm6,%xmm1
+
+# qhasm: uint32323232                                                        a3 += diag1
+# asm 1: paddd <diag1=int6464#2,<a3=int6464#6
+# asm 2: paddd <diag1=%xmm1,<a3=%xmm5
+paddd %xmm1,%xmm5
+
+# qhasm:                 a4 = diag3
+# asm 1: movdqa <diag3=int6464#4,>a4=int6464#5
+# asm 2: movdqa <diag3=%xmm3,>a4=%xmm4
+movdqa %xmm3,%xmm4
+
+# qhasm:                                                                     b3 = a3
+# asm 1: movdqa <a3=int6464#6,>b3=int6464#7
+# asm 2: movdqa <a3=%xmm5,>b3=%xmm6
+movdqa %xmm5,%xmm6
+
+# qhasm: uint32323232                                                        a3 <<= 18
+# asm 1: pslld $18,<a3=int6464#6
+# asm 2: pslld $18,<a3=%xmm5
+pslld $18,%xmm5
+
+# qhasm: uint32323232                                                        b3 >>= 14
+# asm 1: psrld $14,<b3=int6464#7
+# asm 2: psrld $14,<b3=%xmm6
+psrld $14,%xmm6
+
+# qhasm:                                                                 diag0 ^= a3
+# asm 1: pxor  <a3=int6464#6,<diag0=int6464#1
+# asm 2: pxor  <a3=%xmm5,<diag0=%xmm0
+pxor  %xmm5,%xmm0
+
+# qhasm:                                                 diag1 <<<= 96
+# asm 1: pshufd $0x39,<diag1=int6464#2,<diag1=int6464#2
+# asm 2: pshufd $0x39,<diag1=%xmm1,<diag1=%xmm1
+pshufd $0x39,%xmm1,%xmm1
+
+# qhasm:                                                                 diag0 ^= b3
+# asm 1: pxor  <b3=int6464#7,<diag0=int6464#1
+# asm 2: pxor  <b3=%xmm6,<diag0=%xmm0
+pxor  %xmm6,%xmm0
+
+# qhasm: uint32323232        a4 += diag0
+# asm 1: paddd <diag0=int6464#1,<a4=int6464#5
+# asm 2: paddd <diag0=%xmm0,<a4=%xmm4
+paddd %xmm0,%xmm4
+
+# qhasm:                                 a5 = diag0
+# asm 1: movdqa <diag0=int6464#1,>a5=int6464#6
+# asm 2: movdqa <diag0=%xmm0,>a5=%xmm5
+movdqa %xmm0,%xmm5
+
+# qhasm:                     b4 = a4
+# asm 1: movdqa <a4=int6464#5,>b4=int6464#7
+# asm 2: movdqa <a4=%xmm4,>b4=%xmm6
+movdqa %xmm4,%xmm6
+
+# qhasm: uint32323232        a4 <<= 7
+# asm 1: pslld $7,<a4=int6464#5
+# asm 2: pslld $7,<a4=%xmm4
+pslld $7,%xmm4
+
+# qhasm: uint32323232        b4 >>= 25
+# asm 1: psrld $25,<b4=int6464#7
+# asm 2: psrld $25,<b4=%xmm6
+psrld $25,%xmm6
+
+# qhasm:                 diag1 ^= a4
+# asm 1: pxor  <a4=int6464#5,<diag1=int6464#2
+# asm 2: pxor  <a4=%xmm4,<diag1=%xmm1
+pxor  %xmm4,%xmm1
+
+# qhasm:                 diag1 ^= b4
+# asm 1: pxor  <b4=int6464#7,<diag1=int6464#2
+# asm 2: pxor  <b4=%xmm6,<diag1=%xmm1
+pxor  %xmm6,%xmm1
+
+# qhasm: uint32323232                        a5 += diag1
+# asm 1: paddd <diag1=int6464#2,<a5=int6464#6
+# asm 2: paddd <diag1=%xmm1,<a5=%xmm5
+paddd %xmm1,%xmm5
+
+# qhasm:                                                 a6 = diag1
+# asm 1: movdqa <diag1=int6464#2,>a6=int6464#5
+# asm 2: movdqa <diag1=%xmm1,>a6=%xmm4
+movdqa %xmm1,%xmm4
+
+# qhasm:                                     b5 = a5
+# asm 1: movdqa <a5=int6464#6,>b5=int6464#7
+# asm 2: movdqa <a5=%xmm5,>b5=%xmm6
+movdqa %xmm5,%xmm6
+
+# qhasm: uint32323232                        a5 <<= 9
+# asm 1: pslld $9,<a5=int6464#6
+# asm 2: pslld $9,<a5=%xmm5
+pslld $9,%xmm5
+
+# qhasm: uint32323232                        b5 >>= 23
+# asm 1: psrld $23,<b5=int6464#7
+# asm 2: psrld $23,<b5=%xmm6
+psrld $23,%xmm6
+
+# qhasm:                                 diag2 ^= a5
+# asm 1: pxor  <a5=int6464#6,<diag2=int6464#3
+# asm 2: pxor  <a5=%xmm5,<diag2=%xmm2
+pxor  %xmm5,%xmm2
+
+# qhasm:                 diag1 <<<= 32
+# asm 1: pshufd $0x93,<diag1=int6464#2,<diag1=int6464#2
+# asm 2: pshufd $0x93,<diag1=%xmm1,<diag1=%xmm1
+pshufd $0x93,%xmm1,%xmm1
+
+# qhasm:                                 diag2 ^= b5
+# asm 1: pxor  <b5=int6464#7,<diag2=int6464#3
+# asm 2: pxor  <b5=%xmm6,<diag2=%xmm2
+pxor  %xmm6,%xmm2
+
+# qhasm: uint32323232                                        a6 += diag2
+# asm 1: paddd <diag2=int6464#3,<a6=int6464#5
+# asm 2: paddd <diag2=%xmm2,<a6=%xmm4
+paddd %xmm2,%xmm4
+
+# qhasm:                                                                 a7 = diag2
+# asm 1: movdqa <diag2=int6464#3,>a7=int6464#6
+# asm 2: movdqa <diag2=%xmm2,>a7=%xmm5
+movdqa %xmm2,%xmm5
+
+# qhasm:                                                     b6 = a6
+# asm 1: movdqa <a6=int6464#5,>b6=int6464#7
+# asm 2: movdqa <a6=%xmm4,>b6=%xmm6
+movdqa %xmm4,%xmm6
+
+# qhasm: uint32323232                                        a6 <<= 13
+# asm 1: pslld $13,<a6=int6464#5
+# asm 2: pslld $13,<a6=%xmm4
+pslld $13,%xmm4
+
+# qhasm: uint32323232                                        b6 >>= 19
+# asm 1: psrld $19,<b6=int6464#7
+# asm 2: psrld $19,<b6=%xmm6
+psrld $19,%xmm6
+
+# qhasm:                                                 diag3 ^= a6
+# asm 1: pxor  <a6=int6464#5,<diag3=int6464#4
+# asm 2: pxor  <a6=%xmm4,<diag3=%xmm3
+pxor  %xmm4,%xmm3
+
+# qhasm:                                 diag2 <<<= 64
+# asm 1: pshufd $0x4e,<diag2=int6464#3,<diag2=int6464#3
+# asm 2: pshufd $0x4e,<diag2=%xmm2,<diag2=%xmm2
+pshufd $0x4e,%xmm2,%xmm2
+
+# qhasm:                                                 diag3 ^= b6
+# asm 1: pxor  <b6=int6464#7,<diag3=int6464#4
+# asm 2: pxor  <b6=%xmm6,<diag3=%xmm3
+pxor  %xmm6,%xmm3
+
+# qhasm:                  unsigned>? i -= 4
+# asm 1: sub  $4,<i=int64#4
+# asm 2: sub  $4,<i=%rcx
+sub  $4,%rcx
+
+# qhasm: uint32323232                                                        a7 += diag3
+# asm 1: paddd <diag3=int6464#4,<a7=int6464#6
+# asm 2: paddd <diag3=%xmm3,<a7=%xmm5
+paddd %xmm3,%xmm5
+
+# qhasm:                 a0 = diag1
+# asm 1: movdqa <diag1=int6464#2,>a0=int6464#5
+# asm 2: movdqa <diag1=%xmm1,>a0=%xmm4
+movdqa %xmm1,%xmm4
+
+# qhasm:                                                                     b7 = a7
+# asm 1: movdqa <a7=int6464#6,>b7=int6464#7
+# asm 2: movdqa <a7=%xmm5,>b7=%xmm6
+movdqa %xmm5,%xmm6
+
+# qhasm: uint32323232                                                        a7 <<= 18
+# asm 1: pslld $18,<a7=int6464#6
+# asm 2: pslld $18,<a7=%xmm5
+pslld $18,%xmm5
+
+# qhasm:                 b0 = 0
+# asm 1: pxor   >b0=int6464#8,>b0=int6464#8
+# asm 2: pxor   >b0=%xmm7,>b0=%xmm7
+pxor   %xmm7,%xmm7
+
+# qhasm: uint32323232                                                        b7 >>= 14
+# asm 1: psrld $14,<b7=int6464#7
+# asm 2: psrld $14,<b7=%xmm6
+psrld $14,%xmm6
+
+# qhasm:                                                                 diag0 ^= a7
+# asm 1: pxor  <a7=int6464#6,<diag0=int6464#1
+# asm 2: pxor  <a7=%xmm5,<diag0=%xmm0
+pxor  %xmm5,%xmm0
+
+# qhasm:                                                 diag3 <<<= 96
+# asm 1: pshufd $0x39,<diag3=int6464#4,<diag3=int6464#4
+# asm 2: pshufd $0x39,<diag3=%xmm3,<diag3=%xmm3
+pshufd $0x39,%xmm3,%xmm3
+
+# qhasm:                                                                 diag0 ^= b7
+# asm 1: pxor  <b7=int6464#7,<diag0=int6464#1
+# asm 2: pxor  <b7=%xmm6,<diag0=%xmm0
+pxor  %xmm6,%xmm0
+# comment:fp stack unchanged by jump
+
+# qhasm: goto mainloop2 if unsigned>
+ja ._mainloop2
+
+# qhasm: uint32323232 diag0 += x0
+# asm 1: paddd <x0=stack128#4,<diag0=int6464#1
+# asm 2: paddd <x0=48(%rsp),<diag0=%xmm0
+paddd 48(%rsp),%xmm0
+
+# qhasm: uint32323232 diag1 += x1
+# asm 1: paddd <x1=stack128#1,<diag1=int6464#2
+# asm 2: paddd <x1=0(%rsp),<diag1=%xmm1
+paddd 0(%rsp),%xmm1
+
+# qhasm: uint32323232 diag2 += x2
+# asm 1: paddd <x2=stack128#2,<diag2=int6464#3
+# asm 2: paddd <x2=16(%rsp),<diag2=%xmm2
+paddd 16(%rsp),%xmm2
+
+# qhasm: uint32323232 diag3 += x3
+# asm 1: paddd <x3=stack128#3,<diag3=int6464#4
+# asm 2: paddd <x3=32(%rsp),<diag3=%xmm3
+paddd 32(%rsp),%xmm3
+
+# qhasm: in0 = diag0
+# asm 1: movd   <diag0=int6464#1,>in0=int64#4
+# asm 2: movd   <diag0=%xmm0,>in0=%rcx
+movd   %xmm0,%rcx
+
+# qhasm: in12 = diag1
+# asm 1: movd   <diag1=int6464#2,>in12=int64#5
+# asm 2: movd   <diag1=%xmm1,>in12=%r8
+movd   %xmm1,%r8
+
+# qhasm: in8 = diag2
+# asm 1: movd   <diag2=int6464#3,>in8=int64#6
+# asm 2: movd   <diag2=%xmm2,>in8=%r9
+movd   %xmm2,%r9
+
+# qhasm: in4 = diag3
+# asm 1: movd   <diag3=int6464#4,>in4=int64#7
+# asm 2: movd   <diag3=%xmm3,>in4=%rax
+movd   %xmm3,%rax
+
+# qhasm: diag0 <<<= 96
+# asm 1: pshufd $0x39,<diag0=int6464#1,<diag0=int6464#1
+# asm 2: pshufd $0x39,<diag0=%xmm0,<diag0=%xmm0
+pshufd $0x39,%xmm0,%xmm0
+
+# qhasm: diag1 <<<= 96
+# asm 1: pshufd $0x39,<diag1=int6464#2,<diag1=int6464#2
+# asm 2: pshufd $0x39,<diag1=%xmm1,<diag1=%xmm1
+pshufd $0x39,%xmm1,%xmm1
+
+# qhasm: diag2 <<<= 96
+# asm 1: pshufd $0x39,<diag2=int6464#3,<diag2=int6464#3
+# asm 2: pshufd $0x39,<diag2=%xmm2,<diag2=%xmm2
+pshufd $0x39,%xmm2,%xmm2
+
+# qhasm: diag3 <<<= 96
+# asm 1: pshufd $0x39,<diag3=int6464#4,<diag3=int6464#4
+# asm 2: pshufd $0x39,<diag3=%xmm3,<diag3=%xmm3
+pshufd $0x39,%xmm3,%xmm3
+
+# qhasm: (uint32) in0 ^= *(uint32 *) (m + 0)
+# asm 1: xorl 0(<m=int64#2),<in0=int64#4d
+# asm 2: xorl 0(<m=%rsi),<in0=%ecx
+xorl 0(%rsi),%ecx
+
+# qhasm: (uint32) in12 ^= *(uint32 *) (m + 48)
+# asm 1: xorl 48(<m=int64#2),<in12=int64#5d
+# asm 2: xorl 48(<m=%rsi),<in12=%r8d
+xorl 48(%rsi),%r8d
+
+# qhasm: (uint32) in8 ^= *(uint32 *) (m + 32)
+# asm 1: xorl 32(<m=int64#2),<in8=int64#6d
+# asm 2: xorl 32(<m=%rsi),<in8=%r9d
+xorl 32(%rsi),%r9d
+
+# qhasm: (uint32) in4 ^= *(uint32 *) (m + 16)
+# asm 1: xorl 16(<m=int64#2),<in4=int64#7d
+# asm 2: xorl 16(<m=%rsi),<in4=%eax
+xorl 16(%rsi),%eax
+
+# qhasm: *(uint32 *) (out + 0) = in0
+# asm 1: movl   <in0=int64#4d,0(<out=int64#1)
+# asm 2: movl   <in0=%ecx,0(<out=%rdi)
+movl   %ecx,0(%rdi)
+
+# qhasm: *(uint32 *) (out + 48) = in12
+# asm 1: movl   <in12=int64#5d,48(<out=int64#1)
+# asm 2: movl   <in12=%r8d,48(<out=%rdi)
+movl   %r8d,48(%rdi)
+
+# qhasm: *(uint32 *) (out + 32) = in8
+# asm 1: movl   <in8=int64#6d,32(<out=int64#1)
+# asm 2: movl   <in8=%r9d,32(<out=%rdi)
+movl   %r9d,32(%rdi)
+
+# qhasm: *(uint32 *) (out + 16) = in4
+# asm 1: movl   <in4=int64#7d,16(<out=int64#1)
+# asm 2: movl   <in4=%eax,16(<out=%rdi)
+movl   %eax,16(%rdi)
+
+# qhasm: in5 = diag0
+# asm 1: movd   <diag0=int6464#1,>in5=int64#4
+# asm 2: movd   <diag0=%xmm0,>in5=%rcx
+movd   %xmm0,%rcx
+
+# qhasm: in1 = diag1
+# asm 1: movd   <diag1=int6464#2,>in1=int64#5
+# asm 2: movd   <diag1=%xmm1,>in1=%r8
+movd   %xmm1,%r8
+
+# qhasm: in13 = diag2
+# asm 1: movd   <diag2=int6464#3,>in13=int64#6
+# asm 2: movd   <diag2=%xmm2,>in13=%r9
+movd   %xmm2,%r9
+
+# qhasm: in9 = diag3
+# asm 1: movd   <diag3=int6464#4,>in9=int64#7
+# asm 2: movd   <diag3=%xmm3,>in9=%rax
+movd   %xmm3,%rax
+
+# qhasm: diag0 <<<= 96
+# asm 1: pshufd $0x39,<diag0=int6464#1,<diag0=int6464#1
+# asm 2: pshufd $0x39,<diag0=%xmm0,<diag0=%xmm0
+pshufd $0x39,%xmm0,%xmm0
+
+# qhasm: diag1 <<<= 96
+# asm 1: pshufd $0x39,<diag1=int6464#2,<diag1=int6464#2
+# asm 2: pshufd $0x39,<diag1=%xmm1,<diag1=%xmm1
+pshufd $0x39,%xmm1,%xmm1
+
+# qhasm: diag2 <<<= 96
+# asm 1: pshufd $0x39,<diag2=int6464#3,<diag2=int6464#3
+# asm 2: pshufd $0x39,<diag2=%xmm2,<diag2=%xmm2
+pshufd $0x39,%xmm2,%xmm2
+
+# qhasm: diag3 <<<= 96
+# asm 1: pshufd $0x39,<diag3=int6464#4,<diag3=int6464#4
+# asm 2: pshufd $0x39,<diag3=%xmm3,<diag3=%xmm3
+pshufd $0x39,%xmm3,%xmm3
+
+# qhasm: (uint32) in5 ^= *(uint32 *) (m + 20)
+# asm 1: xorl 20(<m=int64#2),<in5=int64#4d
+# asm 2: xorl 20(<m=%rsi),<in5=%ecx
+xorl 20(%rsi),%ecx
+
+# qhasm: (uint32) in1 ^= *(uint32 *) (m + 4)
+# asm 1: xorl 4(<m=int64#2),<in1=int64#5d
+# asm 2: xorl 4(<m=%rsi),<in1=%r8d
+xorl 4(%rsi),%r8d
+
+# qhasm: (uint32) in13 ^= *(uint32 *) (m + 52)
+# asm 1: xorl 52(<m=int64#2),<in13=int64#6d
+# asm 2: xorl 52(<m=%rsi),<in13=%r9d
+xorl 52(%rsi),%r9d
+
+# qhasm: (uint32) in9 ^= *(uint32 *) (m + 36)
+# asm 1: xorl 36(<m=int64#2),<in9=int64#7d
+# asm 2: xorl 36(<m=%rsi),<in9=%eax
+xorl 36(%rsi),%eax
+
+# qhasm: *(uint32 *) (out + 20) = in5
+# asm 1: movl   <in5=int64#4d,20(<out=int64#1)
+# asm 2: movl   <in5=%ecx,20(<out=%rdi)
+movl   %ecx,20(%rdi)
+
+# qhasm: *(uint32 *) (out + 4) = in1
+# asm 1: movl   <in1=int64#5d,4(<out=int64#1)
+# asm 2: movl   <in1=%r8d,4(<out=%rdi)
+movl   %r8d,4(%rdi)
+
+# qhasm: *(uint32 *) (out + 52) = in13
+# asm 1: movl   <in13=int64#6d,52(<out=int64#1)
+# asm 2: movl   <in13=%r9d,52(<out=%rdi)
+movl   %r9d,52(%rdi)
+
+# qhasm: *(uint32 *) (out + 36) = in9
+# asm 1: movl   <in9=int64#7d,36(<out=int64#1)
+# asm 2: movl   <in9=%eax,36(<out=%rdi)
+movl   %eax,36(%rdi)
+
+# qhasm: in10 = diag0
+# asm 1: movd   <diag0=int6464#1,>in10=int64#4
+# asm 2: movd   <diag0=%xmm0,>in10=%rcx
+movd   %xmm0,%rcx
+
+# qhasm: in6 = diag1
+# asm 1: movd   <diag1=int6464#2,>in6=int64#5
+# asm 2: movd   <diag1=%xmm1,>in6=%r8
+movd   %xmm1,%r8
+
+# qhasm: in2 = diag2
+# asm 1: movd   <diag2=int6464#3,>in2=int64#6
+# asm 2: movd   <diag2=%xmm2,>in2=%r9
+movd   %xmm2,%r9
+
+# qhasm: in14 = diag3
+# asm 1: movd   <diag3=int6464#4,>in14=int64#7
+# asm 2: movd   <diag3=%xmm3,>in14=%rax
+movd   %xmm3,%rax
+
+# qhasm: diag0 <<<= 96
+# asm 1: pshufd $0x39,<diag0=int6464#1,<diag0=int6464#1
+# asm 2: pshufd $0x39,<diag0=%xmm0,<diag0=%xmm0
+pshufd $0x39,%xmm0,%xmm0
+
+# qhasm: diag1 <<<= 96
+# asm 1: pshufd $0x39,<diag1=int6464#2,<diag1=int6464#2
+# asm 2: pshufd $0x39,<diag1=%xmm1,<diag1=%xmm1
+pshufd $0x39,%xmm1,%xmm1
+
+# qhasm: diag2 <<<= 96
+# asm 1: pshufd $0x39,<diag2=int6464#3,<diag2=int6464#3
+# asm 2: pshufd $0x39,<diag2=%xmm2,<diag2=%xmm2
+pshufd $0x39,%xmm2,%xmm2
+
+# qhasm: diag3 <<<= 96
+# asm 1: pshufd $0x39,<diag3=int6464#4,<diag3=int6464#4
+# asm 2: pshufd $0x39,<diag3=%xmm3,<diag3=%xmm3
+pshufd $0x39,%xmm3,%xmm3
+
+# qhasm: (uint32) in10 ^= *(uint32 *) (m + 40)
+# asm 1: xorl 40(<m=int64#2),<in10=int64#4d
+# asm 2: xorl 40(<m=%rsi),<in10=%ecx
+xorl 40(%rsi),%ecx
+
+# qhasm: (uint32) in6 ^= *(uint32 *) (m + 24)
+# asm 1: xorl 24(<m=int64#2),<in6=int64#5d
+# asm 2: xorl 24(<m=%rsi),<in6=%r8d
+xorl 24(%rsi),%r8d
+
+# qhasm: (uint32) in2 ^= *(uint32 *) (m + 8)
+# asm 1: xorl 8(<m=int64#2),<in2=int64#6d
+# asm 2: xorl 8(<m=%rsi),<in2=%r9d
+xorl 8(%rsi),%r9d
+
+# qhasm: (uint32) in14 ^= *(uint32 *) (m + 56)
+# asm 1: xorl 56(<m=int64#2),<in14=int64#7d
+# asm 2: xorl 56(<m=%rsi),<in14=%eax
+xorl 56(%rsi),%eax
+
+# qhasm: *(uint32 *) (out + 40) = in10
+# asm 1: movl   <in10=int64#4d,40(<out=int64#1)
+# asm 2: movl   <in10=%ecx,40(<out=%rdi)
+movl   %ecx,40(%rdi)
+
+# qhasm: *(uint32 *) (out + 24) = in6
+# asm 1: movl   <in6=int64#5d,24(<out=int64#1)
+# asm 2: movl   <in6=%r8d,24(<out=%rdi)
+movl   %r8d,24(%rdi)
+
+# qhasm: *(uint32 *) (out + 8) = in2
+# asm 1: movl   <in2=int64#6d,8(<out=int64#1)
+# asm 2: movl   <in2=%r9d,8(<out=%rdi)
+movl   %r9d,8(%rdi)
+
+# qhasm: *(uint32 *) (out + 56) = in14
+# asm 1: movl   <in14=int64#7d,56(<out=int64#1)
+# asm 2: movl   <in14=%eax,56(<out=%rdi)
+movl   %eax,56(%rdi)
+
+# qhasm: in15 = diag0
+# asm 1: movd   <diag0=int6464#1,>in15=int64#4
+# asm 2: movd   <diag0=%xmm0,>in15=%rcx
+movd   %xmm0,%rcx
+
+# qhasm: in11 = diag1
+# asm 1: movd   <diag1=int6464#2,>in11=int64#5
+# asm 2: movd   <diag1=%xmm1,>in11=%r8
+movd   %xmm1,%r8
+
+# qhasm: in7 = diag2
+# asm 1: movd   <diag2=int6464#3,>in7=int64#6
+# asm 2: movd   <diag2=%xmm2,>in7=%r9
+movd   %xmm2,%r9
+
+# qhasm: in3 = diag3
+# asm 1: movd   <diag3=int6464#4,>in3=int64#7
+# asm 2: movd   <diag3=%xmm3,>in3=%rax
+movd   %xmm3,%rax
+
+# qhasm: (uint32) in15 ^= *(uint32 *) (m + 60)
+# asm 1: xorl 60(<m=int64#2),<in15=int64#4d
+# asm 2: xorl 60(<m=%rsi),<in15=%ecx
+xorl 60(%rsi),%ecx
+
+# qhasm: (uint32) in11 ^= *(uint32 *) (m + 44)
+# asm 1: xorl 44(<m=int64#2),<in11=int64#5d
+# asm 2: xorl 44(<m=%rsi),<in11=%r8d
+xorl 44(%rsi),%r8d
+
+# qhasm: (uint32) in7 ^= *(uint32 *) (m + 28)
+# asm 1: xorl 28(<m=int64#2),<in7=int64#6d
+# asm 2: xorl 28(<m=%rsi),<in7=%r9d
+xorl 28(%rsi),%r9d
+
+# qhasm: (uint32) in3 ^= *(uint32 *) (m + 12)
+# asm 1: xorl 12(<m=int64#2),<in3=int64#7d
+# asm 2: xorl 12(<m=%rsi),<in3=%eax
+xorl 12(%rsi),%eax
+
+# qhasm: *(uint32 *) (out + 60) = in15
+# asm 1: movl   <in15=int64#4d,60(<out=int64#1)
+# asm 2: movl   <in15=%ecx,60(<out=%rdi)
+movl   %ecx,60(%rdi)
+
+# qhasm: *(uint32 *) (out + 44) = in11
+# asm 1: movl   <in11=int64#5d,44(<out=int64#1)
+# asm 2: movl   <in11=%r8d,44(<out=%rdi)
+movl   %r8d,44(%rdi)
+
+# qhasm: *(uint32 *) (out + 28) = in7
+# asm 1: movl   <in7=int64#6d,28(<out=int64#1)
+# asm 2: movl   <in7=%r9d,28(<out=%rdi)
+movl   %r9d,28(%rdi)
+
+# qhasm: *(uint32 *) (out + 12) = in3
+# asm 1: movl   <in3=int64#7d,12(<out=int64#1)
+# asm 2: movl   <in3=%eax,12(<out=%rdi)
+movl   %eax,12(%rdi)
+
+# qhasm:   bytes = bytes_backup
+# asm 1: movq <bytes_backup=stack64#8,>bytes=int64#6
+# asm 2: movq <bytes_backup=408(%rsp),>bytes=%r9
+movq 408(%rsp),%r9
+
+# qhasm:   in8 = ((uint32 *)&x2)[0]
+# asm 1: movl <x2=stack128#2,>in8=int64#4d
+# asm 2: movl <x2=16(%rsp),>in8=%ecx
+movl 16(%rsp),%ecx
+
+# qhasm:   in9 = ((uint32 *)&x3)[1]
+# asm 1: movl 4+<x3=stack128#3,>in9=int64#5d
+# asm 2: movl 4+<x3=32(%rsp),>in9=%r8d
+movl 4+32(%rsp),%r8d
+
+# qhasm:   in8 += 1
+# asm 1: add  $1,<in8=int64#4
+# asm 2: add  $1,<in8=%rcx
+add  $1,%rcx
+
+# qhasm:   in9 <<= 32
+# asm 1: shl  $32,<in9=int64#5
+# asm 2: shl  $32,<in9=%r8
+shl  $32,%r8
+
+# qhasm:   in8 += in9
+# asm 1: add  <in9=int64#5,<in8=int64#4
+# asm 2: add  <in9=%r8,<in8=%rcx
+add  %r8,%rcx
+
+# qhasm:   in9 = in8
+# asm 1: mov  <in8=int64#4,>in9=int64#5
+# asm 2: mov  <in8=%rcx,>in9=%r8
+mov  %rcx,%r8
+
+# qhasm:   (uint64) in9 >>= 32
+# asm 1: shr  $32,<in9=int64#5
+# asm 2: shr  $32,<in9=%r8
+shr  $32,%r8
+
+# qhasm:   ((uint32 *)&x2)[0] = in8
+# asm 1: movl <in8=int64#4d,>x2=stack128#2
+# asm 2: movl <in8=%ecx,>x2=16(%rsp)
+movl %ecx,16(%rsp)
+
+# qhasm:   ((uint32 *)&x3)[1] = in9
+# asm 1: movl <in9=int64#5d,4+<x3=stack128#3
+# asm 2: movl <in9=%r8d,4+<x3=32(%rsp)
+movl %r8d,4+32(%rsp)
+
+# qhasm:                          unsigned>? unsigned<? bytes - 64
+# asm 1: cmp  $64,<bytes=int64#6
+# asm 2: cmp  $64,<bytes=%r9
+cmp  $64,%r9
+# comment:fp stack unchanged by jump
+
+# qhasm:   goto bytesatleast65 if unsigned>
+ja ._bytesatleast65
+# comment:fp stack unchanged by jump
+
+# qhasm:     goto bytesatleast64 if !unsigned<
+jae ._bytesatleast64
+
+# qhasm:       m = out
+# asm 1: mov  <out=int64#1,>m=int64#2
+# asm 2: mov  <out=%rdi,>m=%rsi
+mov  %rdi,%rsi
+
+# qhasm:       out = ctarget
+# asm 1: mov  <ctarget=int64#3,>out=int64#1
+# asm 2: mov  <ctarget=%rdx,>out=%rdi
+mov  %rdx,%rdi
+
+# qhasm:       i = bytes
+# asm 1: mov  <bytes=int64#6,>i=int64#4
+# asm 2: mov  <bytes=%r9,>i=%rcx
+mov  %r9,%rcx
+
+# qhasm:       while (i) { *out++ = *m++; --i }
+rep movsb
+# comment:fp stack unchanged by fallthrough
+
+# qhasm:     bytesatleast64:
+._bytesatleast64:
+# comment:fp stack unchanged by fallthrough
+
+# qhasm:     done:
+._done:
+
+# qhasm:     r11_caller = r11_stack
+# asm 1: movq <r11_stack=stack64#1,>r11_caller=int64#9
+# asm 2: movq <r11_stack=352(%rsp),>r11_caller=%r11
+movq 352(%rsp),%r11
+
+# qhasm:     r12_caller = r12_stack
+# asm 1: movq <r12_stack=stack64#2,>r12_caller=int64#10
+# asm 2: movq <r12_stack=360(%rsp),>r12_caller=%r12
+movq 360(%rsp),%r12
+
+# qhasm:     r13_caller = r13_stack
+# asm 1: movq <r13_stack=stack64#3,>r13_caller=int64#11
+# asm 2: movq <r13_stack=368(%rsp),>r13_caller=%r13
+movq 368(%rsp),%r13
+
+# qhasm:     r14_caller = r14_stack
+# asm 1: movq <r14_stack=stack64#4,>r14_caller=int64#12
+# asm 2: movq <r14_stack=376(%rsp),>r14_caller=%r14
+movq 376(%rsp),%r14
+
+# qhasm:     r15_caller = r15_stack
+# asm 1: movq <r15_stack=stack64#5,>r15_caller=int64#13
+# asm 2: movq <r15_stack=384(%rsp),>r15_caller=%r15
+movq 384(%rsp),%r15
+
+# qhasm:     rbx_caller = rbx_stack
+# asm 1: movq <rbx_stack=stack64#6,>rbx_caller=int64#14
+# asm 2: movq <rbx_stack=392(%rsp),>rbx_caller=%rbx
+movq 392(%rsp),%rbx
+
+# qhasm:     rbp_caller = rbp_stack
+# asm 1: movq <rbp_stack=stack64#7,>rbp_caller=int64#15
+# asm 2: movq <rbp_stack=400(%rsp),>rbp_caller=%rbp
+movq 400(%rsp),%rbp
+
+# qhasm:     leave
+add %r11,%rsp
+xor %rax,%rax
+xor %rdx,%rdx
+ret
+
+# qhasm:   bytesatleast65:
+._bytesatleast65:
+
+# qhasm:   bytes -= 64
+# asm 1: sub  $64,<bytes=int64#6
+# asm 2: sub  $64,<bytes=%r9
+sub  $64,%r9
+
+# qhasm:   out += 64
+# asm 1: add  $64,<out=int64#1
+# asm 2: add  $64,<out=%rdi
+add  $64,%rdi
+
+# qhasm:   m += 64
+# asm 1: add  $64,<m=int64#2
+# asm 2: add  $64,<m=%rsi
+add  $64,%rsi
+# comment:fp stack unchanged by jump
+
+# qhasm: goto bytesbetween1and255
+jmp ._bytesbetween1and255
diff --git a/src/crypto/cipher/salsa2012/xmm/salsa2012_x86_xmm5.s b/src/crypto/cipher/salsa2012/xmm/salsa2012_x86_xmm5.s
new file mode 100644
index 0000000..c511b0d
--- /dev/null
+++ b/src/crypto/cipher/salsa2012/xmm/salsa2012_x86_xmm5.s
@@ -0,0 +1,5078 @@
+
+# qhasm: int32 a
+
+# qhasm: stack32 arg1
+
+# qhasm: stack32 arg2
+
+# qhasm: stack32 arg3
+
+# qhasm: stack32 arg4
+
+# qhasm: stack32 arg5
+
+# qhasm: stack32 arg6
+
+# qhasm: input arg1
+
+# qhasm: input arg2
+
+# qhasm: input arg3
+
+# qhasm: input arg4
+
+# qhasm: input arg5
+
+# qhasm: input arg6
+
+# qhasm: int32 eax
+
+# qhasm: int32 ebx
+
+# qhasm: int32 esi
+
+# qhasm: int32 edi
+
+# qhasm: int32 ebp
+
+# qhasm: caller eax
+
+# qhasm: caller ebx
+
+# qhasm: caller esi
+
+# qhasm: caller edi
+
+# qhasm: caller ebp
+
+# qhasm: int32 k
+
+# qhasm: int32 kbits
+
+# qhasm: int32 iv
+
+# qhasm: int32 i
+
+# qhasm: stack128 x0
+
+# qhasm: stack128 x1
+
+# qhasm: stack128 x2
+
+# qhasm: stack128 x3
+
+# qhasm: int32 m
+
+# qhasm: stack32 out_stack
+
+# qhasm: int32 out
+
+# qhasm: stack32 bytes_stack
+
+# qhasm: int32 bytes
+
+# qhasm: stack32 eax_stack
+
+# qhasm: stack32 ebx_stack
+
+# qhasm: stack32 esi_stack
+
+# qhasm: stack32 edi_stack
+
+# qhasm: stack32 ebp_stack
+
+# qhasm: int6464 diag0
+
+# qhasm: int6464 diag1
+
+# qhasm: int6464 diag2
+
+# qhasm: int6464 diag3
+
+# qhasm: int6464 a0
+
+# qhasm: int6464 a1
+
+# qhasm: int6464 a2
+
+# qhasm: int6464 a3
+
+# qhasm: int6464 a4
+
+# qhasm: int6464 a5
+
+# qhasm: int6464 a6
+
+# qhasm: int6464 a7
+
+# qhasm: int6464 b0
+
+# qhasm: int6464 b1
+
+# qhasm: int6464 b2
+
+# qhasm: int6464 b3
+
+# qhasm: int6464 b4
+
+# qhasm: int6464 b5
+
+# qhasm: int6464 b6
+
+# qhasm: int6464 b7
+
+# qhasm: int6464 z0
+
+# qhasm: int6464 z1
+
+# qhasm: int6464 z2
+
+# qhasm: int6464 z3
+
+# qhasm: int6464 z4
+
+# qhasm: int6464 z5
+
+# qhasm: int6464 z6
+
+# qhasm: int6464 z7
+
+# qhasm: int6464 z8
+
+# qhasm: int6464 z9
+
+# qhasm: int6464 z10
+
+# qhasm: int6464 z11
+
+# qhasm: int6464 z12
+
+# qhasm: int6464 z13
+
+# qhasm: int6464 z14
+
+# qhasm: int6464 z15
+
+# qhasm: stack128 z0_stack
+
+# qhasm: stack128 z1_stack
+
+# qhasm: stack128 z2_stack
+
+# qhasm: stack128 z3_stack
+
+# qhasm: stack128 z4_stack
+
+# qhasm: stack128 z5_stack
+
+# qhasm: stack128 z6_stack
+
+# qhasm: stack128 z7_stack
+
+# qhasm: stack128 z8_stack
+
+# qhasm: stack128 z9_stack
+
+# qhasm: stack128 z10_stack
+
+# qhasm: stack128 z11_stack
+
+# qhasm: stack128 z12_stack
+
+# qhasm: stack128 z13_stack
+
+# qhasm: stack128 z14_stack
+
+# qhasm: stack128 z15_stack
+
+# qhasm: stack128 orig0
+
+# qhasm: stack128 orig1
+
+# qhasm: stack128 orig2
+
+# qhasm: stack128 orig3
+
+# qhasm: stack128 orig4
+
+# qhasm: stack128 orig5
+
+# qhasm: stack128 orig6
+
+# qhasm: stack128 orig7
+
+# qhasm: stack128 orig8
+
+# qhasm: stack128 orig9
+
+# qhasm: stack128 orig10
+
+# qhasm: stack128 orig11
+
+# qhasm: stack128 orig12
+
+# qhasm: stack128 orig13
+
+# qhasm: stack128 orig14
+
+# qhasm: stack128 orig15
+
+# qhasm: int6464 p
+
+# qhasm: int6464 q
+
+# qhasm: int6464 r
+
+# qhasm: int6464 s
+
+# qhasm: int6464 t
+
+# qhasm: int6464 u
+
+# qhasm: int6464 v
+
+# qhasm: int6464 w
+
+# qhasm: int6464 mp
+
+# qhasm: int6464 mq
+
+# qhasm: int6464 mr
+
+# qhasm: int6464 ms
+
+# qhasm: int6464 mt
+
+# qhasm: int6464 mu
+
+# qhasm: int6464 mv
+
+# qhasm: int6464 mw
+
+# qhasm: int32 in0
+
+# qhasm: int32 in1
+
+# qhasm: int32 in2
+
+# qhasm: int32 in3
+
+# qhasm: int32 in4
+
+# qhasm: int32 in5
+
+# qhasm: int32 in6
+
+# qhasm: int32 in7
+
+# qhasm: int32 in8
+
+# qhasm: int32 in9
+
+# qhasm: int32 in10
+
+# qhasm: int32 in11
+
+# qhasm: int32 in12
+
+# qhasm: int32 in13
+
+# qhasm: int32 in14
+
+# qhasm: int32 in15
+
+# qhasm: stack512 tmp
+
+# qhasm: stack32 ctarget
+
+# qhasm: enter crypto_stream_salsa2012_x86_xmm5
+.text
+.p2align 5
+.globl _crypto_stream_salsa2012_x86_xmm5
+.globl crypto_stream_salsa2012_x86_xmm5
+_crypto_stream_salsa2012_x86_xmm5:
+crypto_stream_salsa2012_x86_xmm5:
+mov %esp,%eax
+and $31,%eax
+add $704,%eax
+sub %eax,%esp
+
+# qhasm: eax_stack = eax
+# asm 1: movl <eax=int32#1,>eax_stack=stack32#1
+# asm 2: movl <eax=%eax,>eax_stack=0(%esp)
+movl %eax,0(%esp)
+
+# qhasm: ebx_stack = ebx
+# asm 1: movl <ebx=int32#4,>ebx_stack=stack32#2
+# asm 2: movl <ebx=%ebx,>ebx_stack=4(%esp)
+movl %ebx,4(%esp)
+
+# qhasm: esi_stack = esi
+# asm 1: movl <esi=int32#5,>esi_stack=stack32#3
+# asm 2: movl <esi=%esi,>esi_stack=8(%esp)
+movl %esi,8(%esp)
+
+# qhasm: edi_stack = edi
+# asm 1: movl <edi=int32#6,>edi_stack=stack32#4
+# asm 2: movl <edi=%edi,>edi_stack=12(%esp)
+movl %edi,12(%esp)
+
+# qhasm: ebp_stack = ebp
+# asm 1: movl <ebp=int32#7,>ebp_stack=stack32#5
+# asm 2: movl <ebp=%ebp,>ebp_stack=16(%esp)
+movl %ebp,16(%esp)
+
+# qhasm: bytes = arg2
+# asm 1: movl <arg2=stack32#-2,>bytes=int32#3
+# asm 2: movl <arg2=8(%esp,%eax),>bytes=%edx
+movl 8(%esp,%eax),%edx
+
+# qhasm: out = arg1
+# asm 1: movl <arg1=stack32#-1,>out=int32#6
+# asm 2: movl <arg1=4(%esp,%eax),>out=%edi
+movl 4(%esp,%eax),%edi
+
+# qhasm: m = out
+# asm 1: mov  <out=int32#6,>m=int32#5
+# asm 2: mov  <out=%edi,>m=%esi
+mov  %edi,%esi
+
+# qhasm: iv = arg4
+# asm 1: movl <arg4=stack32#-4,>iv=int32#4
+# asm 2: movl <arg4=16(%esp,%eax),>iv=%ebx
+movl 16(%esp,%eax),%ebx
+
+# qhasm: k = arg5
+# asm 1: movl <arg5=stack32#-5,>k=int32#7
+# asm 2: movl <arg5=20(%esp,%eax),>k=%ebp
+movl 20(%esp,%eax),%ebp
+
+# qhasm:               unsigned>? bytes - 0
+# asm 1: cmp  $0,<bytes=int32#3
+# asm 2: cmp  $0,<bytes=%edx
+cmp  $0,%edx
+# comment:fp stack unchanged by jump
+
+# qhasm: goto done if !unsigned>
+jbe ._done
+
+# qhasm: a = 0
+# asm 1: mov  $0,>a=int32#1
+# asm 2: mov  $0,>a=%eax
+mov  $0,%eax
+
+# qhasm: i = bytes
+# asm 1: mov  <bytes=int32#3,>i=int32#2
+# asm 2: mov  <bytes=%edx,>i=%ecx
+mov  %edx,%ecx
+
+# qhasm: while (i) { *out++ = a; --i }
+rep stosb
+
+# qhasm: out -= bytes
+# asm 1: subl <bytes=int32#3,<out=int32#6
+# asm 2: subl <bytes=%edx,<out=%edi
+subl %edx,%edi
+# comment:fp stack unchanged by jump
+
+# qhasm: goto start
+jmp ._start
+
+# qhasm: enter crypto_stream_salsa2012_x86_xmm5_xor
+.text
+.p2align 5
+.globl _crypto_stream_salsa2012_x86_xmm5_xor
+.globl crypto_stream_salsa2012_x86_xmm5_xor
+_crypto_stream_salsa2012_x86_xmm5_xor:
+crypto_stream_salsa2012_x86_xmm5_xor:
+mov %esp,%eax
+and $31,%eax
+add $704,%eax
+sub %eax,%esp
+
+# qhasm: eax_stack = eax
+# asm 1: movl <eax=int32#1,>eax_stack=stack32#1
+# asm 2: movl <eax=%eax,>eax_stack=0(%esp)
+movl %eax,0(%esp)
+
+# qhasm: ebx_stack = ebx
+# asm 1: movl <ebx=int32#4,>ebx_stack=stack32#2
+# asm 2: movl <ebx=%ebx,>ebx_stack=4(%esp)
+movl %ebx,4(%esp)
+
+# qhasm: esi_stack = esi
+# asm 1: movl <esi=int32#5,>esi_stack=stack32#3
+# asm 2: movl <esi=%esi,>esi_stack=8(%esp)
+movl %esi,8(%esp)
+
+# qhasm: edi_stack = edi
+# asm 1: movl <edi=int32#6,>edi_stack=stack32#4
+# asm 2: movl <edi=%edi,>edi_stack=12(%esp)
+movl %edi,12(%esp)
+
+# qhasm: ebp_stack = ebp
+# asm 1: movl <ebp=int32#7,>ebp_stack=stack32#5
+# asm 2: movl <ebp=%ebp,>ebp_stack=16(%esp)
+movl %ebp,16(%esp)
+
+# qhasm: out = arg1
+# asm 1: movl <arg1=stack32#-1,>out=int32#6
+# asm 2: movl <arg1=4(%esp,%eax),>out=%edi
+movl 4(%esp,%eax),%edi
+
+# qhasm: m = arg2
+# asm 1: movl <arg2=stack32#-2,>m=int32#5
+# asm 2: movl <arg2=8(%esp,%eax),>m=%esi
+movl 8(%esp,%eax),%esi
+
+# qhasm: bytes = arg3
+# asm 1: movl <arg3=stack32#-3,>bytes=int32#3
+# asm 2: movl <arg3=12(%esp,%eax),>bytes=%edx
+movl 12(%esp,%eax),%edx
+
+# qhasm: iv = arg5
+# asm 1: movl <arg5=stack32#-5,>iv=int32#4
+# asm 2: movl <arg5=20(%esp,%eax),>iv=%ebx
+movl 20(%esp,%eax),%ebx
+
+# qhasm: k = arg6
+# asm 1: movl <arg6=stack32#-6,>k=int32#7
+# asm 2: movl <arg6=24(%esp,%eax),>k=%ebp
+movl 24(%esp,%eax),%ebp
+
+# qhasm:               unsigned>? bytes - 0
+# asm 1: cmp  $0,<bytes=int32#3
+# asm 2: cmp  $0,<bytes=%edx
+cmp  $0,%edx
+# comment:fp stack unchanged by jump
+
+# qhasm: goto done if !unsigned>
+jbe ._done
+# comment:fp stack unchanged by fallthrough
+
+# qhasm: start:
+._start:
+
+# qhasm:   out_stack = out
+# asm 1: movl <out=int32#6,>out_stack=stack32#6
+# asm 2: movl <out=%edi,>out_stack=20(%esp)
+movl %edi,20(%esp)
+
+# qhasm:   bytes_stack = bytes
+# asm 1: movl <bytes=int32#3,>bytes_stack=stack32#7
+# asm 2: movl <bytes=%edx,>bytes_stack=24(%esp)
+movl %edx,24(%esp)
+
+# qhasm:   in4 = *(uint32 *) (k + 12)
+# asm 1: movl 12(<k=int32#7),>in4=int32#1
+# asm 2: movl 12(<k=%ebp),>in4=%eax
+movl 12(%ebp),%eax
+
+# qhasm:   in12 = *(uint32 *) (k + 20)
+# asm 1: movl 20(<k=int32#7),>in12=int32#2
+# asm 2: movl 20(<k=%ebp),>in12=%ecx
+movl 20(%ebp),%ecx
+
+# qhasm:   ((uint32 *)&x3)[0] = in4
+# asm 1: movl <in4=int32#1,>x3=stack128#1
+# asm 2: movl <in4=%eax,>x3=32(%esp)
+movl %eax,32(%esp)
+
+# qhasm:   ((uint32 *)&x1)[0] = in12
+# asm 1: movl <in12=int32#2,>x1=stack128#2
+# asm 2: movl <in12=%ecx,>x1=48(%esp)
+movl %ecx,48(%esp)
+
+# qhasm:   in0 = 1634760805
+# asm 1: mov  $1634760805,>in0=int32#1
+# asm 2: mov  $1634760805,>in0=%eax
+mov  $1634760805,%eax
+
+# qhasm:   in8 = 0
+# asm 1: mov  $0,>in8=int32#2
+# asm 2: mov  $0,>in8=%ecx
+mov  $0,%ecx
+
+# qhasm:   ((uint32 *)&x0)[0] = in0
+# asm 1: movl <in0=int32#1,>x0=stack128#3
+# asm 2: movl <in0=%eax,>x0=64(%esp)
+movl %eax,64(%esp)
+
+# qhasm:   ((uint32 *)&x2)[0] = in8
+# asm 1: movl <in8=int32#2,>x2=stack128#4
+# asm 2: movl <in8=%ecx,>x2=80(%esp)
+movl %ecx,80(%esp)
+
+# qhasm:   in6 = *(uint32 *) (iv + 0)
+# asm 1: movl 0(<iv=int32#4),>in6=int32#1
+# asm 2: movl 0(<iv=%ebx),>in6=%eax
+movl 0(%ebx),%eax
+
+# qhasm:   in7 = *(uint32 *) (iv + 4)
+# asm 1: movl 4(<iv=int32#4),>in7=int32#2
+# asm 2: movl 4(<iv=%ebx),>in7=%ecx
+movl 4(%ebx),%ecx
+
+# qhasm:   ((uint32 *)&x1)[2] = in6
+# asm 1: movl <in6=int32#1,8+<x1=stack128#2
+# asm 2: movl <in6=%eax,8+<x1=48(%esp)
+movl %eax,8+48(%esp)
+
+# qhasm:   ((uint32 *)&x2)[3] = in7
+# asm 1: movl <in7=int32#2,12+<x2=stack128#4
+# asm 2: movl <in7=%ecx,12+<x2=80(%esp)
+movl %ecx,12+80(%esp)
+
+# qhasm:   in9 = 0
+# asm 1: mov  $0,>in9=int32#1
+# asm 2: mov  $0,>in9=%eax
+mov  $0,%eax
+
+# qhasm:   in10 = 2036477234
+# asm 1: mov  $2036477234,>in10=int32#2
+# asm 2: mov  $2036477234,>in10=%ecx
+mov  $2036477234,%ecx
+
+# qhasm:   ((uint32 *)&x3)[1] = in9
+# asm 1: movl <in9=int32#1,4+<x3=stack128#1
+# asm 2: movl <in9=%eax,4+<x3=32(%esp)
+movl %eax,4+32(%esp)
+
+# qhasm:   ((uint32 *)&x0)[2] = in10
+# asm 1: movl <in10=int32#2,8+<x0=stack128#3
+# asm 2: movl <in10=%ecx,8+<x0=64(%esp)
+movl %ecx,8+64(%esp)
+
+# qhasm:   in1 = *(uint32 *) (k + 0)
+# asm 1: movl 0(<k=int32#7),>in1=int32#1
+# asm 2: movl 0(<k=%ebp),>in1=%eax
+movl 0(%ebp),%eax
+
+# qhasm:   in2 = *(uint32 *) (k + 4)
+# asm 1: movl 4(<k=int32#7),>in2=int32#2
+# asm 2: movl 4(<k=%ebp),>in2=%ecx
+movl 4(%ebp),%ecx
+
+# qhasm:   in3 = *(uint32 *) (k + 8)
+# asm 1: movl 8(<k=int32#7),>in3=int32#3
+# asm 2: movl 8(<k=%ebp),>in3=%edx
+movl 8(%ebp),%edx
+
+# qhasm:   in5 = 857760878
+# asm 1: mov  $857760878,>in5=int32#4
+# asm 2: mov  $857760878,>in5=%ebx
+mov  $857760878,%ebx
+
+# qhasm:   ((uint32 *)&x1)[1] = in1
+# asm 1: movl <in1=int32#1,4+<x1=stack128#2
+# asm 2: movl <in1=%eax,4+<x1=48(%esp)
+movl %eax,4+48(%esp)
+
+# qhasm:   ((uint32 *)&x2)[2] = in2
+# asm 1: movl <in2=int32#2,8+<x2=stack128#4
+# asm 2: movl <in2=%ecx,8+<x2=80(%esp)
+movl %ecx,8+80(%esp)
+
+# qhasm:   ((uint32 *)&x3)[3] = in3
+# asm 1: movl <in3=int32#3,12+<x3=stack128#1
+# asm 2: movl <in3=%edx,12+<x3=32(%esp)
+movl %edx,12+32(%esp)
+
+# qhasm:   ((uint32 *)&x0)[1] = in5
+# asm 1: movl <in5=int32#4,4+<x0=stack128#3
+# asm 2: movl <in5=%ebx,4+<x0=64(%esp)
+movl %ebx,4+64(%esp)
+
+# qhasm:   in11 = *(uint32 *) (k + 16)
+# asm 1: movl 16(<k=int32#7),>in11=int32#1
+# asm 2: movl 16(<k=%ebp),>in11=%eax
+movl 16(%ebp),%eax
+
+# qhasm:   in13 = *(uint32 *) (k + 24)
+# asm 1: movl 24(<k=int32#7),>in13=int32#2
+# asm 2: movl 24(<k=%ebp),>in13=%ecx
+movl 24(%ebp),%ecx
+
+# qhasm:   in14 = *(uint32 *) (k + 28)
+# asm 1: movl 28(<k=int32#7),>in14=int32#3
+# asm 2: movl 28(<k=%ebp),>in14=%edx
+movl 28(%ebp),%edx
+
+# qhasm:   in15 = 1797285236
+# asm 1: mov  $1797285236,>in15=int32#4
+# asm 2: mov  $1797285236,>in15=%ebx
+mov  $1797285236,%ebx
+
+# qhasm:   ((uint32 *)&x1)[3] = in11
+# asm 1: movl <in11=int32#1,12+<x1=stack128#2
+# asm 2: movl <in11=%eax,12+<x1=48(%esp)
+movl %eax,12+48(%esp)
+
+# qhasm:   ((uint32 *)&x2)[1] = in13
+# asm 1: movl <in13=int32#2,4+<x2=stack128#4
+# asm 2: movl <in13=%ecx,4+<x2=80(%esp)
+movl %ecx,4+80(%esp)
+
+# qhasm:   ((uint32 *)&x3)[2] = in14
+# asm 1: movl <in14=int32#3,8+<x3=stack128#1
+# asm 2: movl <in14=%edx,8+<x3=32(%esp)
+movl %edx,8+32(%esp)
+
+# qhasm:   ((uint32 *)&x0)[3] = in15
+# asm 1: movl <in15=int32#4,12+<x0=stack128#3
+# asm 2: movl <in15=%ebx,12+<x0=64(%esp)
+movl %ebx,12+64(%esp)
+
+# qhasm:   bytes = bytes_stack
+# asm 1: movl <bytes_stack=stack32#7,>bytes=int32#1
+# asm 2: movl <bytes_stack=24(%esp),>bytes=%eax
+movl 24(%esp),%eax
+
+# qhasm:                               unsigned<? bytes - 256
+# asm 1: cmp  $256,<bytes=int32#1
+# asm 2: cmp  $256,<bytes=%eax
+cmp  $256,%eax
+# comment:fp stack unchanged by jump
+
+# qhasm:   goto bytesbetween1and255 if unsigned<
+jb ._bytesbetween1and255
+
+# qhasm:   z0 = x0
+# asm 1: movdqa <x0=stack128#3,>z0=int6464#1
+# asm 2: movdqa <x0=64(%esp),>z0=%xmm0
+movdqa 64(%esp),%xmm0
+
+# qhasm:   z5 = z0[1,1,1,1]
+# asm 1: pshufd $0x55,<z0=int6464#1,>z5=int6464#2
+# asm 2: pshufd $0x55,<z0=%xmm0,>z5=%xmm1
+pshufd $0x55,%xmm0,%xmm1
+
+# qhasm:   z10 = z0[2,2,2,2]
+# asm 1: pshufd $0xaa,<z0=int6464#1,>z10=int6464#3
+# asm 2: pshufd $0xaa,<z0=%xmm0,>z10=%xmm2
+pshufd $0xaa,%xmm0,%xmm2
+
+# qhasm:   z15 = z0[3,3,3,3]
+# asm 1: pshufd $0xff,<z0=int6464#1,>z15=int6464#4
+# asm 2: pshufd $0xff,<z0=%xmm0,>z15=%xmm3
+pshufd $0xff,%xmm0,%xmm3
+
+# qhasm:   z0 = z0[0,0,0,0]
+# asm 1: pshufd $0x00,<z0=int6464#1,>z0=int6464#1
+# asm 2: pshufd $0x00,<z0=%xmm0,>z0=%xmm0
+pshufd $0x00,%xmm0,%xmm0
+
+# qhasm:   orig5 = z5
+# asm 1: movdqa <z5=int6464#2,>orig5=stack128#5
+# asm 2: movdqa <z5=%xmm1,>orig5=96(%esp)
+movdqa %xmm1,96(%esp)
+
+# qhasm:   orig10 = z10
+# asm 1: movdqa <z10=int6464#3,>orig10=stack128#6
+# asm 2: movdqa <z10=%xmm2,>orig10=112(%esp)
+movdqa %xmm2,112(%esp)
+
+# qhasm:   orig15 = z15
+# asm 1: movdqa <z15=int6464#4,>orig15=stack128#7
+# asm 2: movdqa <z15=%xmm3,>orig15=128(%esp)
+movdqa %xmm3,128(%esp)
+
+# qhasm:   orig0 = z0
+# asm 1: movdqa <z0=int6464#1,>orig0=stack128#8
+# asm 2: movdqa <z0=%xmm0,>orig0=144(%esp)
+movdqa %xmm0,144(%esp)
+
+# qhasm:   z1 = x1
+# asm 1: movdqa <x1=stack128#2,>z1=int6464#1
+# asm 2: movdqa <x1=48(%esp),>z1=%xmm0
+movdqa 48(%esp),%xmm0
+
+# qhasm:   z6 = z1[2,2,2,2]
+# asm 1: pshufd $0xaa,<z1=int6464#1,>z6=int6464#2
+# asm 2: pshufd $0xaa,<z1=%xmm0,>z6=%xmm1
+pshufd $0xaa,%xmm0,%xmm1
+
+# qhasm:   z11 = z1[3,3,3,3]
+# asm 1: pshufd $0xff,<z1=int6464#1,>z11=int6464#3
+# asm 2: pshufd $0xff,<z1=%xmm0,>z11=%xmm2
+pshufd $0xff,%xmm0,%xmm2
+
+# qhasm:   z12 = z1[0,0,0,0]
+# asm 1: pshufd $0x00,<z1=int6464#1,>z12=int6464#4
+# asm 2: pshufd $0x00,<z1=%xmm0,>z12=%xmm3
+pshufd $0x00,%xmm0,%xmm3
+
+# qhasm:   z1 = z1[1,1,1,1]
+# asm 1: pshufd $0x55,<z1=int6464#1,>z1=int6464#1
+# asm 2: pshufd $0x55,<z1=%xmm0,>z1=%xmm0
+pshufd $0x55,%xmm0,%xmm0
+
+# qhasm:   orig6 = z6
+# asm 1: movdqa <z6=int6464#2,>orig6=stack128#9
+# asm 2: movdqa <z6=%xmm1,>orig6=160(%esp)
+movdqa %xmm1,160(%esp)
+
+# qhasm:   orig11 = z11
+# asm 1: movdqa <z11=int6464#3,>orig11=stack128#10
+# asm 2: movdqa <z11=%xmm2,>orig11=176(%esp)
+movdqa %xmm2,176(%esp)
+
+# qhasm:   orig12 = z12
+# asm 1: movdqa <z12=int6464#4,>orig12=stack128#11
+# asm 2: movdqa <z12=%xmm3,>orig12=192(%esp)
+movdqa %xmm3,192(%esp)
+
+# qhasm:   orig1 = z1
+# asm 1: movdqa <z1=int6464#1,>orig1=stack128#12
+# asm 2: movdqa <z1=%xmm0,>orig1=208(%esp)
+movdqa %xmm0,208(%esp)
+
+# qhasm:   z2 = x2
+# asm 1: movdqa <x2=stack128#4,>z2=int6464#1
+# asm 2: movdqa <x2=80(%esp),>z2=%xmm0
+movdqa 80(%esp),%xmm0
+
+# qhasm:   z7 = z2[3,3,3,3]
+# asm 1: pshufd $0xff,<z2=int6464#1,>z7=int6464#2
+# asm 2: pshufd $0xff,<z2=%xmm0,>z7=%xmm1
+pshufd $0xff,%xmm0,%xmm1
+
+# qhasm:   z13 = z2[1,1,1,1]
+# asm 1: pshufd $0x55,<z2=int6464#1,>z13=int6464#3
+# asm 2: pshufd $0x55,<z2=%xmm0,>z13=%xmm2
+pshufd $0x55,%xmm0,%xmm2
+
+# qhasm:   z2 = z2[2,2,2,2]
+# asm 1: pshufd $0xaa,<z2=int6464#1,>z2=int6464#1
+# asm 2: pshufd $0xaa,<z2=%xmm0,>z2=%xmm0
+pshufd $0xaa,%xmm0,%xmm0
+
+# qhasm:   orig7 = z7
+# asm 1: movdqa <z7=int6464#2,>orig7=stack128#13
+# asm 2: movdqa <z7=%xmm1,>orig7=224(%esp)
+movdqa %xmm1,224(%esp)
+
+# qhasm:   orig13 = z13
+# asm 1: movdqa <z13=int6464#3,>orig13=stack128#14
+# asm 2: movdqa <z13=%xmm2,>orig13=240(%esp)
+movdqa %xmm2,240(%esp)
+
+# qhasm:   orig2 = z2
+# asm 1: movdqa <z2=int6464#1,>orig2=stack128#15
+# asm 2: movdqa <z2=%xmm0,>orig2=256(%esp)
+movdqa %xmm0,256(%esp)
+
+# qhasm:   z3 = x3
+# asm 1: movdqa <x3=stack128#1,>z3=int6464#1
+# asm 2: movdqa <x3=32(%esp),>z3=%xmm0
+movdqa 32(%esp),%xmm0
+
+# qhasm:   z4 = z3[0,0,0,0]
+# asm 1: pshufd $0x00,<z3=int6464#1,>z4=int6464#2
+# asm 2: pshufd $0x00,<z3=%xmm0,>z4=%xmm1
+pshufd $0x00,%xmm0,%xmm1
+
+# qhasm:   z14 = z3[2,2,2,2]
+# asm 1: pshufd $0xaa,<z3=int6464#1,>z14=int6464#3
+# asm 2: pshufd $0xaa,<z3=%xmm0,>z14=%xmm2
+pshufd $0xaa,%xmm0,%xmm2
+
+# qhasm:   z3 = z3[3,3,3,3]
+# asm 1: pshufd $0xff,<z3=int6464#1,>z3=int6464#1
+# asm 2: pshufd $0xff,<z3=%xmm0,>z3=%xmm0
+pshufd $0xff,%xmm0,%xmm0
+
+# qhasm:   orig4 = z4
+# asm 1: movdqa <z4=int6464#2,>orig4=stack128#16
+# asm 2: movdqa <z4=%xmm1,>orig4=272(%esp)
+movdqa %xmm1,272(%esp)
+
+# qhasm:   orig14 = z14
+# asm 1: movdqa <z14=int6464#3,>orig14=stack128#17
+# asm 2: movdqa <z14=%xmm2,>orig14=288(%esp)
+movdqa %xmm2,288(%esp)
+
+# qhasm:   orig3 = z3
+# asm 1: movdqa <z3=int6464#1,>orig3=stack128#18
+# asm 2: movdqa <z3=%xmm0,>orig3=304(%esp)
+movdqa %xmm0,304(%esp)
+
+# qhasm: bytesatleast256:
+._bytesatleast256:
+
+# qhasm:   in8 = ((uint32 *)&x2)[0]
+# asm 1: movl <x2=stack128#4,>in8=int32#2
+# asm 2: movl <x2=80(%esp),>in8=%ecx
+movl 80(%esp),%ecx
+
+# qhasm:   in9 = ((uint32 *)&x3)[1]
+# asm 1: movl 4+<x3=stack128#1,>in9=int32#3
+# asm 2: movl 4+<x3=32(%esp),>in9=%edx
+movl 4+32(%esp),%edx
+
+# qhasm:   ((uint32 *) &orig8)[0] = in8
+# asm 1: movl <in8=int32#2,>orig8=stack128#19
+# asm 2: movl <in8=%ecx,>orig8=320(%esp)
+movl %ecx,320(%esp)
+
+# qhasm:   ((uint32 *) &orig9)[0] = in9
+# asm 1: movl <in9=int32#3,>orig9=stack128#20
+# asm 2: movl <in9=%edx,>orig9=336(%esp)
+movl %edx,336(%esp)
+
+# qhasm:   carry? in8 += 1
+# asm 1: add  $1,<in8=int32#2
+# asm 2: add  $1,<in8=%ecx
+add  $1,%ecx
+
+# qhasm:   in9 += 0 + carry
+# asm 1: adc $0,<in9=int32#3
+# asm 2: adc $0,<in9=%edx
+adc $0,%edx
+
+# qhasm:   ((uint32 *) &orig8)[1] = in8
+# asm 1: movl <in8=int32#2,4+<orig8=stack128#19
+# asm 2: movl <in8=%ecx,4+<orig8=320(%esp)
+movl %ecx,4+320(%esp)
+
+# qhasm:   ((uint32 *) &orig9)[1] = in9
+# asm 1: movl <in9=int32#3,4+<orig9=stack128#20
+# asm 2: movl <in9=%edx,4+<orig9=336(%esp)
+movl %edx,4+336(%esp)
+
+# qhasm:   carry? in8 += 1
+# asm 1: add  $1,<in8=int32#2
+# asm 2: add  $1,<in8=%ecx
+add  $1,%ecx
+
+# qhasm:   in9 += 0 + carry
+# asm 1: adc $0,<in9=int32#3
+# asm 2: adc $0,<in9=%edx
+adc $0,%edx
+
+# qhasm:   ((uint32 *) &orig8)[2] = in8
+# asm 1: movl <in8=int32#2,8+<orig8=stack128#19
+# asm 2: movl <in8=%ecx,8+<orig8=320(%esp)
+movl %ecx,8+320(%esp)
+
+# qhasm:   ((uint32 *) &orig9)[2] = in9
+# asm 1: movl <in9=int32#3,8+<orig9=stack128#20
+# asm 2: movl <in9=%edx,8+<orig9=336(%esp)
+movl %edx,8+336(%esp)
+
+# qhasm:   carry? in8 += 1
+# asm 1: add  $1,<in8=int32#2
+# asm 2: add  $1,<in8=%ecx
+add  $1,%ecx
+
+# qhasm:   in9 += 0 + carry
+# asm 1: adc $0,<in9=int32#3
+# asm 2: adc $0,<in9=%edx
+adc $0,%edx
+
+# qhasm:   ((uint32 *) &orig8)[3] = in8
+# asm 1: movl <in8=int32#2,12+<orig8=stack128#19
+# asm 2: movl <in8=%ecx,12+<orig8=320(%esp)
+movl %ecx,12+320(%esp)
+
+# qhasm:   ((uint32 *) &orig9)[3] = in9
+# asm 1: movl <in9=int32#3,12+<orig9=stack128#20
+# asm 2: movl <in9=%edx,12+<orig9=336(%esp)
+movl %edx,12+336(%esp)
+
+# qhasm:   carry? in8 += 1
+# asm 1: add  $1,<in8=int32#2
+# asm 2: add  $1,<in8=%ecx
+add  $1,%ecx
+
+# qhasm:   in9 += 0 + carry
+# asm 1: adc $0,<in9=int32#3
+# asm 2: adc $0,<in9=%edx
+adc $0,%edx
+
+# qhasm:   ((uint32 *)&x2)[0] = in8
+# asm 1: movl <in8=int32#2,>x2=stack128#4
+# asm 2: movl <in8=%ecx,>x2=80(%esp)
+movl %ecx,80(%esp)
+
+# qhasm:   ((uint32 *)&x3)[1] = in9
+# asm 1: movl <in9=int32#3,4+<x3=stack128#1
+# asm 2: movl <in9=%edx,4+<x3=32(%esp)
+movl %edx,4+32(%esp)
+
+# qhasm:   bytes_stack = bytes
+# asm 1: movl <bytes=int32#1,>bytes_stack=stack32#7
+# asm 2: movl <bytes=%eax,>bytes_stack=24(%esp)
+movl %eax,24(%esp)
+
+# qhasm: i = 12
+# asm 1: mov  $12,>i=int32#1
+# asm 2: mov  $12,>i=%eax
+mov  $12,%eax
+
+# qhasm:   z5 = orig5
+# asm 1: movdqa <orig5=stack128#5,>z5=int6464#1
+# asm 2: movdqa <orig5=96(%esp),>z5=%xmm0
+movdqa 96(%esp),%xmm0
+
+# qhasm:   z10 = orig10
+# asm 1: movdqa <orig10=stack128#6,>z10=int6464#2
+# asm 2: movdqa <orig10=112(%esp),>z10=%xmm1
+movdqa 112(%esp),%xmm1
+
+# qhasm:   z15 = orig15
+# asm 1: movdqa <orig15=stack128#7,>z15=int6464#3
+# asm 2: movdqa <orig15=128(%esp),>z15=%xmm2
+movdqa 128(%esp),%xmm2
+
+# qhasm:   z14 = orig14
+# asm 1: movdqa <orig14=stack128#17,>z14=int6464#4
+# asm 2: movdqa <orig14=288(%esp),>z14=%xmm3
+movdqa 288(%esp),%xmm3
+
+# qhasm:   z3 = orig3
+# asm 1: movdqa <orig3=stack128#18,>z3=int6464#5
+# asm 2: movdqa <orig3=304(%esp),>z3=%xmm4
+movdqa 304(%esp),%xmm4
+
+# qhasm:   z6 = orig6
+# asm 1: movdqa <orig6=stack128#9,>z6=int6464#6
+# asm 2: movdqa <orig6=160(%esp),>z6=%xmm5
+movdqa 160(%esp),%xmm5
+
+# qhasm:   z11 = orig11
+# asm 1: movdqa <orig11=stack128#10,>z11=int6464#7
+# asm 2: movdqa <orig11=176(%esp),>z11=%xmm6
+movdqa 176(%esp),%xmm6
+
+# qhasm:   z1 = orig1
+# asm 1: movdqa <orig1=stack128#12,>z1=int6464#8
+# asm 2: movdqa <orig1=208(%esp),>z1=%xmm7
+movdqa 208(%esp),%xmm7
+
+# qhasm:   z5_stack = z5
+# asm 1: movdqa <z5=int6464#1,>z5_stack=stack128#21
+# asm 2: movdqa <z5=%xmm0,>z5_stack=352(%esp)
+movdqa %xmm0,352(%esp)
+
+# qhasm:   z10_stack = z10
+# asm 1: movdqa <z10=int6464#2,>z10_stack=stack128#22
+# asm 2: movdqa <z10=%xmm1,>z10_stack=368(%esp)
+movdqa %xmm1,368(%esp)
+
+# qhasm:   z15_stack = z15
+# asm 1: movdqa <z15=int6464#3,>z15_stack=stack128#23
+# asm 2: movdqa <z15=%xmm2,>z15_stack=384(%esp)
+movdqa %xmm2,384(%esp)
+
+# qhasm:   z14_stack = z14
+# asm 1: movdqa <z14=int6464#4,>z14_stack=stack128#24
+# asm 2: movdqa <z14=%xmm3,>z14_stack=400(%esp)
+movdqa %xmm3,400(%esp)
+
+# qhasm:   z3_stack = z3
+# asm 1: movdqa <z3=int6464#5,>z3_stack=stack128#25
+# asm 2: movdqa <z3=%xmm4,>z3_stack=416(%esp)
+movdqa %xmm4,416(%esp)
+
+# qhasm:   z6_stack = z6
+# asm 1: movdqa <z6=int6464#6,>z6_stack=stack128#26
+# asm 2: movdqa <z6=%xmm5,>z6_stack=432(%esp)
+movdqa %xmm5,432(%esp)
+
+# qhasm:   z11_stack = z11
+# asm 1: movdqa <z11=int6464#7,>z11_stack=stack128#27
+# asm 2: movdqa <z11=%xmm6,>z11_stack=448(%esp)
+movdqa %xmm6,448(%esp)
+
+# qhasm:   z1_stack = z1
+# asm 1: movdqa <z1=int6464#8,>z1_stack=stack128#28
+# asm 2: movdqa <z1=%xmm7,>z1_stack=464(%esp)
+movdqa %xmm7,464(%esp)
+
+# qhasm:   z7 = orig7
+# asm 1: movdqa <orig7=stack128#13,>z7=int6464#5
+# asm 2: movdqa <orig7=224(%esp),>z7=%xmm4
+movdqa 224(%esp),%xmm4
+
+# qhasm:   z13 = orig13
+# asm 1: movdqa <orig13=stack128#14,>z13=int6464#6
+# asm 2: movdqa <orig13=240(%esp),>z13=%xmm5
+movdqa 240(%esp),%xmm5
+
+# qhasm:   z2 = orig2
+# asm 1: movdqa <orig2=stack128#15,>z2=int6464#7
+# asm 2: movdqa <orig2=256(%esp),>z2=%xmm6
+movdqa 256(%esp),%xmm6
+
+# qhasm:   z9 = orig9
+# asm 1: movdqa <orig9=stack128#20,>z9=int6464#8
+# asm 2: movdqa <orig9=336(%esp),>z9=%xmm7
+movdqa 336(%esp),%xmm7
+
+# qhasm:                   p = orig0
+# asm 1: movdqa <orig0=stack128#8,>p=int6464#1
+# asm 2: movdqa <orig0=144(%esp),>p=%xmm0
+movdqa 144(%esp),%xmm0
+
+# qhasm:                   t = orig12
+# asm 1: movdqa <orig12=stack128#11,>t=int6464#3
+# asm 2: movdqa <orig12=192(%esp),>t=%xmm2
+movdqa 192(%esp),%xmm2
+
+# qhasm:                   q = orig4
+# asm 1: movdqa <orig4=stack128#16,>q=int6464#4
+# asm 2: movdqa <orig4=272(%esp),>q=%xmm3
+movdqa 272(%esp),%xmm3
+
+# qhasm:                   r = orig8
+# asm 1: movdqa <orig8=stack128#19,>r=int6464#2
+# asm 2: movdqa <orig8=320(%esp),>r=%xmm1
+movdqa 320(%esp),%xmm1
+
+# qhasm:   z7_stack = z7
+# asm 1: movdqa <z7=int6464#5,>z7_stack=stack128#29
+# asm 2: movdqa <z7=%xmm4,>z7_stack=480(%esp)
+movdqa %xmm4,480(%esp)
+
+# qhasm:   z13_stack = z13
+# asm 1: movdqa <z13=int6464#6,>z13_stack=stack128#30
+# asm 2: movdqa <z13=%xmm5,>z13_stack=496(%esp)
+movdqa %xmm5,496(%esp)
+
+# qhasm:   z2_stack = z2
+# asm 1: movdqa <z2=int6464#7,>z2_stack=stack128#31
+# asm 2: movdqa <z2=%xmm6,>z2_stack=512(%esp)
+movdqa %xmm6,512(%esp)
+
+# qhasm:   z9_stack = z9
+# asm 1: movdqa <z9=int6464#8,>z9_stack=stack128#32
+# asm 2: movdqa <z9=%xmm7,>z9_stack=528(%esp)
+movdqa %xmm7,528(%esp)
+
+# qhasm:   z0_stack = p
+# asm 1: movdqa <p=int6464#1,>z0_stack=stack128#33
+# asm 2: movdqa <p=%xmm0,>z0_stack=544(%esp)
+movdqa %xmm0,544(%esp)
+
+# qhasm:   z12_stack = t
+# asm 1: movdqa <t=int6464#3,>z12_stack=stack128#34
+# asm 2: movdqa <t=%xmm2,>z12_stack=560(%esp)
+movdqa %xmm2,560(%esp)
+
+# qhasm:   z4_stack = q
+# asm 1: movdqa <q=int6464#4,>z4_stack=stack128#35
+# asm 2: movdqa <q=%xmm3,>z4_stack=576(%esp)
+movdqa %xmm3,576(%esp)
+
+# qhasm:   z8_stack = r
+# asm 1: movdqa <r=int6464#2,>z8_stack=stack128#36
+# asm 2: movdqa <r=%xmm1,>z8_stack=592(%esp)
+movdqa %xmm1,592(%esp)
+
+# qhasm: mainloop1:
+._mainloop1:
+
+# qhasm:                   assign xmm0 to p
+
+# qhasm:                   assign xmm1 to r
+
+# qhasm:                   assign xmm2 to t
+
+# qhasm:                   assign xmm3 to q
+
+# qhasm:                   s = t
+# asm 1: movdqa <t=int6464#3,>s=int6464#7
+# asm 2: movdqa <t=%xmm2,>s=%xmm6
+movdqa %xmm2,%xmm6
+
+# qhasm: uint32323232      t += p
+# asm 1: paddd <p=int6464#1,<t=int6464#3
+# asm 2: paddd <p=%xmm0,<t=%xmm2
+paddd %xmm0,%xmm2
+
+# qhasm:                   u = t
+# asm 1: movdqa <t=int6464#3,>u=int6464#5
+# asm 2: movdqa <t=%xmm2,>u=%xmm4
+movdqa %xmm2,%xmm4
+
+# qhasm: uint32323232      t >>= 25
+# asm 1: psrld $25,<t=int6464#3
+# asm 2: psrld $25,<t=%xmm2
+psrld $25,%xmm2
+
+# qhasm:                   q ^= t
+# asm 1: pxor  <t=int6464#3,<q=int6464#4
+# asm 2: pxor  <t=%xmm2,<q=%xmm3
+pxor  %xmm2,%xmm3
+
+# qhasm: uint32323232      u <<= 7
+# asm 1: pslld $7,<u=int6464#5
+# asm 2: pslld $7,<u=%xmm4
+pslld $7,%xmm4
+
+# qhasm:                   q ^= u
+# asm 1: pxor  <u=int6464#5,<q=int6464#4
+# asm 2: pxor  <u=%xmm4,<q=%xmm3
+pxor  %xmm4,%xmm3
+
+# qhasm:                   z4_stack = q
+# asm 1: movdqa <q=int6464#4,>z4_stack=stack128#33
+# asm 2: movdqa <q=%xmm3,>z4_stack=544(%esp)
+movdqa %xmm3,544(%esp)
+
+# qhasm:                   t = p
+# asm 1: movdqa <p=int6464#1,>t=int6464#3
+# asm 2: movdqa <p=%xmm0,>t=%xmm2
+movdqa %xmm0,%xmm2
+
+# qhasm: uint32323232      t += q
+# asm 1: paddd <q=int6464#4,<t=int6464#3
+# asm 2: paddd <q=%xmm3,<t=%xmm2
+paddd %xmm3,%xmm2
+
+# qhasm:                   u = t
+# asm 1: movdqa <t=int6464#3,>u=int6464#5
+# asm 2: movdqa <t=%xmm2,>u=%xmm4
+movdqa %xmm2,%xmm4
+
+# qhasm: uint32323232      t >>= 23
+# asm 1: psrld $23,<t=int6464#3
+# asm 2: psrld $23,<t=%xmm2
+psrld $23,%xmm2
+
+# qhasm:                   r ^= t
+# asm 1: pxor  <t=int6464#3,<r=int6464#2
+# asm 2: pxor  <t=%xmm2,<r=%xmm1
+pxor  %xmm2,%xmm1
+
+# qhasm: uint32323232      u <<= 9
+# asm 1: pslld $9,<u=int6464#5
+# asm 2: pslld $9,<u=%xmm4
+pslld $9,%xmm4
+
+# qhasm:                   r ^= u
+# asm 1: pxor  <u=int6464#5,<r=int6464#2
+# asm 2: pxor  <u=%xmm4,<r=%xmm1
+pxor  %xmm4,%xmm1
+
+# qhasm:                   z8_stack = r
+# asm 1: movdqa <r=int6464#2,>z8_stack=stack128#34
+# asm 2: movdqa <r=%xmm1,>z8_stack=560(%esp)
+movdqa %xmm1,560(%esp)
+
+# qhasm: uint32323232      q += r
+# asm 1: paddd <r=int6464#2,<q=int6464#4
+# asm 2: paddd <r=%xmm1,<q=%xmm3
+paddd %xmm1,%xmm3
+
+# qhasm:                   u = q
+# asm 1: movdqa <q=int6464#4,>u=int6464#3
+# asm 2: movdqa <q=%xmm3,>u=%xmm2
+movdqa %xmm3,%xmm2
+
+# qhasm: uint32323232      q >>= 19
+# asm 1: psrld $19,<q=int6464#4
+# asm 2: psrld $19,<q=%xmm3
+psrld $19,%xmm3
+
+# qhasm:                   s ^= q
+# asm 1: pxor  <q=int6464#4,<s=int6464#7
+# asm 2: pxor  <q=%xmm3,<s=%xmm6
+pxor  %xmm3,%xmm6
+
+# qhasm: uint32323232      u <<= 13
+# asm 1: pslld $13,<u=int6464#3
+# asm 2: pslld $13,<u=%xmm2
+pslld $13,%xmm2
+
+# qhasm:                   s ^= u
+# asm 1: pxor  <u=int6464#3,<s=int6464#7
+# asm 2: pxor  <u=%xmm2,<s=%xmm6
+pxor  %xmm2,%xmm6
+
+# qhasm:                                   mt = z1_stack
+# asm 1: movdqa <z1_stack=stack128#28,>mt=int6464#3
+# asm 2: movdqa <z1_stack=464(%esp),>mt=%xmm2
+movdqa 464(%esp),%xmm2
+
+# qhasm:                                   mp = z5_stack
+# asm 1: movdqa <z5_stack=stack128#21,>mp=int6464#5
+# asm 2: movdqa <z5_stack=352(%esp),>mp=%xmm4
+movdqa 352(%esp),%xmm4
+
+# qhasm:                                   mq = z9_stack
+# asm 1: movdqa <z9_stack=stack128#32,>mq=int6464#4
+# asm 2: movdqa <z9_stack=528(%esp),>mq=%xmm3
+movdqa 528(%esp),%xmm3
+
+# qhasm:                                   mr = z13_stack
+# asm 1: movdqa <z13_stack=stack128#30,>mr=int6464#6
+# asm 2: movdqa <z13_stack=496(%esp),>mr=%xmm5
+movdqa 496(%esp),%xmm5
+
+# qhasm:                   z12_stack = s
+# asm 1: movdqa <s=int6464#7,>z12_stack=stack128#30
+# asm 2: movdqa <s=%xmm6,>z12_stack=496(%esp)
+movdqa %xmm6,496(%esp)
+
+# qhasm: uint32323232      r += s
+# asm 1: paddd <s=int6464#7,<r=int6464#2
+# asm 2: paddd <s=%xmm6,<r=%xmm1
+paddd %xmm6,%xmm1
+
+# qhasm:                   u = r
+# asm 1: movdqa <r=int6464#2,>u=int6464#7
+# asm 2: movdqa <r=%xmm1,>u=%xmm6
+movdqa %xmm1,%xmm6
+
+# qhasm: uint32323232      r >>= 14
+# asm 1: psrld $14,<r=int6464#2
+# asm 2: psrld $14,<r=%xmm1
+psrld $14,%xmm1
+
+# qhasm:                   p ^= r
+# asm 1: pxor  <r=int6464#2,<p=int6464#1
+# asm 2: pxor  <r=%xmm1,<p=%xmm0
+pxor  %xmm1,%xmm0
+
+# qhasm: uint32323232      u <<= 18
+# asm 1: pslld $18,<u=int6464#7
+# asm 2: pslld $18,<u=%xmm6
+pslld $18,%xmm6
+
+# qhasm:                   p ^= u
+# asm 1: pxor  <u=int6464#7,<p=int6464#1
+# asm 2: pxor  <u=%xmm6,<p=%xmm0
+pxor  %xmm6,%xmm0
+
+# qhasm:                   z0_stack = p
+# asm 1: movdqa <p=int6464#1,>z0_stack=stack128#21
+# asm 2: movdqa <p=%xmm0,>z0_stack=352(%esp)
+movdqa %xmm0,352(%esp)
+
+# qhasm:                                   assign xmm2 to mt
+
+# qhasm:                                   assign xmm3 to mq
+
+# qhasm:                                   assign xmm4 to mp
+
+# qhasm:                                   assign xmm5 to mr
+
+# qhasm:                                   ms = mt
+# asm 1: movdqa <mt=int6464#3,>ms=int6464#7
+# asm 2: movdqa <mt=%xmm2,>ms=%xmm6
+movdqa %xmm2,%xmm6
+
+# qhasm: uint32323232                      mt += mp
+# asm 1: paddd <mp=int6464#5,<mt=int6464#3
+# asm 2: paddd <mp=%xmm4,<mt=%xmm2
+paddd %xmm4,%xmm2
+
+# qhasm:                                   mu = mt
+# asm 1: movdqa <mt=int6464#3,>mu=int6464#1
+# asm 2: movdqa <mt=%xmm2,>mu=%xmm0
+movdqa %xmm2,%xmm0
+
+# qhasm: uint32323232                      mt >>= 25
+# asm 1: psrld $25,<mt=int6464#3
+# asm 2: psrld $25,<mt=%xmm2
+psrld $25,%xmm2
+
+# qhasm:                                   mq ^= mt
+# asm 1: pxor  <mt=int6464#3,<mq=int6464#4
+# asm 2: pxor  <mt=%xmm2,<mq=%xmm3
+pxor  %xmm2,%xmm3
+
+# qhasm: uint32323232                      mu <<= 7
+# asm 1: pslld $7,<mu=int6464#1
+# asm 2: pslld $7,<mu=%xmm0
+pslld $7,%xmm0
+
+# qhasm:                                   mq ^= mu
+# asm 1: pxor  <mu=int6464#1,<mq=int6464#4
+# asm 2: pxor  <mu=%xmm0,<mq=%xmm3
+pxor  %xmm0,%xmm3
+
+# qhasm:                                   z9_stack = mq
+# asm 1: movdqa <mq=int6464#4,>z9_stack=stack128#32
+# asm 2: movdqa <mq=%xmm3,>z9_stack=528(%esp)
+movdqa %xmm3,528(%esp)
+
+# qhasm:                                   mt = mp
+# asm 1: movdqa <mp=int6464#5,>mt=int6464#1
+# asm 2: movdqa <mp=%xmm4,>mt=%xmm0
+movdqa %xmm4,%xmm0
+
+# qhasm: uint32323232                      mt += mq
+# asm 1: paddd <mq=int6464#4,<mt=int6464#1
+# asm 2: paddd <mq=%xmm3,<mt=%xmm0
+paddd %xmm3,%xmm0
+
+# qhasm:                                   mu = mt
+# asm 1: movdqa <mt=int6464#1,>mu=int6464#2
+# asm 2: movdqa <mt=%xmm0,>mu=%xmm1
+movdqa %xmm0,%xmm1
+
+# qhasm: uint32323232                      mt >>= 23
+# asm 1: psrld $23,<mt=int6464#1
+# asm 2: psrld $23,<mt=%xmm0
+psrld $23,%xmm0
+
+# qhasm:                                   mr ^= mt
+# asm 1: pxor  <mt=int6464#1,<mr=int6464#6
+# asm 2: pxor  <mt=%xmm0,<mr=%xmm5
+pxor  %xmm0,%xmm5
+
+# qhasm: uint32323232                      mu <<= 9
+# asm 1: pslld $9,<mu=int6464#2
+# asm 2: pslld $9,<mu=%xmm1
+pslld $9,%xmm1
+
+# qhasm:                                   mr ^= mu
+# asm 1: pxor  <mu=int6464#2,<mr=int6464#6
+# asm 2: pxor  <mu=%xmm1,<mr=%xmm5
+pxor  %xmm1,%xmm5
+
+# qhasm:                                   z13_stack = mr
+# asm 1: movdqa <mr=int6464#6,>z13_stack=stack128#35
+# asm 2: movdqa <mr=%xmm5,>z13_stack=576(%esp)
+movdqa %xmm5,576(%esp)
+
+# qhasm: uint32323232                      mq += mr
+# asm 1: paddd <mr=int6464#6,<mq=int6464#4
+# asm 2: paddd <mr=%xmm5,<mq=%xmm3
+paddd %xmm5,%xmm3
+
+# qhasm:                                   mu = mq
+# asm 1: movdqa <mq=int6464#4,>mu=int6464#1
+# asm 2: movdqa <mq=%xmm3,>mu=%xmm0
+movdqa %xmm3,%xmm0
+
+# qhasm: uint32323232                      mq >>= 19
+# asm 1: psrld $19,<mq=int6464#4
+# asm 2: psrld $19,<mq=%xmm3
+psrld $19,%xmm3
+
+# qhasm:                                   ms ^= mq
+# asm 1: pxor  <mq=int6464#4,<ms=int6464#7
+# asm 2: pxor  <mq=%xmm3,<ms=%xmm6
+pxor  %xmm3,%xmm6
+
+# qhasm: uint32323232                      mu <<= 13
+# asm 1: pslld $13,<mu=int6464#1
+# asm 2: pslld $13,<mu=%xmm0
+pslld $13,%xmm0
+
+# qhasm:                                   ms ^= mu
+# asm 1: pxor  <mu=int6464#1,<ms=int6464#7
+# asm 2: pxor  <mu=%xmm0,<ms=%xmm6
+pxor  %xmm0,%xmm6
+
+# qhasm:                                                   t = z6_stack
+# asm 1: movdqa <z6_stack=stack128#26,>t=int6464#3
+# asm 2: movdqa <z6_stack=432(%esp),>t=%xmm2
+movdqa 432(%esp),%xmm2
+
+# qhasm:                                                   p = z10_stack
+# asm 1: movdqa <z10_stack=stack128#22,>p=int6464#1
+# asm 2: movdqa <z10_stack=368(%esp),>p=%xmm0
+movdqa 368(%esp),%xmm0
+
+# qhasm:                                                   q = z14_stack
+# asm 1: movdqa <z14_stack=stack128#24,>q=int6464#4
+# asm 2: movdqa <z14_stack=400(%esp),>q=%xmm3
+movdqa 400(%esp),%xmm3
+
+# qhasm:                                                   r = z2_stack
+# asm 1: movdqa <z2_stack=stack128#31,>r=int6464#2
+# asm 2: movdqa <z2_stack=512(%esp),>r=%xmm1
+movdqa 512(%esp),%xmm1
+
+# qhasm:                                   z1_stack = ms
+# asm 1: movdqa <ms=int6464#7,>z1_stack=stack128#22
+# asm 2: movdqa <ms=%xmm6,>z1_stack=368(%esp)
+movdqa %xmm6,368(%esp)
+
+# qhasm: uint32323232                      mr += ms
+# asm 1: paddd <ms=int6464#7,<mr=int6464#6
+# asm 2: paddd <ms=%xmm6,<mr=%xmm5
+paddd %xmm6,%xmm5
+
+# qhasm:                                   mu = mr
+# asm 1: movdqa <mr=int6464#6,>mu=int6464#7
+# asm 2: movdqa <mr=%xmm5,>mu=%xmm6
+movdqa %xmm5,%xmm6
+
+# qhasm: uint32323232                      mr >>= 14
+# asm 1: psrld $14,<mr=int6464#6
+# asm 2: psrld $14,<mr=%xmm5
+psrld $14,%xmm5
+
+# qhasm:                                   mp ^= mr
+# asm 1: pxor  <mr=int6464#6,<mp=int6464#5
+# asm 2: pxor  <mr=%xmm5,<mp=%xmm4
+pxor  %xmm5,%xmm4
+
+# qhasm: uint32323232                      mu <<= 18
+# asm 1: pslld $18,<mu=int6464#7
+# asm 2: pslld $18,<mu=%xmm6
+pslld $18,%xmm6
+
+# qhasm:                                   mp ^= mu
+# asm 1: pxor  <mu=int6464#7,<mp=int6464#5
+# asm 2: pxor  <mu=%xmm6,<mp=%xmm4
+pxor  %xmm6,%xmm4
+
+# qhasm:                                   z5_stack = mp
+# asm 1: movdqa <mp=int6464#5,>z5_stack=stack128#24
+# asm 2: movdqa <mp=%xmm4,>z5_stack=400(%esp)
+movdqa %xmm4,400(%esp)
+
+# qhasm:                                                   assign xmm0 to p
+
+# qhasm:                                                   assign xmm1 to r
+
+# qhasm:                                                   assign xmm2 to t
+
+# qhasm:                                                   assign xmm3 to q
+
+# qhasm:                                                   s = t
+# asm 1: movdqa <t=int6464#3,>s=int6464#7
+# asm 2: movdqa <t=%xmm2,>s=%xmm6
+movdqa %xmm2,%xmm6
+
+# qhasm: uint32323232                                      t += p
+# asm 1: paddd <p=int6464#1,<t=int6464#3
+# asm 2: paddd <p=%xmm0,<t=%xmm2
+paddd %xmm0,%xmm2
+
+# qhasm:                                                   u = t
+# asm 1: movdqa <t=int6464#3,>u=int6464#5
+# asm 2: movdqa <t=%xmm2,>u=%xmm4
+movdqa %xmm2,%xmm4
+
+# qhasm: uint32323232                                      t >>= 25
+# asm 1: psrld $25,<t=int6464#3
+# asm 2: psrld $25,<t=%xmm2
+psrld $25,%xmm2
+
+# qhasm:                                                   q ^= t
+# asm 1: pxor  <t=int6464#3,<q=int6464#4
+# asm 2: pxor  <t=%xmm2,<q=%xmm3
+pxor  %xmm2,%xmm3
+
+# qhasm: uint32323232                                      u <<= 7
+# asm 1: pslld $7,<u=int6464#5
+# asm 2: pslld $7,<u=%xmm4
+pslld $7,%xmm4
+
+# qhasm:                                                   q ^= u
+# asm 1: pxor  <u=int6464#5,<q=int6464#4
+# asm 2: pxor  <u=%xmm4,<q=%xmm3
+pxor  %xmm4,%xmm3
+
+# qhasm:                                                   z14_stack = q
+# asm 1: movdqa <q=int6464#4,>z14_stack=stack128#36
+# asm 2: movdqa <q=%xmm3,>z14_stack=592(%esp)
+movdqa %xmm3,592(%esp)
+
+# qhasm:                                                   t = p
+# asm 1: movdqa <p=int6464#1,>t=int6464#3
+# asm 2: movdqa <p=%xmm0,>t=%xmm2
+movdqa %xmm0,%xmm2
+
+# qhasm: uint32323232                                      t += q
+# asm 1: paddd <q=int6464#4,<t=int6464#3
+# asm 2: paddd <q=%xmm3,<t=%xmm2
+paddd %xmm3,%xmm2
+
+# qhasm:                                                   u = t
+# asm 1: movdqa <t=int6464#3,>u=int6464#5
+# asm 2: movdqa <t=%xmm2,>u=%xmm4
+movdqa %xmm2,%xmm4
+
+# qhasm: uint32323232                                      t >>= 23
+# asm 1: psrld $23,<t=int6464#3
+# asm 2: psrld $23,<t=%xmm2
+psrld $23,%xmm2
+
+# qhasm:                                                   r ^= t
+# asm 1: pxor  <t=int6464#3,<r=int6464#2
+# asm 2: pxor  <t=%xmm2,<r=%xmm1
+pxor  %xmm2,%xmm1
+
+# qhasm: uint32323232                                      u <<= 9
+# asm 1: pslld $9,<u=int6464#5
+# asm 2: pslld $9,<u=%xmm4
+pslld $9,%xmm4
+
+# qhasm:                                                   r ^= u
+# asm 1: pxor  <u=int6464#5,<r=int6464#2
+# asm 2: pxor  <u=%xmm4,<r=%xmm1
+pxor  %xmm4,%xmm1
+
+# qhasm:                                                   z2_stack = r
+# asm 1: movdqa <r=int6464#2,>z2_stack=stack128#26
+# asm 2: movdqa <r=%xmm1,>z2_stack=432(%esp)
+movdqa %xmm1,432(%esp)
+
+# qhasm: uint32323232                                      q += r
+# asm 1: paddd <r=int6464#2,<q=int6464#4
+# asm 2: paddd <r=%xmm1,<q=%xmm3
+paddd %xmm1,%xmm3
+
+# qhasm:                                                   u = q
+# asm 1: movdqa <q=int6464#4,>u=int6464#3
+# asm 2: movdqa <q=%xmm3,>u=%xmm2
+movdqa %xmm3,%xmm2
+
+# qhasm: uint32323232                                      q >>= 19
+# asm 1: psrld $19,<q=int6464#4
+# asm 2: psrld $19,<q=%xmm3
+psrld $19,%xmm3
+
+# qhasm:                                                   s ^= q
+# asm 1: pxor  <q=int6464#4,<s=int6464#7
+# asm 2: pxor  <q=%xmm3,<s=%xmm6
+pxor  %xmm3,%xmm6
+
+# qhasm: uint32323232                                      u <<= 13
+# asm 1: pslld $13,<u=int6464#3
+# asm 2: pslld $13,<u=%xmm2
+pslld $13,%xmm2
+
+# qhasm:                                                   s ^= u
+# asm 1: pxor  <u=int6464#3,<s=int6464#7
+# asm 2: pxor  <u=%xmm2,<s=%xmm6
+pxor  %xmm2,%xmm6
+
+# qhasm:                                                                   mt = z11_stack
+# asm 1: movdqa <z11_stack=stack128#27,>mt=int6464#3
+# asm 2: movdqa <z11_stack=448(%esp),>mt=%xmm2
+movdqa 448(%esp),%xmm2
+
+# qhasm:                                                                   mp = z15_stack
+# asm 1: movdqa <z15_stack=stack128#23,>mp=int6464#5
+# asm 2: movdqa <z15_stack=384(%esp),>mp=%xmm4
+movdqa 384(%esp),%xmm4
+
+# qhasm:                                                                   mq = z3_stack
+# asm 1: movdqa <z3_stack=stack128#25,>mq=int6464#4
+# asm 2: movdqa <z3_stack=416(%esp),>mq=%xmm3
+movdqa 416(%esp),%xmm3
+
+# qhasm:                                                                   mr = z7_stack
+# asm 1: movdqa <z7_stack=stack128#29,>mr=int6464#6
+# asm 2: movdqa <z7_stack=480(%esp),>mr=%xmm5
+movdqa 480(%esp),%xmm5
+
+# qhasm:                                                   z6_stack = s
+# asm 1: movdqa <s=int6464#7,>z6_stack=stack128#23
+# asm 2: movdqa <s=%xmm6,>z6_stack=384(%esp)
+movdqa %xmm6,384(%esp)
+
+# qhasm: uint32323232                                      r += s
+# asm 1: paddd <s=int6464#7,<r=int6464#2
+# asm 2: paddd <s=%xmm6,<r=%xmm1
+paddd %xmm6,%xmm1
+
+# qhasm:                                                   u = r
+# asm 1: movdqa <r=int6464#2,>u=int6464#7
+# asm 2: movdqa <r=%xmm1,>u=%xmm6
+movdqa %xmm1,%xmm6
+
+# qhasm: uint32323232                                      r >>= 14
+# asm 1: psrld $14,<r=int6464#2
+# asm 2: psrld $14,<r=%xmm1
+psrld $14,%xmm1
+
+# qhasm:                                                   p ^= r
+# asm 1: pxor  <r=int6464#2,<p=int6464#1
+# asm 2: pxor  <r=%xmm1,<p=%xmm0
+pxor  %xmm1,%xmm0
+
+# qhasm: uint32323232                                      u <<= 18
+# asm 1: pslld $18,<u=int6464#7
+# asm 2: pslld $18,<u=%xmm6
+pslld $18,%xmm6
+
+# qhasm:                                                   p ^= u
+# asm 1: pxor  <u=int6464#7,<p=int6464#1
+# asm 2: pxor  <u=%xmm6,<p=%xmm0
+pxor  %xmm6,%xmm0
+
+# qhasm:                                                   z10_stack = p
+# asm 1: movdqa <p=int6464#1,>z10_stack=stack128#27
+# asm 2: movdqa <p=%xmm0,>z10_stack=448(%esp)
+movdqa %xmm0,448(%esp)
+
+# qhasm:                                                                   assign xmm2 to mt
+
+# qhasm:                                                                   assign xmm3 to mq
+
+# qhasm:                                                                   assign xmm4 to mp
+
+# qhasm:                                                                   assign xmm5 to mr
+
+# qhasm:                                                                   ms = mt
+# asm 1: movdqa <mt=int6464#3,>ms=int6464#7
+# asm 2: movdqa <mt=%xmm2,>ms=%xmm6
+movdqa %xmm2,%xmm6
+
+# qhasm: uint32323232                                                      mt += mp
+# asm 1: paddd <mp=int6464#5,<mt=int6464#3
+# asm 2: paddd <mp=%xmm4,<mt=%xmm2
+paddd %xmm4,%xmm2
+
+# qhasm:                                                                   mu = mt
+# asm 1: movdqa <mt=int6464#3,>mu=int6464#1
+# asm 2: movdqa <mt=%xmm2,>mu=%xmm0
+movdqa %xmm2,%xmm0
+
+# qhasm: uint32323232                                                      mt >>= 25
+# asm 1: psrld $25,<mt=int6464#3
+# asm 2: psrld $25,<mt=%xmm2
+psrld $25,%xmm2
+
+# qhasm:                                                                   mq ^= mt
+# asm 1: pxor  <mt=int6464#3,<mq=int6464#4
+# asm 2: pxor  <mt=%xmm2,<mq=%xmm3
+pxor  %xmm2,%xmm3
+
+# qhasm: uint32323232                                                      mu <<= 7
+# asm 1: pslld $7,<mu=int6464#1
+# asm 2: pslld $7,<mu=%xmm0
+pslld $7,%xmm0
+
+# qhasm:                                                                   mq ^= mu
+# asm 1: pxor  <mu=int6464#1,<mq=int6464#4
+# asm 2: pxor  <mu=%xmm0,<mq=%xmm3
+pxor  %xmm0,%xmm3
+
+# qhasm:                                                                   z3_stack = mq
+# asm 1: movdqa <mq=int6464#4,>z3_stack=stack128#25
+# asm 2: movdqa <mq=%xmm3,>z3_stack=416(%esp)
+movdqa %xmm3,416(%esp)
+
+# qhasm:                                                                   mt = mp
+# asm 1: movdqa <mp=int6464#5,>mt=int6464#1
+# asm 2: movdqa <mp=%xmm4,>mt=%xmm0
+movdqa %xmm4,%xmm0
+
+# qhasm: uint32323232                                                      mt += mq
+# asm 1: paddd <mq=int6464#4,<mt=int6464#1
+# asm 2: paddd <mq=%xmm3,<mt=%xmm0
+paddd %xmm3,%xmm0
+
+# qhasm:                                                                   mu = mt
+# asm 1: movdqa <mt=int6464#1,>mu=int6464#2
+# asm 2: movdqa <mt=%xmm0,>mu=%xmm1
+movdqa %xmm0,%xmm1
+
+# qhasm: uint32323232                                                      mt >>= 23
+# asm 1: psrld $23,<mt=int6464#1
+# asm 2: psrld $23,<mt=%xmm0
+psrld $23,%xmm0
+
+# qhasm:                                                                   mr ^= mt
+# asm 1: pxor  <mt=int6464#1,<mr=int6464#6
+# asm 2: pxor  <mt=%xmm0,<mr=%xmm5
+pxor  %xmm0,%xmm5
+
+# qhasm: uint32323232                                                      mu <<= 9
+# asm 1: pslld $9,<mu=int6464#2
+# asm 2: pslld $9,<mu=%xmm1
+pslld $9,%xmm1
+
+# qhasm:                                                                   mr ^= mu
+# asm 1: pxor  <mu=int6464#2,<mr=int6464#6
+# asm 2: pxor  <mu=%xmm1,<mr=%xmm5
+pxor  %xmm1,%xmm5
+
+# qhasm:                                                                   z7_stack = mr
+# asm 1: movdqa <mr=int6464#6,>z7_stack=stack128#29
+# asm 2: movdqa <mr=%xmm5,>z7_stack=480(%esp)
+movdqa %xmm5,480(%esp)
+
+# qhasm: uint32323232                                                      mq += mr
+# asm 1: paddd <mr=int6464#6,<mq=int6464#4
+# asm 2: paddd <mr=%xmm5,<mq=%xmm3
+paddd %xmm5,%xmm3
+
+# qhasm:                                                                   mu = mq
+# asm 1: movdqa <mq=int6464#4,>mu=int6464#1
+# asm 2: movdqa <mq=%xmm3,>mu=%xmm0
+movdqa %xmm3,%xmm0
+
+# qhasm: uint32323232                                                      mq >>= 19
+# asm 1: psrld $19,<mq=int6464#4
+# asm 2: psrld $19,<mq=%xmm3
+psrld $19,%xmm3
+
+# qhasm:                                                                   ms ^= mq
+# asm 1: pxor  <mq=int6464#4,<ms=int6464#7
+# asm 2: pxor  <mq=%xmm3,<ms=%xmm6
+pxor  %xmm3,%xmm6
+
+# qhasm: uint32323232                                                      mu <<= 13
+# asm 1: pslld $13,<mu=int6464#1
+# asm 2: pslld $13,<mu=%xmm0
+pslld $13,%xmm0
+
+# qhasm:                                                                   ms ^= mu
+# asm 1: pxor  <mu=int6464#1,<ms=int6464#7
+# asm 2: pxor  <mu=%xmm0,<ms=%xmm6
+pxor  %xmm0,%xmm6
+
+# qhasm:                   t = z3_stack
+# asm 1: movdqa <z3_stack=stack128#25,>t=int6464#3
+# asm 2: movdqa <z3_stack=416(%esp),>t=%xmm2
+movdqa 416(%esp),%xmm2
+
+# qhasm:                   p = z0_stack
+# asm 1: movdqa <z0_stack=stack128#21,>p=int6464#1
+# asm 2: movdqa <z0_stack=352(%esp),>p=%xmm0
+movdqa 352(%esp),%xmm0
+
+# qhasm:                   q = z1_stack
+# asm 1: movdqa <z1_stack=stack128#22,>q=int6464#4
+# asm 2: movdqa <z1_stack=368(%esp),>q=%xmm3
+movdqa 368(%esp),%xmm3
+
+# qhasm:                   r = z2_stack
+# asm 1: movdqa <z2_stack=stack128#26,>r=int6464#2
+# asm 2: movdqa <z2_stack=432(%esp),>r=%xmm1
+movdqa 432(%esp),%xmm1
+
+# qhasm:                                                                   z11_stack = ms
+# asm 1: movdqa <ms=int6464#7,>z11_stack=stack128#21
+# asm 2: movdqa <ms=%xmm6,>z11_stack=352(%esp)
+movdqa %xmm6,352(%esp)
+
+# qhasm: uint32323232                                                      mr += ms
+# asm 1: paddd <ms=int6464#7,<mr=int6464#6
+# asm 2: paddd <ms=%xmm6,<mr=%xmm5
+paddd %xmm6,%xmm5
+
+# qhasm:                                                                   mu = mr
+# asm 1: movdqa <mr=int6464#6,>mu=int6464#7
+# asm 2: movdqa <mr=%xmm5,>mu=%xmm6
+movdqa %xmm5,%xmm6
+
+# qhasm: uint32323232                                                      mr >>= 14
+# asm 1: psrld $14,<mr=int6464#6
+# asm 2: psrld $14,<mr=%xmm5
+psrld $14,%xmm5
+
+# qhasm:                                                                   mp ^= mr
+# asm 1: pxor  <mr=int6464#6,<mp=int6464#5
+# asm 2: pxor  <mr=%xmm5,<mp=%xmm4
+pxor  %xmm5,%xmm4
+
+# qhasm: uint32323232                                                      mu <<= 18
+# asm 1: pslld $18,<mu=int6464#7
+# asm 2: pslld $18,<mu=%xmm6
+pslld $18,%xmm6
+
+# qhasm:                                                                   mp ^= mu
+# asm 1: pxor  <mu=int6464#7,<mp=int6464#5
+# asm 2: pxor  <mu=%xmm6,<mp=%xmm4
+pxor  %xmm6,%xmm4
+
+# qhasm:                                                                   z15_stack = mp
+# asm 1: movdqa <mp=int6464#5,>z15_stack=stack128#22
+# asm 2: movdqa <mp=%xmm4,>z15_stack=368(%esp)
+movdqa %xmm4,368(%esp)
+
+# qhasm:                   assign xmm0 to p
+
+# qhasm:                   assign xmm1 to r
+
+# qhasm:                   assign xmm2 to t
+
+# qhasm:                   assign xmm3 to q
+
+# qhasm:                   s = t
+# asm 1: movdqa <t=int6464#3,>s=int6464#7
+# asm 2: movdqa <t=%xmm2,>s=%xmm6
+movdqa %xmm2,%xmm6
+
+# qhasm: uint32323232      t += p
+# asm 1: paddd <p=int6464#1,<t=int6464#3
+# asm 2: paddd <p=%xmm0,<t=%xmm2
+paddd %xmm0,%xmm2
+
+# qhasm:                   u = t
+# asm 1: movdqa <t=int6464#3,>u=int6464#5
+# asm 2: movdqa <t=%xmm2,>u=%xmm4
+movdqa %xmm2,%xmm4
+
+# qhasm: uint32323232      t >>= 25
+# asm 1: psrld $25,<t=int6464#3
+# asm 2: psrld $25,<t=%xmm2
+psrld $25,%xmm2
+
+# qhasm:                   q ^= t
+# asm 1: pxor  <t=int6464#3,<q=int6464#4
+# asm 2: pxor  <t=%xmm2,<q=%xmm3
+pxor  %xmm2,%xmm3
+
+# qhasm: uint32323232      u <<= 7
+# asm 1: pslld $7,<u=int6464#5
+# asm 2: pslld $7,<u=%xmm4
+pslld $7,%xmm4
+
+# qhasm:                   q ^= u
+# asm 1: pxor  <u=int6464#5,<q=int6464#4
+# asm 2: pxor  <u=%xmm4,<q=%xmm3
+pxor  %xmm4,%xmm3
+
+# qhasm:                   z1_stack = q
+# asm 1: movdqa <q=int6464#4,>z1_stack=stack128#28
+# asm 2: movdqa <q=%xmm3,>z1_stack=464(%esp)
+movdqa %xmm3,464(%esp)
+
+# qhasm:                   t = p
+# asm 1: movdqa <p=int6464#1,>t=int6464#3
+# asm 2: movdqa <p=%xmm0,>t=%xmm2
+movdqa %xmm0,%xmm2
+
+# qhasm: uint32323232      t += q
+# asm 1: paddd <q=int6464#4,<t=int6464#3
+# asm 2: paddd <q=%xmm3,<t=%xmm2
+paddd %xmm3,%xmm2
+
+# qhasm:                   u = t
+# asm 1: movdqa <t=int6464#3,>u=int6464#5
+# asm 2: movdqa <t=%xmm2,>u=%xmm4
+movdqa %xmm2,%xmm4
+
+# qhasm: uint32323232      t >>= 23
+# asm 1: psrld $23,<t=int6464#3
+# asm 2: psrld $23,<t=%xmm2
+psrld $23,%xmm2
+
+# qhasm:                   r ^= t
+# asm 1: pxor  <t=int6464#3,<r=int6464#2
+# asm 2: pxor  <t=%xmm2,<r=%xmm1
+pxor  %xmm2,%xmm1
+
+# qhasm: uint32323232      u <<= 9
+# asm 1: pslld $9,<u=int6464#5
+# asm 2: pslld $9,<u=%xmm4
+pslld $9,%xmm4
+
+# qhasm:                   r ^= u
+# asm 1: pxor  <u=int6464#5,<r=int6464#2
+# asm 2: pxor  <u=%xmm4,<r=%xmm1
+pxor  %xmm4,%xmm1
+
+# qhasm:                   z2_stack = r
+# asm 1: movdqa <r=int6464#2,>z2_stack=stack128#31
+# asm 2: movdqa <r=%xmm1,>z2_stack=512(%esp)
+movdqa %xmm1,512(%esp)
+
+# qhasm: uint32323232      q += r
+# asm 1: paddd <r=int6464#2,<q=int6464#4
+# asm 2: paddd <r=%xmm1,<q=%xmm3
+paddd %xmm1,%xmm3
+
+# qhasm:                   u = q
+# asm 1: movdqa <q=int6464#4,>u=int6464#3
+# asm 2: movdqa <q=%xmm3,>u=%xmm2
+movdqa %xmm3,%xmm2
+
+# qhasm: uint32323232      q >>= 19
+# asm 1: psrld $19,<q=int6464#4
+# asm 2: psrld $19,<q=%xmm3
+psrld $19,%xmm3
+
+# qhasm:                   s ^= q
+# asm 1: pxor  <q=int6464#4,<s=int6464#7
+# asm 2: pxor  <q=%xmm3,<s=%xmm6
+pxor  %xmm3,%xmm6
+
+# qhasm: uint32323232      u <<= 13
+# asm 1: pslld $13,<u=int6464#3
+# asm 2: pslld $13,<u=%xmm2
+pslld $13,%xmm2
+
+# qhasm:                   s ^= u
+# asm 1: pxor  <u=int6464#3,<s=int6464#7
+# asm 2: pxor  <u=%xmm2,<s=%xmm6
+pxor  %xmm2,%xmm6
+
+# qhasm:                                   mt = z4_stack
+# asm 1: movdqa <z4_stack=stack128#33,>mt=int6464#3
+# asm 2: movdqa <z4_stack=544(%esp),>mt=%xmm2
+movdqa 544(%esp),%xmm2
+
+# qhasm:                                   mp = z5_stack
+# asm 1: movdqa <z5_stack=stack128#24,>mp=int6464#5
+# asm 2: movdqa <z5_stack=400(%esp),>mp=%xmm4
+movdqa 400(%esp),%xmm4
+
+# qhasm:                                   mq = z6_stack
+# asm 1: movdqa <z6_stack=stack128#23,>mq=int6464#4
+# asm 2: movdqa <z6_stack=384(%esp),>mq=%xmm3
+movdqa 384(%esp),%xmm3
+
+# qhasm:                                   mr = z7_stack
+# asm 1: movdqa <z7_stack=stack128#29,>mr=int6464#6
+# asm 2: movdqa <z7_stack=480(%esp),>mr=%xmm5
+movdqa 480(%esp),%xmm5
+
+# qhasm:                   z3_stack = s
+# asm 1: movdqa <s=int6464#7,>z3_stack=stack128#25
+# asm 2: movdqa <s=%xmm6,>z3_stack=416(%esp)
+movdqa %xmm6,416(%esp)
+
+# qhasm: uint32323232      r += s
+# asm 1: paddd <s=int6464#7,<r=int6464#2
+# asm 2: paddd <s=%xmm6,<r=%xmm1
+paddd %xmm6,%xmm1
+
+# qhasm:                   u = r
+# asm 1: movdqa <r=int6464#2,>u=int6464#7
+# asm 2: movdqa <r=%xmm1,>u=%xmm6
+movdqa %xmm1,%xmm6
+
+# qhasm: uint32323232      r >>= 14
+# asm 1: psrld $14,<r=int6464#2
+# asm 2: psrld $14,<r=%xmm1
+psrld $14,%xmm1
+
+# qhasm:                   p ^= r
+# asm 1: pxor  <r=int6464#2,<p=int6464#1
+# asm 2: pxor  <r=%xmm1,<p=%xmm0
+pxor  %xmm1,%xmm0
+
+# qhasm: uint32323232      u <<= 18
+# asm 1: pslld $18,<u=int6464#7
+# asm 2: pslld $18,<u=%xmm6
+pslld $18,%xmm6
+
+# qhasm:                   p ^= u
+# asm 1: pxor  <u=int6464#7,<p=int6464#1
+# asm 2: pxor  <u=%xmm6,<p=%xmm0
+pxor  %xmm6,%xmm0
+
+# qhasm:                   z0_stack = p
+# asm 1: movdqa <p=int6464#1,>z0_stack=stack128#33
+# asm 2: movdqa <p=%xmm0,>z0_stack=544(%esp)
+movdqa %xmm0,544(%esp)
+
+# qhasm:                                   assign xmm2 to mt
+
+# qhasm:                                   assign xmm3 to mq
+
+# qhasm:                                   assign xmm4 to mp
+
+# qhasm:                                   assign xmm5 to mr
+
+# qhasm:                                   ms = mt
+# asm 1: movdqa <mt=int6464#3,>ms=int6464#7
+# asm 2: movdqa <mt=%xmm2,>ms=%xmm6
+movdqa %xmm2,%xmm6
+
+# qhasm: uint32323232                      mt += mp
+# asm 1: paddd <mp=int6464#5,<mt=int6464#3
+# asm 2: paddd <mp=%xmm4,<mt=%xmm2
+paddd %xmm4,%xmm2
+
+# qhasm:                                   mu = mt
+# asm 1: movdqa <mt=int6464#3,>mu=int6464#1
+# asm 2: movdqa <mt=%xmm2,>mu=%xmm0
+movdqa %xmm2,%xmm0
+
+# qhasm: uint32323232                      mt >>= 25
+# asm 1: psrld $25,<mt=int6464#3
+# asm 2: psrld $25,<mt=%xmm2
+psrld $25,%xmm2
+
+# qhasm:                                   mq ^= mt
+# asm 1: pxor  <mt=int6464#3,<mq=int6464#4
+# asm 2: pxor  <mt=%xmm2,<mq=%xmm3
+pxor  %xmm2,%xmm3
+
+# qhasm: uint32323232                      mu <<= 7
+# asm 1: pslld $7,<mu=int6464#1
+# asm 2: pslld $7,<mu=%xmm0
+pslld $7,%xmm0
+
+# qhasm:                                   mq ^= mu
+# asm 1: pxor  <mu=int6464#1,<mq=int6464#4
+# asm 2: pxor  <mu=%xmm0,<mq=%xmm3
+pxor  %xmm0,%xmm3
+
+# qhasm:                                   z6_stack = mq
+# asm 1: movdqa <mq=int6464#4,>z6_stack=stack128#26
+# asm 2: movdqa <mq=%xmm3,>z6_stack=432(%esp)
+movdqa %xmm3,432(%esp)
+
+# qhasm:                                   mt = mp
+# asm 1: movdqa <mp=int6464#5,>mt=int6464#1
+# asm 2: movdqa <mp=%xmm4,>mt=%xmm0
+movdqa %xmm4,%xmm0
+
+# qhasm: uint32323232                      mt += mq
+# asm 1: paddd <mq=int6464#4,<mt=int6464#1
+# asm 2: paddd <mq=%xmm3,<mt=%xmm0
+paddd %xmm3,%xmm0
+
+# qhasm:                                   mu = mt
+# asm 1: movdqa <mt=int6464#1,>mu=int6464#2
+# asm 2: movdqa <mt=%xmm0,>mu=%xmm1
+movdqa %xmm0,%xmm1
+
+# qhasm: uint32323232                      mt >>= 23
+# asm 1: psrld $23,<mt=int6464#1
+# asm 2: psrld $23,<mt=%xmm0
+psrld $23,%xmm0
+
+# qhasm:                                   mr ^= mt
+# asm 1: pxor  <mt=int6464#1,<mr=int6464#6
+# asm 2: pxor  <mt=%xmm0,<mr=%xmm5
+pxor  %xmm0,%xmm5
+
+# qhasm: uint32323232                      mu <<= 9
+# asm 1: pslld $9,<mu=int6464#2
+# asm 2: pslld $9,<mu=%xmm1
+pslld $9,%xmm1
+
+# qhasm:                                   mr ^= mu
+# asm 1: pxor  <mu=int6464#2,<mr=int6464#6
+# asm 2: pxor  <mu=%xmm1,<mr=%xmm5
+pxor  %xmm1,%xmm5
+
+# qhasm:                                   z7_stack = mr
+# asm 1: movdqa <mr=int6464#6,>z7_stack=stack128#29
+# asm 2: movdqa <mr=%xmm5,>z7_stack=480(%esp)
+movdqa %xmm5,480(%esp)
+
+# qhasm: uint32323232                      mq += mr
+# asm 1: paddd <mr=int6464#6,<mq=int6464#4
+# asm 2: paddd <mr=%xmm5,<mq=%xmm3
+paddd %xmm5,%xmm3
+
+# qhasm:                                   mu = mq
+# asm 1: movdqa <mq=int6464#4,>mu=int6464#1
+# asm 2: movdqa <mq=%xmm3,>mu=%xmm0
+movdqa %xmm3,%xmm0
+
+# qhasm: uint32323232                      mq >>= 19
+# asm 1: psrld $19,<mq=int6464#4
+# asm 2: psrld $19,<mq=%xmm3
+psrld $19,%xmm3
+
+# qhasm:                                   ms ^= mq
+# asm 1: pxor  <mq=int6464#4,<ms=int6464#7
+# asm 2: pxor  <mq=%xmm3,<ms=%xmm6
+pxor  %xmm3,%xmm6
+
+# qhasm: uint32323232                      mu <<= 13
+# asm 1: pslld $13,<mu=int6464#1
+# asm 2: pslld $13,<mu=%xmm0
+pslld $13,%xmm0
+
+# qhasm:                                   ms ^= mu
+# asm 1: pxor  <mu=int6464#1,<ms=int6464#7
+# asm 2: pxor  <mu=%xmm0,<ms=%xmm6
+pxor  %xmm0,%xmm6
+
+# qhasm:                                                   t = z9_stack
+# asm 1: movdqa <z9_stack=stack128#32,>t=int6464#3
+# asm 2: movdqa <z9_stack=528(%esp),>t=%xmm2
+movdqa 528(%esp),%xmm2
+
+# qhasm:                                                   p = z10_stack
+# asm 1: movdqa <z10_stack=stack128#27,>p=int6464#1
+# asm 2: movdqa <z10_stack=448(%esp),>p=%xmm0
+movdqa 448(%esp),%xmm0
+
+# qhasm:                                                   q = z11_stack
+# asm 1: movdqa <z11_stack=stack128#21,>q=int6464#4
+# asm 2: movdqa <z11_stack=352(%esp),>q=%xmm3
+movdqa 352(%esp),%xmm3
+
+# qhasm:                                                   r = z8_stack
+# asm 1: movdqa <z8_stack=stack128#34,>r=int6464#2
+# asm 2: movdqa <z8_stack=560(%esp),>r=%xmm1
+movdqa 560(%esp),%xmm1
+
+# qhasm:                                   z4_stack = ms
+# asm 1: movdqa <ms=int6464#7,>z4_stack=stack128#34
+# asm 2: movdqa <ms=%xmm6,>z4_stack=560(%esp)
+movdqa %xmm6,560(%esp)
+
+# qhasm: uint32323232                      mr += ms
+# asm 1: paddd <ms=int6464#7,<mr=int6464#6
+# asm 2: paddd <ms=%xmm6,<mr=%xmm5
+paddd %xmm6,%xmm5
+
+# qhasm:                                   mu = mr
+# asm 1: movdqa <mr=int6464#6,>mu=int6464#7
+# asm 2: movdqa <mr=%xmm5,>mu=%xmm6
+movdqa %xmm5,%xmm6
+
+# qhasm: uint32323232                      mr >>= 14
+# asm 1: psrld $14,<mr=int6464#6
+# asm 2: psrld $14,<mr=%xmm5
+psrld $14,%xmm5
+
+# qhasm:                                   mp ^= mr
+# asm 1: pxor  <mr=int6464#6,<mp=int6464#5
+# asm 2: pxor  <mr=%xmm5,<mp=%xmm4
+pxor  %xmm5,%xmm4
+
+# qhasm: uint32323232                      mu <<= 18
+# asm 1: pslld $18,<mu=int6464#7
+# asm 2: pslld $18,<mu=%xmm6
+pslld $18,%xmm6
+
+# qhasm:                                   mp ^= mu
+# asm 1: pxor  <mu=int6464#7,<mp=int6464#5
+# asm 2: pxor  <mu=%xmm6,<mp=%xmm4
+pxor  %xmm6,%xmm4
+
+# qhasm:                                   z5_stack = mp
+# asm 1: movdqa <mp=int6464#5,>z5_stack=stack128#21
+# asm 2: movdqa <mp=%xmm4,>z5_stack=352(%esp)
+movdqa %xmm4,352(%esp)
+
+# qhasm:                                                   assign xmm0 to p
+
+# qhasm:                                                   assign xmm1 to r
+
+# qhasm:                                                   assign xmm2 to t
+
+# qhasm:                                                   assign xmm3 to q
+
+# qhasm:                                                   s = t
+# asm 1: movdqa <t=int6464#3,>s=int6464#7
+# asm 2: movdqa <t=%xmm2,>s=%xmm6
+movdqa %xmm2,%xmm6
+
+# qhasm: uint32323232                                      t += p
+# asm 1: paddd <p=int6464#1,<t=int6464#3
+# asm 2: paddd <p=%xmm0,<t=%xmm2
+paddd %xmm0,%xmm2
+
+# qhasm:                                                   u = t
+# asm 1: movdqa <t=int6464#3,>u=int6464#5
+# asm 2: movdqa <t=%xmm2,>u=%xmm4
+movdqa %xmm2,%xmm4
+
+# qhasm: uint32323232                                      t >>= 25
+# asm 1: psrld $25,<t=int6464#3
+# asm 2: psrld $25,<t=%xmm2
+psrld $25,%xmm2
+
+# qhasm:                                                   q ^= t
+# asm 1: pxor  <t=int6464#3,<q=int6464#4
+# asm 2: pxor  <t=%xmm2,<q=%xmm3
+pxor  %xmm2,%xmm3
+
+# qhasm: uint32323232                                      u <<= 7
+# asm 1: pslld $7,<u=int6464#5
+# asm 2: pslld $7,<u=%xmm4
+pslld $7,%xmm4
+
+# qhasm:                                                   q ^= u
+# asm 1: pxor  <u=int6464#5,<q=int6464#4
+# asm 2: pxor  <u=%xmm4,<q=%xmm3
+pxor  %xmm4,%xmm3
+
+# qhasm:                                                   z11_stack = q
+# asm 1: movdqa <q=int6464#4,>z11_stack=stack128#27
+# asm 2: movdqa <q=%xmm3,>z11_stack=448(%esp)
+movdqa %xmm3,448(%esp)
+
+# qhasm:                                                   t = p
+# asm 1: movdqa <p=int6464#1,>t=int6464#3
+# asm 2: movdqa <p=%xmm0,>t=%xmm2
+movdqa %xmm0,%xmm2
+
+# qhasm: uint32323232                                      t += q
+# asm 1: paddd <q=int6464#4,<t=int6464#3
+# asm 2: paddd <q=%xmm3,<t=%xmm2
+paddd %xmm3,%xmm2
+
+# qhasm:                                                   u = t
+# asm 1: movdqa <t=int6464#3,>u=int6464#5
+# asm 2: movdqa <t=%xmm2,>u=%xmm4
+movdqa %xmm2,%xmm4
+
+# qhasm: uint32323232                                      t >>= 23
+# asm 1: psrld $23,<t=int6464#3
+# asm 2: psrld $23,<t=%xmm2
+psrld $23,%xmm2
+
+# qhasm:                                                   r ^= t
+# asm 1: pxor  <t=int6464#3,<r=int6464#2
+# asm 2: pxor  <t=%xmm2,<r=%xmm1
+pxor  %xmm2,%xmm1
+
+# qhasm: uint32323232                                      u <<= 9
+# asm 1: pslld $9,<u=int6464#5
+# asm 2: pslld $9,<u=%xmm4
+pslld $9,%xmm4
+
+# qhasm:                                                   r ^= u
+# asm 1: pxor  <u=int6464#5,<r=int6464#2
+# asm 2: pxor  <u=%xmm4,<r=%xmm1
+pxor  %xmm4,%xmm1
+
+# qhasm:                                                   z8_stack = r
+# asm 1: movdqa <r=int6464#2,>z8_stack=stack128#37
+# asm 2: movdqa <r=%xmm1,>z8_stack=608(%esp)
+movdqa %xmm1,608(%esp)
+
+# qhasm: uint32323232                                      q += r
+# asm 1: paddd <r=int6464#2,<q=int6464#4
+# asm 2: paddd <r=%xmm1,<q=%xmm3
+paddd %xmm1,%xmm3
+
+# qhasm:                                                   u = q
+# asm 1: movdqa <q=int6464#4,>u=int6464#3
+# asm 2: movdqa <q=%xmm3,>u=%xmm2
+movdqa %xmm3,%xmm2
+
+# qhasm: uint32323232                                      q >>= 19
+# asm 1: psrld $19,<q=int6464#4
+# asm 2: psrld $19,<q=%xmm3
+psrld $19,%xmm3
+
+# qhasm:                                                   s ^= q
+# asm 1: pxor  <q=int6464#4,<s=int6464#7
+# asm 2: pxor  <q=%xmm3,<s=%xmm6
+pxor  %xmm3,%xmm6
+
+# qhasm: uint32323232                                      u <<= 13
+# asm 1: pslld $13,<u=int6464#3
+# asm 2: pslld $13,<u=%xmm2
+pslld $13,%xmm2
+
+# qhasm:                                                   s ^= u
+# asm 1: pxor  <u=int6464#3,<s=int6464#7
+# asm 2: pxor  <u=%xmm2,<s=%xmm6
+pxor  %xmm2,%xmm6
+
+# qhasm:                                                                   mt = z14_stack
+# asm 1: movdqa <z14_stack=stack128#36,>mt=int6464#3
+# asm 2: movdqa <z14_stack=592(%esp),>mt=%xmm2
+movdqa 592(%esp),%xmm2
+
+# qhasm:                                                                   mp = z15_stack
+# asm 1: movdqa <z15_stack=stack128#22,>mp=int6464#5
+# asm 2: movdqa <z15_stack=368(%esp),>mp=%xmm4
+movdqa 368(%esp),%xmm4
+
+# qhasm:                                                                   mq = z12_stack
+# asm 1: movdqa <z12_stack=stack128#30,>mq=int6464#4
+# asm 2: movdqa <z12_stack=496(%esp),>mq=%xmm3
+movdqa 496(%esp),%xmm3
+
+# qhasm:                                                                   mr = z13_stack
+# asm 1: movdqa <z13_stack=stack128#35,>mr=int6464#6
+# asm 2: movdqa <z13_stack=576(%esp),>mr=%xmm5
+movdqa 576(%esp),%xmm5
+
+# qhasm:                                                   z9_stack = s
+# asm 1: movdqa <s=int6464#7,>z9_stack=stack128#32
+# asm 2: movdqa <s=%xmm6,>z9_stack=528(%esp)
+movdqa %xmm6,528(%esp)
+
+# qhasm: uint32323232                                      r += s
+# asm 1: paddd <s=int6464#7,<r=int6464#2
+# asm 2: paddd <s=%xmm6,<r=%xmm1
+paddd %xmm6,%xmm1
+
+# qhasm:                                                   u = r
+# asm 1: movdqa <r=int6464#2,>u=int6464#7
+# asm 2: movdqa <r=%xmm1,>u=%xmm6
+movdqa %xmm1,%xmm6
+
+# qhasm: uint32323232                                      r >>= 14
+# asm 1: psrld $14,<r=int6464#2
+# asm 2: psrld $14,<r=%xmm1
+psrld $14,%xmm1
+
+# qhasm:                                                   p ^= r
+# asm 1: pxor  <r=int6464#2,<p=int6464#1
+# asm 2: pxor  <r=%xmm1,<p=%xmm0
+pxor  %xmm1,%xmm0
+
+# qhasm: uint32323232                                      u <<= 18
+# asm 1: pslld $18,<u=int6464#7
+# asm 2: pslld $18,<u=%xmm6
+pslld $18,%xmm6
+
+# qhasm:                                                   p ^= u
+# asm 1: pxor  <u=int6464#7,<p=int6464#1
+# asm 2: pxor  <u=%xmm6,<p=%xmm0
+pxor  %xmm6,%xmm0
+
+# qhasm:                                                   z10_stack = p
+# asm 1: movdqa <p=int6464#1,>z10_stack=stack128#22
+# asm 2: movdqa <p=%xmm0,>z10_stack=368(%esp)
+movdqa %xmm0,368(%esp)
+
+# qhasm:                                                                   assign xmm2 to mt
+
+# qhasm:                                                                   assign xmm3 to mq
+
+# qhasm:                                                                   assign xmm4 to mp
+
+# qhasm:                                                                   assign xmm5 to mr
+
+# qhasm:                                                                   ms = mt
+# asm 1: movdqa <mt=int6464#3,>ms=int6464#7
+# asm 2: movdqa <mt=%xmm2,>ms=%xmm6
+movdqa %xmm2,%xmm6
+
+# qhasm: uint32323232                                                      mt += mp
+# asm 1: paddd <mp=int6464#5,<mt=int6464#3
+# asm 2: paddd <mp=%xmm4,<mt=%xmm2
+paddd %xmm4,%xmm2
+
+# qhasm:                                                                   mu = mt
+# asm 1: movdqa <mt=int6464#3,>mu=int6464#1
+# asm 2: movdqa <mt=%xmm2,>mu=%xmm0
+movdqa %xmm2,%xmm0
+
+# qhasm: uint32323232                                                      mt >>= 25
+# asm 1: psrld $25,<mt=int6464#3
+# asm 2: psrld $25,<mt=%xmm2
+psrld $25,%xmm2
+
+# qhasm:                                                                   mq ^= mt
+# asm 1: pxor  <mt=int6464#3,<mq=int6464#4
+# asm 2: pxor  <mt=%xmm2,<mq=%xmm3
+pxor  %xmm2,%xmm3
+
+# qhasm: uint32323232                                                      mu <<= 7
+# asm 1: pslld $7,<mu=int6464#1
+# asm 2: pslld $7,<mu=%xmm0
+pslld $7,%xmm0
+
+# qhasm:                                                                   mq ^= mu
+# asm 1: pxor  <mu=int6464#1,<mq=int6464#4
+# asm 2: pxor  <mu=%xmm0,<mq=%xmm3
+pxor  %xmm0,%xmm3
+
+# qhasm:                                                                   z12_stack = mq
+# asm 1: movdqa <mq=int6464#4,>z12_stack=stack128#35
+# asm 2: movdqa <mq=%xmm3,>z12_stack=576(%esp)
+movdqa %xmm3,576(%esp)
+
+# qhasm:                                                                   mt = mp
+# asm 1: movdqa <mp=int6464#5,>mt=int6464#1
+# asm 2: movdqa <mp=%xmm4,>mt=%xmm0
+movdqa %xmm4,%xmm0
+
+# qhasm: uint32323232                                                      mt += mq
+# asm 1: paddd <mq=int6464#4,<mt=int6464#1
+# asm 2: paddd <mq=%xmm3,<mt=%xmm0
+paddd %xmm3,%xmm0
+
+# qhasm:                                                                   mu = mt
+# asm 1: movdqa <mt=int6464#1,>mu=int6464#2
+# asm 2: movdqa <mt=%xmm0,>mu=%xmm1
+movdqa %xmm0,%xmm1
+
+# qhasm: uint32323232                                                      mt >>= 23
+# asm 1: psrld $23,<mt=int6464#1
+# asm 2: psrld $23,<mt=%xmm0
+psrld $23,%xmm0
+
+# qhasm:                                                                   mr ^= mt
+# asm 1: pxor  <mt=int6464#1,<mr=int6464#6
+# asm 2: pxor  <mt=%xmm0,<mr=%xmm5
+pxor  %xmm0,%xmm5
+
+# qhasm: uint32323232                                                      mu <<= 9
+# asm 1: pslld $9,<mu=int6464#2
+# asm 2: pslld $9,<mu=%xmm1
+pslld $9,%xmm1
+
+# qhasm:                                                                   mr ^= mu
+# asm 1: pxor  <mu=int6464#2,<mr=int6464#6
+# asm 2: pxor  <mu=%xmm1,<mr=%xmm5
+pxor  %xmm1,%xmm5
+
+# qhasm:                                                                   z13_stack = mr
+# asm 1: movdqa <mr=int6464#6,>z13_stack=stack128#30
+# asm 2: movdqa <mr=%xmm5,>z13_stack=496(%esp)
+movdqa %xmm5,496(%esp)
+
+# qhasm: uint32323232                                                      mq += mr
+# asm 1: paddd <mr=int6464#6,<mq=int6464#4
+# asm 2: paddd <mr=%xmm5,<mq=%xmm3
+paddd %xmm5,%xmm3
+
+# qhasm:                                                                   mu = mq
+# asm 1: movdqa <mq=int6464#4,>mu=int6464#1
+# asm 2: movdqa <mq=%xmm3,>mu=%xmm0
+movdqa %xmm3,%xmm0
+
+# qhasm: uint32323232                                                      mq >>= 19
+# asm 1: psrld $19,<mq=int6464#4
+# asm 2: psrld $19,<mq=%xmm3
+psrld $19,%xmm3
+
+# qhasm:                                                                   ms ^= mq
+# asm 1: pxor  <mq=int6464#4,<ms=int6464#7
+# asm 2: pxor  <mq=%xmm3,<ms=%xmm6
+pxor  %xmm3,%xmm6
+
+# qhasm: uint32323232                                                      mu <<= 13
+# asm 1: pslld $13,<mu=int6464#1
+# asm 2: pslld $13,<mu=%xmm0
+pslld $13,%xmm0
+
+# qhasm:                                                                   ms ^= mu
+# asm 1: pxor  <mu=int6464#1,<ms=int6464#7
+# asm 2: pxor  <mu=%xmm0,<ms=%xmm6
+pxor  %xmm0,%xmm6
+
+# qhasm:                   t = z12_stack
+# asm 1: movdqa <z12_stack=stack128#35,>t=int6464#3
+# asm 2: movdqa <z12_stack=576(%esp),>t=%xmm2
+movdqa 576(%esp),%xmm2
+
+# qhasm:                   p = z0_stack
+# asm 1: movdqa <z0_stack=stack128#33,>p=int6464#1
+# asm 2: movdqa <z0_stack=544(%esp),>p=%xmm0
+movdqa 544(%esp),%xmm0
+
+# qhasm:                   q = z4_stack
+# asm 1: movdqa <z4_stack=stack128#34,>q=int6464#4
+# asm 2: movdqa <z4_stack=560(%esp),>q=%xmm3
+movdqa 560(%esp),%xmm3
+
+# qhasm:                   r = z8_stack
+# asm 1: movdqa <z8_stack=stack128#37,>r=int6464#2
+# asm 2: movdqa <z8_stack=608(%esp),>r=%xmm1
+movdqa 608(%esp),%xmm1
+
+# qhasm:                                                                   z14_stack = ms
+# asm 1: movdqa <ms=int6464#7,>z14_stack=stack128#24
+# asm 2: movdqa <ms=%xmm6,>z14_stack=400(%esp)
+movdqa %xmm6,400(%esp)
+
+# qhasm: uint32323232                                                      mr += ms
+# asm 1: paddd <ms=int6464#7,<mr=int6464#6
+# asm 2: paddd <ms=%xmm6,<mr=%xmm5
+paddd %xmm6,%xmm5
+
+# qhasm:                                                                   mu = mr
+# asm 1: movdqa <mr=int6464#6,>mu=int6464#7
+# asm 2: movdqa <mr=%xmm5,>mu=%xmm6
+movdqa %xmm5,%xmm6
+
+# qhasm: uint32323232                                                      mr >>= 14
+# asm 1: psrld $14,<mr=int6464#6
+# asm 2: psrld $14,<mr=%xmm5
+psrld $14,%xmm5
+
+# qhasm:                                                                   mp ^= mr
+# asm 1: pxor  <mr=int6464#6,<mp=int6464#5
+# asm 2: pxor  <mr=%xmm5,<mp=%xmm4
+pxor  %xmm5,%xmm4
+
+# qhasm: uint32323232                                                      mu <<= 18
+# asm 1: pslld $18,<mu=int6464#7
+# asm 2: pslld $18,<mu=%xmm6
+pslld $18,%xmm6
+
+# qhasm:                                                                   mp ^= mu
+# asm 1: pxor  <mu=int6464#7,<mp=int6464#5
+# asm 2: pxor  <mu=%xmm6,<mp=%xmm4
+pxor  %xmm6,%xmm4
+
+# qhasm:                                                                   z15_stack = mp
+# asm 1: movdqa <mp=int6464#5,>z15_stack=stack128#23
+# asm 2: movdqa <mp=%xmm4,>z15_stack=384(%esp)
+movdqa %xmm4,384(%esp)
+
+# qhasm:                   unsigned>? i -= 2
+# asm 1: sub  $2,<i=int32#1
+# asm 2: sub  $2,<i=%eax
+sub  $2,%eax
+# comment:fp stack unchanged by jump
+
+# qhasm: goto mainloop1 if unsigned>
+ja ._mainloop1
+
+# qhasm:   out = out_stack
+# asm 1: movl <out_stack=stack32#6,>out=int32#6
+# asm 2: movl <out_stack=20(%esp),>out=%edi
+movl 20(%esp),%edi
+
+# qhasm:   z0 = z0_stack
+# asm 1: movdqa <z0_stack=stack128#33,>z0=int6464#1
+# asm 2: movdqa <z0_stack=544(%esp),>z0=%xmm0
+movdqa 544(%esp),%xmm0
+
+# qhasm:   z1 = z1_stack
+# asm 1: movdqa <z1_stack=stack128#28,>z1=int6464#2
+# asm 2: movdqa <z1_stack=464(%esp),>z1=%xmm1
+movdqa 464(%esp),%xmm1
+
+# qhasm:   z2 = z2_stack
+# asm 1: movdqa <z2_stack=stack128#31,>z2=int6464#3
+# asm 2: movdqa <z2_stack=512(%esp),>z2=%xmm2
+movdqa 512(%esp),%xmm2
+
+# qhasm:   z3 = z3_stack
+# asm 1: movdqa <z3_stack=stack128#25,>z3=int6464#4
+# asm 2: movdqa <z3_stack=416(%esp),>z3=%xmm3
+movdqa 416(%esp),%xmm3
+
+# qhasm:   uint32323232 z0 += orig0
+# asm 1: paddd <orig0=stack128#8,<z0=int6464#1
+# asm 2: paddd <orig0=144(%esp),<z0=%xmm0
+paddd 144(%esp),%xmm0
+
+# qhasm:   uint32323232 z1 += orig1
+# asm 1: paddd <orig1=stack128#12,<z1=int6464#2
+# asm 2: paddd <orig1=208(%esp),<z1=%xmm1
+paddd 208(%esp),%xmm1
+
+# qhasm:   uint32323232 z2 += orig2
+# asm 1: paddd <orig2=stack128#15,<z2=int6464#3
+# asm 2: paddd <orig2=256(%esp),<z2=%xmm2
+paddd 256(%esp),%xmm2
+
+# qhasm:   uint32323232 z3 += orig3
+# asm 1: paddd <orig3=stack128#18,<z3=int6464#4
+# asm 2: paddd <orig3=304(%esp),<z3=%xmm3
+paddd 304(%esp),%xmm3
+
+# qhasm:   in0 = z0
+# asm 1: movd   <z0=int6464#1,>in0=int32#1
+# asm 2: movd   <z0=%xmm0,>in0=%eax
+movd   %xmm0,%eax
+
+# qhasm:   in1 = z1
+# asm 1: movd   <z1=int6464#2,>in1=int32#2
+# asm 2: movd   <z1=%xmm1,>in1=%ecx
+movd   %xmm1,%ecx
+
+# qhasm:   in2 = z2
+# asm 1: movd   <z2=int6464#3,>in2=int32#3
+# asm 2: movd   <z2=%xmm2,>in2=%edx
+movd   %xmm2,%edx
+
+# qhasm:   in3 = z3
+# asm 1: movd   <z3=int6464#4,>in3=int32#4
+# asm 2: movd   <z3=%xmm3,>in3=%ebx
+movd   %xmm3,%ebx
+
+# qhasm:   z0 <<<= 96
+# asm 1: pshufd $0x39,<z0=int6464#1,<z0=int6464#1
+# asm 2: pshufd $0x39,<z0=%xmm0,<z0=%xmm0
+pshufd $0x39,%xmm0,%xmm0
+
+# qhasm:   z1 <<<= 96
+# asm 1: pshufd $0x39,<z1=int6464#2,<z1=int6464#2
+# asm 2: pshufd $0x39,<z1=%xmm1,<z1=%xmm1
+pshufd $0x39,%xmm1,%xmm1
+
+# qhasm:   z2 <<<= 96
+# asm 1: pshufd $0x39,<z2=int6464#3,<z2=int6464#3
+# asm 2: pshufd $0x39,<z2=%xmm2,<z2=%xmm2
+pshufd $0x39,%xmm2,%xmm2
+
+# qhasm:   z3 <<<= 96
+# asm 1: pshufd $0x39,<z3=int6464#4,<z3=int6464#4
+# asm 2: pshufd $0x39,<z3=%xmm3,<z3=%xmm3
+pshufd $0x39,%xmm3,%xmm3
+
+# qhasm:   in0 ^= *(uint32 *) (m + 0)
+# asm 1: xorl 0(<m=int32#5),<in0=int32#1
+# asm 2: xorl 0(<m=%esi),<in0=%eax
+xorl 0(%esi),%eax
+
+# qhasm:   in1 ^= *(uint32 *) (m + 4)
+# asm 1: xorl 4(<m=int32#5),<in1=int32#2
+# asm 2: xorl 4(<m=%esi),<in1=%ecx
+xorl 4(%esi),%ecx
+
+# qhasm:   in2 ^= *(uint32 *) (m + 8)
+# asm 1: xorl 8(<m=int32#5),<in2=int32#3
+# asm 2: xorl 8(<m=%esi),<in2=%edx
+xorl 8(%esi),%edx
+
+# qhasm:   in3 ^= *(uint32 *) (m + 12)
+# asm 1: xorl 12(<m=int32#5),<in3=int32#4
+# asm 2: xorl 12(<m=%esi),<in3=%ebx
+xorl 12(%esi),%ebx
+
+# qhasm:   *(uint32 *) (out + 0) = in0
+# asm 1: movl <in0=int32#1,0(<out=int32#6)
+# asm 2: movl <in0=%eax,0(<out=%edi)
+movl %eax,0(%edi)
+
+# qhasm:   *(uint32 *) (out + 4) = in1
+# asm 1: movl <in1=int32#2,4(<out=int32#6)
+# asm 2: movl <in1=%ecx,4(<out=%edi)
+movl %ecx,4(%edi)
+
+# qhasm:   *(uint32 *) (out + 8) = in2
+# asm 1: movl <in2=int32#3,8(<out=int32#6)
+# asm 2: movl <in2=%edx,8(<out=%edi)
+movl %edx,8(%edi)
+
+# qhasm:   *(uint32 *) (out + 12) = in3
+# asm 1: movl <in3=int32#4,12(<out=int32#6)
+# asm 2: movl <in3=%ebx,12(<out=%edi)
+movl %ebx,12(%edi)
+
+# qhasm:   in0 = z0
+# asm 1: movd   <z0=int6464#1,>in0=int32#1
+# asm 2: movd   <z0=%xmm0,>in0=%eax
+movd   %xmm0,%eax
+
+# qhasm:   in1 = z1
+# asm 1: movd   <z1=int6464#2,>in1=int32#2
+# asm 2: movd   <z1=%xmm1,>in1=%ecx
+movd   %xmm1,%ecx
+
+# qhasm:   in2 = z2
+# asm 1: movd   <z2=int6464#3,>in2=int32#3
+# asm 2: movd   <z2=%xmm2,>in2=%edx
+movd   %xmm2,%edx
+
+# qhasm:   in3 = z3
+# asm 1: movd   <z3=int6464#4,>in3=int32#4
+# asm 2: movd   <z3=%xmm3,>in3=%ebx
+movd   %xmm3,%ebx
+
+# qhasm:   z0 <<<= 96
+# asm 1: pshufd $0x39,<z0=int6464#1,<z0=int6464#1
+# asm 2: pshufd $0x39,<z0=%xmm0,<z0=%xmm0
+pshufd $0x39,%xmm0,%xmm0
+
+# qhasm:   z1 <<<= 96
+# asm 1: pshufd $0x39,<z1=int6464#2,<z1=int6464#2
+# asm 2: pshufd $0x39,<z1=%xmm1,<z1=%xmm1
+pshufd $0x39,%xmm1,%xmm1
+
+# qhasm:   z2 <<<= 96
+# asm 1: pshufd $0x39,<z2=int6464#3,<z2=int6464#3
+# asm 2: pshufd $0x39,<z2=%xmm2,<z2=%xmm2
+pshufd $0x39,%xmm2,%xmm2
+
+# qhasm:   z3 <<<= 96
+# asm 1: pshufd $0x39,<z3=int6464#4,<z3=int6464#4
+# asm 2: pshufd $0x39,<z3=%xmm3,<z3=%xmm3
+pshufd $0x39,%xmm3,%xmm3
+
+# qhasm:   in0 ^= *(uint32 *) (m + 64)
+# asm 1: xorl 64(<m=int32#5),<in0=int32#1
+# asm 2: xorl 64(<m=%esi),<in0=%eax
+xorl 64(%esi),%eax
+
+# qhasm:   in1 ^= *(uint32 *) (m + 68)
+# asm 1: xorl 68(<m=int32#5),<in1=int32#2
+# asm 2: xorl 68(<m=%esi),<in1=%ecx
+xorl 68(%esi),%ecx
+
+# qhasm:   in2 ^= *(uint32 *) (m + 72)
+# asm 1: xorl 72(<m=int32#5),<in2=int32#3
+# asm 2: xorl 72(<m=%esi),<in2=%edx
+xorl 72(%esi),%edx
+
+# qhasm:   in3 ^= *(uint32 *) (m + 76)
+# asm 1: xorl 76(<m=int32#5),<in3=int32#4
+# asm 2: xorl 76(<m=%esi),<in3=%ebx
+xorl 76(%esi),%ebx
+
+# qhasm:   *(uint32 *) (out + 64) = in0
+# asm 1: movl <in0=int32#1,64(<out=int32#6)
+# asm 2: movl <in0=%eax,64(<out=%edi)
+movl %eax,64(%edi)
+
+# qhasm:   *(uint32 *) (out + 68) = in1
+# asm 1: movl <in1=int32#2,68(<out=int32#6)
+# asm 2: movl <in1=%ecx,68(<out=%edi)
+movl %ecx,68(%edi)
+
+# qhasm:   *(uint32 *) (out + 72) = in2
+# asm 1: movl <in2=int32#3,72(<out=int32#6)
+# asm 2: movl <in2=%edx,72(<out=%edi)
+movl %edx,72(%edi)
+
+# qhasm:   *(uint32 *) (out + 76) = in3
+# asm 1: movl <in3=int32#4,76(<out=int32#6)
+# asm 2: movl <in3=%ebx,76(<out=%edi)
+movl %ebx,76(%edi)
+
+# qhasm:   in0 = z0
+# asm 1: movd   <z0=int6464#1,>in0=int32#1
+# asm 2: movd   <z0=%xmm0,>in0=%eax
+movd   %xmm0,%eax
+
+# qhasm:   in1 = z1
+# asm 1: movd   <z1=int6464#2,>in1=int32#2
+# asm 2: movd   <z1=%xmm1,>in1=%ecx
+movd   %xmm1,%ecx
+
+# qhasm:   in2 = z2
+# asm 1: movd   <z2=int6464#3,>in2=int32#3
+# asm 2: movd   <z2=%xmm2,>in2=%edx
+movd   %xmm2,%edx
+
+# qhasm:   in3 = z3
+# asm 1: movd   <z3=int6464#4,>in3=int32#4
+# asm 2: movd   <z3=%xmm3,>in3=%ebx
+movd   %xmm3,%ebx
+
+# qhasm:   z0 <<<= 96
+# asm 1: pshufd $0x39,<z0=int6464#1,<z0=int6464#1
+# asm 2: pshufd $0x39,<z0=%xmm0,<z0=%xmm0
+pshufd $0x39,%xmm0,%xmm0
+
+# qhasm:   z1 <<<= 96
+# asm 1: pshufd $0x39,<z1=int6464#2,<z1=int6464#2
+# asm 2: pshufd $0x39,<z1=%xmm1,<z1=%xmm1
+pshufd $0x39,%xmm1,%xmm1
+
+# qhasm:   z2 <<<= 96
+# asm 1: pshufd $0x39,<z2=int6464#3,<z2=int6464#3
+# asm 2: pshufd $0x39,<z2=%xmm2,<z2=%xmm2
+pshufd $0x39,%xmm2,%xmm2
+
+# qhasm:   z3 <<<= 96
+# asm 1: pshufd $0x39,<z3=int6464#4,<z3=int6464#4
+# asm 2: pshufd $0x39,<z3=%xmm3,<z3=%xmm3
+pshufd $0x39,%xmm3,%xmm3
+
+# qhasm:   in0 ^= *(uint32 *) (m + 128)
+# asm 1: xorl 128(<m=int32#5),<in0=int32#1
+# asm 2: xorl 128(<m=%esi),<in0=%eax
+xorl 128(%esi),%eax
+
+# qhasm:   in1 ^= *(uint32 *) (m + 132)
+# asm 1: xorl 132(<m=int32#5),<in1=int32#2
+# asm 2: xorl 132(<m=%esi),<in1=%ecx
+xorl 132(%esi),%ecx
+
+# qhasm:   in2 ^= *(uint32 *) (m + 136)
+# asm 1: xorl 136(<m=int32#5),<in2=int32#3
+# asm 2: xorl 136(<m=%esi),<in2=%edx
+xorl 136(%esi),%edx
+
+# qhasm:   in3 ^= *(uint32 *) (m + 140)
+# asm 1: xorl 140(<m=int32#5),<in3=int32#4
+# asm 2: xorl 140(<m=%esi),<in3=%ebx
+xorl 140(%esi),%ebx
+
+# qhasm:   *(uint32 *) (out + 128) = in0
+# asm 1: movl <in0=int32#1,128(<out=int32#6)
+# asm 2: movl <in0=%eax,128(<out=%edi)
+movl %eax,128(%edi)
+
+# qhasm:   *(uint32 *) (out + 132) = in1
+# asm 1: movl <in1=int32#2,132(<out=int32#6)
+# asm 2: movl <in1=%ecx,132(<out=%edi)
+movl %ecx,132(%edi)
+
+# qhasm:   *(uint32 *) (out + 136) = in2
+# asm 1: movl <in2=int32#3,136(<out=int32#6)
+# asm 2: movl <in2=%edx,136(<out=%edi)
+movl %edx,136(%edi)
+
+# qhasm:   *(uint32 *) (out + 140) = in3
+# asm 1: movl <in3=int32#4,140(<out=int32#6)
+# asm 2: movl <in3=%ebx,140(<out=%edi)
+movl %ebx,140(%edi)
+
+# qhasm:   in0 = z0
+# asm 1: movd   <z0=int6464#1,>in0=int32#1
+# asm 2: movd   <z0=%xmm0,>in0=%eax
+movd   %xmm0,%eax
+
+# qhasm:   in1 = z1
+# asm 1: movd   <z1=int6464#2,>in1=int32#2
+# asm 2: movd   <z1=%xmm1,>in1=%ecx
+movd   %xmm1,%ecx
+
+# qhasm:   in2 = z2
+# asm 1: movd   <z2=int6464#3,>in2=int32#3
+# asm 2: movd   <z2=%xmm2,>in2=%edx
+movd   %xmm2,%edx
+
+# qhasm:   in3 = z3
+# asm 1: movd   <z3=int6464#4,>in3=int32#4
+# asm 2: movd   <z3=%xmm3,>in3=%ebx
+movd   %xmm3,%ebx
+
+# qhasm:   in0 ^= *(uint32 *) (m + 192)
+# asm 1: xorl 192(<m=int32#5),<in0=int32#1
+# asm 2: xorl 192(<m=%esi),<in0=%eax
+xorl 192(%esi),%eax
+
+# qhasm:   in1 ^= *(uint32 *) (m + 196)
+# asm 1: xorl 196(<m=int32#5),<in1=int32#2
+# asm 2: xorl 196(<m=%esi),<in1=%ecx
+xorl 196(%esi),%ecx
+
+# qhasm:   in2 ^= *(uint32 *) (m + 200)
+# asm 1: xorl 200(<m=int32#5),<in2=int32#3
+# asm 2: xorl 200(<m=%esi),<in2=%edx
+xorl 200(%esi),%edx
+
+# qhasm:   in3 ^= *(uint32 *) (m + 204)
+# asm 1: xorl 204(<m=int32#5),<in3=int32#4
+# asm 2: xorl 204(<m=%esi),<in3=%ebx
+xorl 204(%esi),%ebx
+
+# qhasm:   *(uint32 *) (out + 192) = in0
+# asm 1: movl <in0=int32#1,192(<out=int32#6)
+# asm 2: movl <in0=%eax,192(<out=%edi)
+movl %eax,192(%edi)
+
+# qhasm:   *(uint32 *) (out + 196) = in1
+# asm 1: movl <in1=int32#2,196(<out=int32#6)
+# asm 2: movl <in1=%ecx,196(<out=%edi)
+movl %ecx,196(%edi)
+
+# qhasm:   *(uint32 *) (out + 200) = in2
+# asm 1: movl <in2=int32#3,200(<out=int32#6)
+# asm 2: movl <in2=%edx,200(<out=%edi)
+movl %edx,200(%edi)
+
+# qhasm:   *(uint32 *) (out + 204) = in3
+# asm 1: movl <in3=int32#4,204(<out=int32#6)
+# asm 2: movl <in3=%ebx,204(<out=%edi)
+movl %ebx,204(%edi)
+
+# qhasm:   z4 = z4_stack
+# asm 1: movdqa <z4_stack=stack128#34,>z4=int6464#1
+# asm 2: movdqa <z4_stack=560(%esp),>z4=%xmm0
+movdqa 560(%esp),%xmm0
+
+# qhasm:   z5 = z5_stack
+# asm 1: movdqa <z5_stack=stack128#21,>z5=int6464#2
+# asm 2: movdqa <z5_stack=352(%esp),>z5=%xmm1
+movdqa 352(%esp),%xmm1
+
+# qhasm:   z6 = z6_stack
+# asm 1: movdqa <z6_stack=stack128#26,>z6=int6464#3
+# asm 2: movdqa <z6_stack=432(%esp),>z6=%xmm2
+movdqa 432(%esp),%xmm2
+
+# qhasm:   z7 = z7_stack
+# asm 1: movdqa <z7_stack=stack128#29,>z7=int6464#4
+# asm 2: movdqa <z7_stack=480(%esp),>z7=%xmm3
+movdqa 480(%esp),%xmm3
+
+# qhasm:   uint32323232 z4 += orig4
+# asm 1: paddd <orig4=stack128#16,<z4=int6464#1
+# asm 2: paddd <orig4=272(%esp),<z4=%xmm0
+paddd 272(%esp),%xmm0
+
+# qhasm:   uint32323232 z5 += orig5
+# asm 1: paddd <orig5=stack128#5,<z5=int6464#2
+# asm 2: paddd <orig5=96(%esp),<z5=%xmm1
+paddd 96(%esp),%xmm1
+
+# qhasm:   uint32323232 z6 += orig6
+# asm 1: paddd <orig6=stack128#9,<z6=int6464#3
+# asm 2: paddd <orig6=160(%esp),<z6=%xmm2
+paddd 160(%esp),%xmm2
+
+# qhasm:   uint32323232 z7 += orig7
+# asm 1: paddd <orig7=stack128#13,<z7=int6464#4
+# asm 2: paddd <orig7=224(%esp),<z7=%xmm3
+paddd 224(%esp),%xmm3
+
+# qhasm:   in4 = z4
+# asm 1: movd   <z4=int6464#1,>in4=int32#1
+# asm 2: movd   <z4=%xmm0,>in4=%eax
+movd   %xmm0,%eax
+
+# qhasm:   in5 = z5
+# asm 1: movd   <z5=int6464#2,>in5=int32#2
+# asm 2: movd   <z5=%xmm1,>in5=%ecx
+movd   %xmm1,%ecx
+
+# qhasm:   in6 = z6
+# asm 1: movd   <z6=int6464#3,>in6=int32#3
+# asm 2: movd   <z6=%xmm2,>in6=%edx
+movd   %xmm2,%edx
+
+# qhasm:   in7 = z7
+# asm 1: movd   <z7=int6464#4,>in7=int32#4
+# asm 2: movd   <z7=%xmm3,>in7=%ebx
+movd   %xmm3,%ebx
+
+# qhasm:   z4 <<<= 96
+# asm 1: pshufd $0x39,<z4=int6464#1,<z4=int6464#1
+# asm 2: pshufd $0x39,<z4=%xmm0,<z4=%xmm0
+pshufd $0x39,%xmm0,%xmm0
+
+# qhasm:   z5 <<<= 96
+# asm 1: pshufd $0x39,<z5=int6464#2,<z5=int6464#2
+# asm 2: pshufd $0x39,<z5=%xmm1,<z5=%xmm1
+pshufd $0x39,%xmm1,%xmm1
+
+# qhasm:   z6 <<<= 96
+# asm 1: pshufd $0x39,<z6=int6464#3,<z6=int6464#3
+# asm 2: pshufd $0x39,<z6=%xmm2,<z6=%xmm2
+pshufd $0x39,%xmm2,%xmm2
+
+# qhasm:   z7 <<<= 96
+# asm 1: pshufd $0x39,<z7=int6464#4,<z7=int6464#4
+# asm 2: pshufd $0x39,<z7=%xmm3,<z7=%xmm3
+pshufd $0x39,%xmm3,%xmm3
+
+# qhasm:   in4 ^= *(uint32 *) (m + 16)
+# asm 1: xorl 16(<m=int32#5),<in4=int32#1
+# asm 2: xorl 16(<m=%esi),<in4=%eax
+xorl 16(%esi),%eax
+
+# qhasm:   in5 ^= *(uint32 *) (m + 20)
+# asm 1: xorl 20(<m=int32#5),<in5=int32#2
+# asm 2: xorl 20(<m=%esi),<in5=%ecx
+xorl 20(%esi),%ecx
+
+# qhasm:   in6 ^= *(uint32 *) (m + 24)
+# asm 1: xorl 24(<m=int32#5),<in6=int32#3
+# asm 2: xorl 24(<m=%esi),<in6=%edx
+xorl 24(%esi),%edx
+
+# qhasm:   in7 ^= *(uint32 *) (m + 28)
+# asm 1: xorl 28(<m=int32#5),<in7=int32#4
+# asm 2: xorl 28(<m=%esi),<in7=%ebx
+xorl 28(%esi),%ebx
+
+# qhasm:   *(uint32 *) (out + 16) = in4
+# asm 1: movl <in4=int32#1,16(<out=int32#6)
+# asm 2: movl <in4=%eax,16(<out=%edi)
+movl %eax,16(%edi)
+
+# qhasm:   *(uint32 *) (out + 20) = in5
+# asm 1: movl <in5=int32#2,20(<out=int32#6)
+# asm 2: movl <in5=%ecx,20(<out=%edi)
+movl %ecx,20(%edi)
+
+# qhasm:   *(uint32 *) (out + 24) = in6
+# asm 1: movl <in6=int32#3,24(<out=int32#6)
+# asm 2: movl <in6=%edx,24(<out=%edi)
+movl %edx,24(%edi)
+
+# qhasm:   *(uint32 *) (out + 28) = in7
+# asm 1: movl <in7=int32#4,28(<out=int32#6)
+# asm 2: movl <in7=%ebx,28(<out=%edi)
+movl %ebx,28(%edi)
+
+# qhasm:   in4 = z4
+# asm 1: movd   <z4=int6464#1,>in4=int32#1
+# asm 2: movd   <z4=%xmm0,>in4=%eax
+movd   %xmm0,%eax
+
+# qhasm:   in5 = z5
+# asm 1: movd   <z5=int6464#2,>in5=int32#2
+# asm 2: movd   <z5=%xmm1,>in5=%ecx
+movd   %xmm1,%ecx
+
+# qhasm:   in6 = z6
+# asm 1: movd   <z6=int6464#3,>in6=int32#3
+# asm 2: movd   <z6=%xmm2,>in6=%edx
+movd   %xmm2,%edx
+
+# qhasm:   in7 = z7
+# asm 1: movd   <z7=int6464#4,>in7=int32#4
+# asm 2: movd   <z7=%xmm3,>in7=%ebx
+movd   %xmm3,%ebx
+
+# qhasm:   z4 <<<= 96
+# asm 1: pshufd $0x39,<z4=int6464#1,<z4=int6464#1
+# asm 2: pshufd $0x39,<z4=%xmm0,<z4=%xmm0
+pshufd $0x39,%xmm0,%xmm0
+
+# qhasm:   z5 <<<= 96
+# asm 1: pshufd $0x39,<z5=int6464#2,<z5=int6464#2
+# asm 2: pshufd $0x39,<z5=%xmm1,<z5=%xmm1
+pshufd $0x39,%xmm1,%xmm1
+
+# qhasm:   z6 <<<= 96
+# asm 1: pshufd $0x39,<z6=int6464#3,<z6=int6464#3
+# asm 2: pshufd $0x39,<z6=%xmm2,<z6=%xmm2
+pshufd $0x39,%xmm2,%xmm2
+
+# qhasm:   z7 <<<= 96
+# asm 1: pshufd $0x39,<z7=int6464#4,<z7=int6464#4
+# asm 2: pshufd $0x39,<z7=%xmm3,<z7=%xmm3
+pshufd $0x39,%xmm3,%xmm3
+
+# qhasm:   in4 ^= *(uint32 *) (m + 80)
+# asm 1: xorl 80(<m=int32#5),<in4=int32#1
+# asm 2: xorl 80(<m=%esi),<in4=%eax
+xorl 80(%esi),%eax
+
+# qhasm:   in5 ^= *(uint32 *) (m + 84)
+# asm 1: xorl 84(<m=int32#5),<in5=int32#2
+# asm 2: xorl 84(<m=%esi),<in5=%ecx
+xorl 84(%esi),%ecx
+
+# qhasm:   in6 ^= *(uint32 *) (m + 88)
+# asm 1: xorl 88(<m=int32#5),<in6=int32#3
+# asm 2: xorl 88(<m=%esi),<in6=%edx
+xorl 88(%esi),%edx
+
+# qhasm:   in7 ^= *(uint32 *) (m + 92)
+# asm 1: xorl 92(<m=int32#5),<in7=int32#4
+# asm 2: xorl 92(<m=%esi),<in7=%ebx
+xorl 92(%esi),%ebx
+
+# qhasm:   *(uint32 *) (out + 80) = in4
+# asm 1: movl <in4=int32#1,80(<out=int32#6)
+# asm 2: movl <in4=%eax,80(<out=%edi)
+movl %eax,80(%edi)
+
+# qhasm:   *(uint32 *) (out + 84) = in5
+# asm 1: movl <in5=int32#2,84(<out=int32#6)
+# asm 2: movl <in5=%ecx,84(<out=%edi)
+movl %ecx,84(%edi)
+
+# qhasm:   *(uint32 *) (out + 88) = in6
+# asm 1: movl <in6=int32#3,88(<out=int32#6)
+# asm 2: movl <in6=%edx,88(<out=%edi)
+movl %edx,88(%edi)
+
+# qhasm:   *(uint32 *) (out + 92) = in7
+# asm 1: movl <in7=int32#4,92(<out=int32#6)
+# asm 2: movl <in7=%ebx,92(<out=%edi)
+movl %ebx,92(%edi)
+
+# qhasm:   in4 = z4
+# asm 1: movd   <z4=int6464#1,>in4=int32#1
+# asm 2: movd   <z4=%xmm0,>in4=%eax
+movd   %xmm0,%eax
+
+# qhasm:   in5 = z5
+# asm 1: movd   <z5=int6464#2,>in5=int32#2
+# asm 2: movd   <z5=%xmm1,>in5=%ecx
+movd   %xmm1,%ecx
+
+# qhasm:   in6 = z6
+# asm 1: movd   <z6=int6464#3,>in6=int32#3
+# asm 2: movd   <z6=%xmm2,>in6=%edx
+movd   %xmm2,%edx
+
+# qhasm:   in7 = z7
+# asm 1: movd   <z7=int6464#4,>in7=int32#4
+# asm 2: movd   <z7=%xmm3,>in7=%ebx
+movd   %xmm3,%ebx
+
+# qhasm:   z4 <<<= 96
+# asm 1: pshufd $0x39,<z4=int6464#1,<z4=int6464#1
+# asm 2: pshufd $0x39,<z4=%xmm0,<z4=%xmm0
+pshufd $0x39,%xmm0,%xmm0
+
+# qhasm:   z5 <<<= 96
+# asm 1: pshufd $0x39,<z5=int6464#2,<z5=int6464#2
+# asm 2: pshufd $0x39,<z5=%xmm1,<z5=%xmm1
+pshufd $0x39,%xmm1,%xmm1
+
+# qhasm:   z6 <<<= 96
+# asm 1: pshufd $0x39,<z6=int6464#3,<z6=int6464#3
+# asm 2: pshufd $0x39,<z6=%xmm2,<z6=%xmm2
+pshufd $0x39,%xmm2,%xmm2
+
+# qhasm:   z7 <<<= 96
+# asm 1: pshufd $0x39,<z7=int6464#4,<z7=int6464#4
+# asm 2: pshufd $0x39,<z7=%xmm3,<z7=%xmm3
+pshufd $0x39,%xmm3,%xmm3
+
+# qhasm:   in4 ^= *(uint32 *) (m + 144)
+# asm 1: xorl 144(<m=int32#5),<in4=int32#1
+# asm 2: xorl 144(<m=%esi),<in4=%eax
+xorl 144(%esi),%eax
+
+# qhasm:   in5 ^= *(uint32 *) (m + 148)
+# asm 1: xorl 148(<m=int32#5),<in5=int32#2
+# asm 2: xorl 148(<m=%esi),<in5=%ecx
+xorl 148(%esi),%ecx
+
+# qhasm:   in6 ^= *(uint32 *) (m + 152)
+# asm 1: xorl 152(<m=int32#5),<in6=int32#3
+# asm 2: xorl 152(<m=%esi),<in6=%edx
+xorl 152(%esi),%edx
+
+# qhasm:   in7 ^= *(uint32 *) (m + 156)
+# asm 1: xorl 156(<m=int32#5),<in7=int32#4
+# asm 2: xorl 156(<m=%esi),<in7=%ebx
+xorl 156(%esi),%ebx
+
+# qhasm:   *(uint32 *) (out + 144) = in4
+# asm 1: movl <in4=int32#1,144(<out=int32#6)
+# asm 2: movl <in4=%eax,144(<out=%edi)
+movl %eax,144(%edi)
+
+# qhasm:   *(uint32 *) (out + 148) = in5
+# asm 1: movl <in5=int32#2,148(<out=int32#6)
+# asm 2: movl <in5=%ecx,148(<out=%edi)
+movl %ecx,148(%edi)
+
+# qhasm:   *(uint32 *) (out + 152) = in6
+# asm 1: movl <in6=int32#3,152(<out=int32#6)
+# asm 2: movl <in6=%edx,152(<out=%edi)
+movl %edx,152(%edi)
+
+# qhasm:   *(uint32 *) (out + 156) = in7
+# asm 1: movl <in7=int32#4,156(<out=int32#6)
+# asm 2: movl <in7=%ebx,156(<out=%edi)
+movl %ebx,156(%edi)
+
+# qhasm:   in4 = z4
+# asm 1: movd   <z4=int6464#1,>in4=int32#1
+# asm 2: movd   <z4=%xmm0,>in4=%eax
+movd   %xmm0,%eax
+
+# qhasm:   in5 = z5
+# asm 1: movd   <z5=int6464#2,>in5=int32#2
+# asm 2: movd   <z5=%xmm1,>in5=%ecx
+movd   %xmm1,%ecx
+
+# qhasm:   in6 = z6
+# asm 1: movd   <z6=int6464#3,>in6=int32#3
+# asm 2: movd   <z6=%xmm2,>in6=%edx
+movd   %xmm2,%edx
+
+# qhasm:   in7 = z7
+# asm 1: movd   <z7=int6464#4,>in7=int32#4
+# asm 2: movd   <z7=%xmm3,>in7=%ebx
+movd   %xmm3,%ebx
+
+# qhasm:   in4 ^= *(uint32 *) (m + 208)
+# asm 1: xorl 208(<m=int32#5),<in4=int32#1
+# asm 2: xorl 208(<m=%esi),<in4=%eax
+xorl 208(%esi),%eax
+
+# qhasm:   in5 ^= *(uint32 *) (m + 212)
+# asm 1: xorl 212(<m=int32#5),<in5=int32#2
+# asm 2: xorl 212(<m=%esi),<in5=%ecx
+xorl 212(%esi),%ecx
+
+# qhasm:   in6 ^= *(uint32 *) (m + 216)
+# asm 1: xorl 216(<m=int32#5),<in6=int32#3
+# asm 2: xorl 216(<m=%esi),<in6=%edx
+xorl 216(%esi),%edx
+
+# qhasm:   in7 ^= *(uint32 *) (m + 220)
+# asm 1: xorl 220(<m=int32#5),<in7=int32#4
+# asm 2: xorl 220(<m=%esi),<in7=%ebx
+xorl 220(%esi),%ebx
+
+# qhasm:   *(uint32 *) (out + 208) = in4
+# asm 1: movl <in4=int32#1,208(<out=int32#6)
+# asm 2: movl <in4=%eax,208(<out=%edi)
+movl %eax,208(%edi)
+
+# qhasm:   *(uint32 *) (out + 212) = in5
+# asm 1: movl <in5=int32#2,212(<out=int32#6)
+# asm 2: movl <in5=%ecx,212(<out=%edi)
+movl %ecx,212(%edi)
+
+# qhasm:   *(uint32 *) (out + 216) = in6
+# asm 1: movl <in6=int32#3,216(<out=int32#6)
+# asm 2: movl <in6=%edx,216(<out=%edi)
+movl %edx,216(%edi)
+
+# qhasm:   *(uint32 *) (out + 220) = in7
+# asm 1: movl <in7=int32#4,220(<out=int32#6)
+# asm 2: movl <in7=%ebx,220(<out=%edi)
+movl %ebx,220(%edi)
+
+# qhasm:   z8 = z8_stack
+# asm 1: movdqa <z8_stack=stack128#37,>z8=int6464#1
+# asm 2: movdqa <z8_stack=608(%esp),>z8=%xmm0
+movdqa 608(%esp),%xmm0
+
+# qhasm:   z9 = z9_stack
+# asm 1: movdqa <z9_stack=stack128#32,>z9=int6464#2
+# asm 2: movdqa <z9_stack=528(%esp),>z9=%xmm1
+movdqa 528(%esp),%xmm1
+
+# qhasm:   z10 = z10_stack
+# asm 1: movdqa <z10_stack=stack128#22,>z10=int6464#3
+# asm 2: movdqa <z10_stack=368(%esp),>z10=%xmm2
+movdqa 368(%esp),%xmm2
+
+# qhasm:   z11 = z11_stack
+# asm 1: movdqa <z11_stack=stack128#27,>z11=int6464#4
+# asm 2: movdqa <z11_stack=448(%esp),>z11=%xmm3
+movdqa 448(%esp),%xmm3
+
+# qhasm:   uint32323232 z8 += orig8
+# asm 1: paddd <orig8=stack128#19,<z8=int6464#1
+# asm 2: paddd <orig8=320(%esp),<z8=%xmm0
+paddd 320(%esp),%xmm0
+
+# qhasm:   uint32323232 z9 += orig9
+# asm 1: paddd <orig9=stack128#20,<z9=int6464#2
+# asm 2: paddd <orig9=336(%esp),<z9=%xmm1
+paddd 336(%esp),%xmm1
+
+# qhasm:   uint32323232 z10 += orig10
+# asm 1: paddd <orig10=stack128#6,<z10=int6464#3
+# asm 2: paddd <orig10=112(%esp),<z10=%xmm2
+paddd 112(%esp),%xmm2
+
+# qhasm:   uint32323232 z11 += orig11
+# asm 1: paddd <orig11=stack128#10,<z11=int6464#4
+# asm 2: paddd <orig11=176(%esp),<z11=%xmm3
+paddd 176(%esp),%xmm3
+
+# qhasm:   in8 = z8
+# asm 1: movd   <z8=int6464#1,>in8=int32#1
+# asm 2: movd   <z8=%xmm0,>in8=%eax
+movd   %xmm0,%eax
+
+# qhasm:   in9 = z9
+# asm 1: movd   <z9=int6464#2,>in9=int32#2
+# asm 2: movd   <z9=%xmm1,>in9=%ecx
+movd   %xmm1,%ecx
+
+# qhasm:   in10 = z10
+# asm 1: movd   <z10=int6464#3,>in10=int32#3
+# asm 2: movd   <z10=%xmm2,>in10=%edx
+movd   %xmm2,%edx
+
+# qhasm:   in11 = z11
+# asm 1: movd   <z11=int6464#4,>in11=int32#4
+# asm 2: movd   <z11=%xmm3,>in11=%ebx
+movd   %xmm3,%ebx
+
+# qhasm:   z8 <<<= 96
+# asm 1: pshufd $0x39,<z8=int6464#1,<z8=int6464#1
+# asm 2: pshufd $0x39,<z8=%xmm0,<z8=%xmm0
+pshufd $0x39,%xmm0,%xmm0
+
+# qhasm:   z9 <<<= 96
+# asm 1: pshufd $0x39,<z9=int6464#2,<z9=int6464#2
+# asm 2: pshufd $0x39,<z9=%xmm1,<z9=%xmm1
+pshufd $0x39,%xmm1,%xmm1
+
+# qhasm:   z10 <<<= 96
+# asm 1: pshufd $0x39,<z10=int6464#3,<z10=int6464#3
+# asm 2: pshufd $0x39,<z10=%xmm2,<z10=%xmm2
+pshufd $0x39,%xmm2,%xmm2
+
+# qhasm:   z11 <<<= 96
+# asm 1: pshufd $0x39,<z11=int6464#4,<z11=int6464#4
+# asm 2: pshufd $0x39,<z11=%xmm3,<z11=%xmm3
+pshufd $0x39,%xmm3,%xmm3
+
+# qhasm:   in8 ^= *(uint32 *) (m + 32)
+# asm 1: xorl 32(<m=int32#5),<in8=int32#1
+# asm 2: xorl 32(<m=%esi),<in8=%eax
+xorl 32(%esi),%eax
+
+# qhasm:   in9 ^= *(uint32 *) (m + 36)
+# asm 1: xorl 36(<m=int32#5),<in9=int32#2
+# asm 2: xorl 36(<m=%esi),<in9=%ecx
+xorl 36(%esi),%ecx
+
+# qhasm:   in10 ^= *(uint32 *) (m + 40)
+# asm 1: xorl 40(<m=int32#5),<in10=int32#3
+# asm 2: xorl 40(<m=%esi),<in10=%edx
+xorl 40(%esi),%edx
+
+# qhasm:   in11 ^= *(uint32 *) (m + 44)
+# asm 1: xorl 44(<m=int32#5),<in11=int32#4
+# asm 2: xorl 44(<m=%esi),<in11=%ebx
+xorl 44(%esi),%ebx
+
+# qhasm:   *(uint32 *) (out + 32) = in8
+# asm 1: movl <in8=int32#1,32(<out=int32#6)
+# asm 2: movl <in8=%eax,32(<out=%edi)
+movl %eax,32(%edi)
+
+# qhasm:   *(uint32 *) (out + 36) = in9
+# asm 1: movl <in9=int32#2,36(<out=int32#6)
+# asm 2: movl <in9=%ecx,36(<out=%edi)
+movl %ecx,36(%edi)
+
+# qhasm:   *(uint32 *) (out + 40) = in10
+# asm 1: movl <in10=int32#3,40(<out=int32#6)
+# asm 2: movl <in10=%edx,40(<out=%edi)
+movl %edx,40(%edi)
+
+# qhasm:   *(uint32 *) (out + 44) = in11
+# asm 1: movl <in11=int32#4,44(<out=int32#6)
+# asm 2: movl <in11=%ebx,44(<out=%edi)
+movl %ebx,44(%edi)
+
+# qhasm:   in8 = z8
+# asm 1: movd   <z8=int6464#1,>in8=int32#1
+# asm 2: movd   <z8=%xmm0,>in8=%eax
+movd   %xmm0,%eax
+
+# qhasm:   in9 = z9
+# asm 1: movd   <z9=int6464#2,>in9=int32#2
+# asm 2: movd   <z9=%xmm1,>in9=%ecx
+movd   %xmm1,%ecx
+
+# qhasm:   in10 = z10
+# asm 1: movd   <z10=int6464#3,>in10=int32#3
+# asm 2: movd   <z10=%xmm2,>in10=%edx
+movd   %xmm2,%edx
+
+# qhasm:   in11 = z11
+# asm 1: movd   <z11=int6464#4,>in11=int32#4
+# asm 2: movd   <z11=%xmm3,>in11=%ebx
+movd   %xmm3,%ebx
+
+# qhasm:   z8 <<<= 96
+# asm 1: pshufd $0x39,<z8=int6464#1,<z8=int6464#1
+# asm 2: pshufd $0x39,<z8=%xmm0,<z8=%xmm0
+pshufd $0x39,%xmm0,%xmm0
+
+# qhasm:   z9 <<<= 96
+# asm 1: pshufd $0x39,<z9=int6464#2,<z9=int6464#2
+# asm 2: pshufd $0x39,<z9=%xmm1,<z9=%xmm1
+pshufd $0x39,%xmm1,%xmm1
+
+# qhasm:   z10 <<<= 96
+# asm 1: pshufd $0x39,<z10=int6464#3,<z10=int6464#3
+# asm 2: pshufd $0x39,<z10=%xmm2,<z10=%xmm2
+pshufd $0x39,%xmm2,%xmm2
+
+# qhasm:   z11 <<<= 96
+# asm 1: pshufd $0x39,<z11=int6464#4,<z11=int6464#4
+# asm 2: pshufd $0x39,<z11=%xmm3,<z11=%xmm3
+pshufd $0x39,%xmm3,%xmm3
+
+# qhasm:   in8 ^= *(uint32 *) (m + 96)
+# asm 1: xorl 96(<m=int32#5),<in8=int32#1
+# asm 2: xorl 96(<m=%esi),<in8=%eax
+xorl 96(%esi),%eax
+
+# qhasm:   in9 ^= *(uint32 *) (m + 100)
+# asm 1: xorl 100(<m=int32#5),<in9=int32#2
+# asm 2: xorl 100(<m=%esi),<in9=%ecx
+xorl 100(%esi),%ecx
+
+# qhasm:   in10 ^= *(uint32 *) (m + 104)
+# asm 1: xorl 104(<m=int32#5),<in10=int32#3
+# asm 2: xorl 104(<m=%esi),<in10=%edx
+xorl 104(%esi),%edx
+
+# qhasm:   in11 ^= *(uint32 *) (m + 108)
+# asm 1: xorl 108(<m=int32#5),<in11=int32#4
+# asm 2: xorl 108(<m=%esi),<in11=%ebx
+xorl 108(%esi),%ebx
+
+# qhasm:   *(uint32 *) (out + 96) = in8
+# asm 1: movl <in8=int32#1,96(<out=int32#6)
+# asm 2: movl <in8=%eax,96(<out=%edi)
+movl %eax,96(%edi)
+
+# qhasm:   *(uint32 *) (out + 100) = in9
+# asm 1: movl <in9=int32#2,100(<out=int32#6)
+# asm 2: movl <in9=%ecx,100(<out=%edi)
+movl %ecx,100(%edi)
+
+# qhasm:   *(uint32 *) (out + 104) = in10
+# asm 1: movl <in10=int32#3,104(<out=int32#6)
+# asm 2: movl <in10=%edx,104(<out=%edi)
+movl %edx,104(%edi)
+
+# qhasm:   *(uint32 *) (out + 108) = in11
+# asm 1: movl <in11=int32#4,108(<out=int32#6)
+# asm 2: movl <in11=%ebx,108(<out=%edi)
+movl %ebx,108(%edi)
+
+# qhasm:   in8 = z8
+# asm 1: movd   <z8=int6464#1,>in8=int32#1
+# asm 2: movd   <z8=%xmm0,>in8=%eax
+movd   %xmm0,%eax
+
+# qhasm:   in9 = z9
+# asm 1: movd   <z9=int6464#2,>in9=int32#2
+# asm 2: movd   <z9=%xmm1,>in9=%ecx
+movd   %xmm1,%ecx
+
+# qhasm:   in10 = z10
+# asm 1: movd   <z10=int6464#3,>in10=int32#3
+# asm 2: movd   <z10=%xmm2,>in10=%edx
+movd   %xmm2,%edx
+
+# qhasm:   in11 = z11
+# asm 1: movd   <z11=int6464#4,>in11=int32#4
+# asm 2: movd   <z11=%xmm3,>in11=%ebx
+movd   %xmm3,%ebx
+
+# qhasm:   z8 <<<= 96
+# asm 1: pshufd $0x39,<z8=int6464#1,<z8=int6464#1
+# asm 2: pshufd $0x39,<z8=%xmm0,<z8=%xmm0
+pshufd $0x39,%xmm0,%xmm0
+
+# qhasm:   z9 <<<= 96
+# asm 1: pshufd $0x39,<z9=int6464#2,<z9=int6464#2
+# asm 2: pshufd $0x39,<z9=%xmm1,<z9=%xmm1
+pshufd $0x39,%xmm1,%xmm1
+
+# qhasm:   z10 <<<= 96
+# asm 1: pshufd $0x39,<z10=int6464#3,<z10=int6464#3
+# asm 2: pshufd $0x39,<z10=%xmm2,<z10=%xmm2
+pshufd $0x39,%xmm2,%xmm2
+
+# qhasm:   z11 <<<= 96
+# asm 1: pshufd $0x39,<z11=int6464#4,<z11=int6464#4
+# asm 2: pshufd $0x39,<z11=%xmm3,<z11=%xmm3
+pshufd $0x39,%xmm3,%xmm3
+
+# qhasm:   in8 ^= *(uint32 *) (m + 160)
+# asm 1: xorl 160(<m=int32#5),<in8=int32#1
+# asm 2: xorl 160(<m=%esi),<in8=%eax
+xorl 160(%esi),%eax
+
+# qhasm:   in9 ^= *(uint32 *) (m + 164)
+# asm 1: xorl 164(<m=int32#5),<in9=int32#2
+# asm 2: xorl 164(<m=%esi),<in9=%ecx
+xorl 164(%esi),%ecx
+
+# qhasm:   in10 ^= *(uint32 *) (m + 168)
+# asm 1: xorl 168(<m=int32#5),<in10=int32#3
+# asm 2: xorl 168(<m=%esi),<in10=%edx
+xorl 168(%esi),%edx
+
+# qhasm:   in11 ^= *(uint32 *) (m + 172)
+# asm 1: xorl 172(<m=int32#5),<in11=int32#4
+# asm 2: xorl 172(<m=%esi),<in11=%ebx
+xorl 172(%esi),%ebx
+
+# qhasm:   *(uint32 *) (out + 160) = in8
+# asm 1: movl <in8=int32#1,160(<out=int32#6)
+# asm 2: movl <in8=%eax,160(<out=%edi)
+movl %eax,160(%edi)
+
+# qhasm:   *(uint32 *) (out + 164) = in9
+# asm 1: movl <in9=int32#2,164(<out=int32#6)
+# asm 2: movl <in9=%ecx,164(<out=%edi)
+movl %ecx,164(%edi)
+
+# qhasm:   *(uint32 *) (out + 168) = in10
+# asm 1: movl <in10=int32#3,168(<out=int32#6)
+# asm 2: movl <in10=%edx,168(<out=%edi)
+movl %edx,168(%edi)
+
+# qhasm:   *(uint32 *) (out + 172) = in11
+# asm 1: movl <in11=int32#4,172(<out=int32#6)
+# asm 2: movl <in11=%ebx,172(<out=%edi)
+movl %ebx,172(%edi)
+
+# qhasm:   in8 = z8
+# asm 1: movd   <z8=int6464#1,>in8=int32#1
+# asm 2: movd   <z8=%xmm0,>in8=%eax
+movd   %xmm0,%eax
+
+# qhasm:   in9 = z9
+# asm 1: movd   <z9=int6464#2,>in9=int32#2
+# asm 2: movd   <z9=%xmm1,>in9=%ecx
+movd   %xmm1,%ecx
+
+# qhasm:   in10 = z10
+# asm 1: movd   <z10=int6464#3,>in10=int32#3
+# asm 2: movd   <z10=%xmm2,>in10=%edx
+movd   %xmm2,%edx
+
+# qhasm:   in11 = z11
+# asm 1: movd   <z11=int6464#4,>in11=int32#4
+# asm 2: movd   <z11=%xmm3,>in11=%ebx
+movd   %xmm3,%ebx
+
+# qhasm:   in8 ^= *(uint32 *) (m + 224)
+# asm 1: xorl 224(<m=int32#5),<in8=int32#1
+# asm 2: xorl 224(<m=%esi),<in8=%eax
+xorl 224(%esi),%eax
+
+# qhasm:   in9 ^= *(uint32 *) (m + 228)
+# asm 1: xorl 228(<m=int32#5),<in9=int32#2
+# asm 2: xorl 228(<m=%esi),<in9=%ecx
+xorl 228(%esi),%ecx
+
+# qhasm:   in10 ^= *(uint32 *) (m + 232)
+# asm 1: xorl 232(<m=int32#5),<in10=int32#3
+# asm 2: xorl 232(<m=%esi),<in10=%edx
+xorl 232(%esi),%edx
+
+# qhasm:   in11 ^= *(uint32 *) (m + 236)
+# asm 1: xorl 236(<m=int32#5),<in11=int32#4
+# asm 2: xorl 236(<m=%esi),<in11=%ebx
+xorl 236(%esi),%ebx
+
+# qhasm:   *(uint32 *) (out + 224) = in8
+# asm 1: movl <in8=int32#1,224(<out=int32#6)
+# asm 2: movl <in8=%eax,224(<out=%edi)
+movl %eax,224(%edi)
+
+# qhasm:   *(uint32 *) (out + 228) = in9
+# asm 1: movl <in9=int32#2,228(<out=int32#6)
+# asm 2: movl <in9=%ecx,228(<out=%edi)
+movl %ecx,228(%edi)
+
+# qhasm:   *(uint32 *) (out + 232) = in10
+# asm 1: movl <in10=int32#3,232(<out=int32#6)
+# asm 2: movl <in10=%edx,232(<out=%edi)
+movl %edx,232(%edi)
+
+# qhasm:   *(uint32 *) (out + 236) = in11
+# asm 1: movl <in11=int32#4,236(<out=int32#6)
+# asm 2: movl <in11=%ebx,236(<out=%edi)
+movl %ebx,236(%edi)
+
+# qhasm:   z12 = z12_stack
+# asm 1: movdqa <z12_stack=stack128#35,>z12=int6464#1
+# asm 2: movdqa <z12_stack=576(%esp),>z12=%xmm0
+movdqa 576(%esp),%xmm0
+
+# qhasm:   z13 = z13_stack
+# asm 1: movdqa <z13_stack=stack128#30,>z13=int6464#2
+# asm 2: movdqa <z13_stack=496(%esp),>z13=%xmm1
+movdqa 496(%esp),%xmm1
+
+# qhasm:   z14 = z14_stack
+# asm 1: movdqa <z14_stack=stack128#24,>z14=int6464#3
+# asm 2: movdqa <z14_stack=400(%esp),>z14=%xmm2
+movdqa 400(%esp),%xmm2
+
+# qhasm:   z15 = z15_stack
+# asm 1: movdqa <z15_stack=stack128#23,>z15=int6464#4
+# asm 2: movdqa <z15_stack=384(%esp),>z15=%xmm3
+movdqa 384(%esp),%xmm3
+
+# qhasm:   uint32323232 z12 += orig12
+# asm 1: paddd <orig12=stack128#11,<z12=int6464#1
+# asm 2: paddd <orig12=192(%esp),<z12=%xmm0
+paddd 192(%esp),%xmm0
+
+# qhasm:   uint32323232 z13 += orig13
+# asm 1: paddd <orig13=stack128#14,<z13=int6464#2
+# asm 2: paddd <orig13=240(%esp),<z13=%xmm1
+paddd 240(%esp),%xmm1
+
+# qhasm:   uint32323232 z14 += orig14
+# asm 1: paddd <orig14=stack128#17,<z14=int6464#3
+# asm 2: paddd <orig14=288(%esp),<z14=%xmm2
+paddd 288(%esp),%xmm2
+
+# qhasm:   uint32323232 z15 += orig15
+# asm 1: paddd <orig15=stack128#7,<z15=int6464#4
+# asm 2: paddd <orig15=128(%esp),<z15=%xmm3
+paddd 128(%esp),%xmm3
+
+# qhasm:   in12 = z12
+# asm 1: movd   <z12=int6464#1,>in12=int32#1
+# asm 2: movd   <z12=%xmm0,>in12=%eax
+movd   %xmm0,%eax
+
+# qhasm:   in13 = z13
+# asm 1: movd   <z13=int6464#2,>in13=int32#2
+# asm 2: movd   <z13=%xmm1,>in13=%ecx
+movd   %xmm1,%ecx
+
+# qhasm:   in14 = z14
+# asm 1: movd   <z14=int6464#3,>in14=int32#3
+# asm 2: movd   <z14=%xmm2,>in14=%edx
+movd   %xmm2,%edx
+
+# qhasm:   in15 = z15
+# asm 1: movd   <z15=int6464#4,>in15=int32#4
+# asm 2: movd   <z15=%xmm3,>in15=%ebx
+movd   %xmm3,%ebx
+
+# qhasm:   z12 <<<= 96
+# asm 1: pshufd $0x39,<z12=int6464#1,<z12=int6464#1
+# asm 2: pshufd $0x39,<z12=%xmm0,<z12=%xmm0
+pshufd $0x39,%xmm0,%xmm0
+
+# qhasm:   z13 <<<= 96
+# asm 1: pshufd $0x39,<z13=int6464#2,<z13=int6464#2
+# asm 2: pshufd $0x39,<z13=%xmm1,<z13=%xmm1
+pshufd $0x39,%xmm1,%xmm1
+
+# qhasm:   z14 <<<= 96
+# asm 1: pshufd $0x39,<z14=int6464#3,<z14=int6464#3
+# asm 2: pshufd $0x39,<z14=%xmm2,<z14=%xmm2
+pshufd $0x39,%xmm2,%xmm2
+
+# qhasm:   z15 <<<= 96
+# asm 1: pshufd $0x39,<z15=int6464#4,<z15=int6464#4
+# asm 2: pshufd $0x39,<z15=%xmm3,<z15=%xmm3
+pshufd $0x39,%xmm3,%xmm3
+
+# qhasm:   in12 ^= *(uint32 *) (m + 48)
+# asm 1: xorl 48(<m=int32#5),<in12=int32#1
+# asm 2: xorl 48(<m=%esi),<in12=%eax
+xorl 48(%esi),%eax
+
+# qhasm:   in13 ^= *(uint32 *) (m + 52)
+# asm 1: xorl 52(<m=int32#5),<in13=int32#2
+# asm 2: xorl 52(<m=%esi),<in13=%ecx
+xorl 52(%esi),%ecx
+
+# qhasm:   in14 ^= *(uint32 *) (m + 56)
+# asm 1: xorl 56(<m=int32#5),<in14=int32#3
+# asm 2: xorl 56(<m=%esi),<in14=%edx
+xorl 56(%esi),%edx
+
+# qhasm:   in15 ^= *(uint32 *) (m + 60)
+# asm 1: xorl 60(<m=int32#5),<in15=int32#4
+# asm 2: xorl 60(<m=%esi),<in15=%ebx
+xorl 60(%esi),%ebx
+
+# qhasm:   *(uint32 *) (out + 48) = in12
+# asm 1: movl <in12=int32#1,48(<out=int32#6)
+# asm 2: movl <in12=%eax,48(<out=%edi)
+movl %eax,48(%edi)
+
+# qhasm:   *(uint32 *) (out + 52) = in13
+# asm 1: movl <in13=int32#2,52(<out=int32#6)
+# asm 2: movl <in13=%ecx,52(<out=%edi)
+movl %ecx,52(%edi)
+
+# qhasm:   *(uint32 *) (out + 56) = in14
+# asm 1: movl <in14=int32#3,56(<out=int32#6)
+# asm 2: movl <in14=%edx,56(<out=%edi)
+movl %edx,56(%edi)
+
+# qhasm:   *(uint32 *) (out + 60) = in15
+# asm 1: movl <in15=int32#4,60(<out=int32#6)
+# asm 2: movl <in15=%ebx,60(<out=%edi)
+movl %ebx,60(%edi)
+
+# qhasm:   in12 = z12
+# asm 1: movd   <z12=int6464#1,>in12=int32#1
+# asm 2: movd   <z12=%xmm0,>in12=%eax
+movd   %xmm0,%eax
+
+# qhasm:   in13 = z13
+# asm 1: movd   <z13=int6464#2,>in13=int32#2
+# asm 2: movd   <z13=%xmm1,>in13=%ecx
+movd   %xmm1,%ecx
+
+# qhasm:   in14 = z14
+# asm 1: movd   <z14=int6464#3,>in14=int32#3
+# asm 2: movd   <z14=%xmm2,>in14=%edx
+movd   %xmm2,%edx
+
+# qhasm:   in15 = z15
+# asm 1: movd   <z15=int6464#4,>in15=int32#4
+# asm 2: movd   <z15=%xmm3,>in15=%ebx
+movd   %xmm3,%ebx
+
+# qhasm:   z12 <<<= 96
+# asm 1: pshufd $0x39,<z12=int6464#1,<z12=int6464#1
+# asm 2: pshufd $0x39,<z12=%xmm0,<z12=%xmm0
+pshufd $0x39,%xmm0,%xmm0
+
+# qhasm:   z13 <<<= 96
+# asm 1: pshufd $0x39,<z13=int6464#2,<z13=int6464#2
+# asm 2: pshufd $0x39,<z13=%xmm1,<z13=%xmm1
+pshufd $0x39,%xmm1,%xmm1
+
+# qhasm:   z14 <<<= 96
+# asm 1: pshufd $0x39,<z14=int6464#3,<z14=int6464#3
+# asm 2: pshufd $0x39,<z14=%xmm2,<z14=%xmm2
+pshufd $0x39,%xmm2,%xmm2
+
+# qhasm:   z15 <<<= 96
+# asm 1: pshufd $0x39,<z15=int6464#4,<z15=int6464#4
+# asm 2: pshufd $0x39,<z15=%xmm3,<z15=%xmm3
+pshufd $0x39,%xmm3,%xmm3
+
+# qhasm:   in12 ^= *(uint32 *) (m + 112)
+# asm 1: xorl 112(<m=int32#5),<in12=int32#1
+# asm 2: xorl 112(<m=%esi),<in12=%eax
+xorl 112(%esi),%eax
+
+# qhasm:   in13 ^= *(uint32 *) (m + 116)
+# asm 1: xorl 116(<m=int32#5),<in13=int32#2
+# asm 2: xorl 116(<m=%esi),<in13=%ecx
+xorl 116(%esi),%ecx
+
+# qhasm:   in14 ^= *(uint32 *) (m + 120)
+# asm 1: xorl 120(<m=int32#5),<in14=int32#3
+# asm 2: xorl 120(<m=%esi),<in14=%edx
+xorl 120(%esi),%edx
+
+# qhasm:   in15 ^= *(uint32 *) (m + 124)
+# asm 1: xorl 124(<m=int32#5),<in15=int32#4
+# asm 2: xorl 124(<m=%esi),<in15=%ebx
+xorl 124(%esi),%ebx
+
+# qhasm:   *(uint32 *) (out + 112) = in12
+# asm 1: movl <in12=int32#1,112(<out=int32#6)
+# asm 2: movl <in12=%eax,112(<out=%edi)
+movl %eax,112(%edi)
+
+# qhasm:   *(uint32 *) (out + 116) = in13
+# asm 1: movl <in13=int32#2,116(<out=int32#6)
+# asm 2: movl <in13=%ecx,116(<out=%edi)
+movl %ecx,116(%edi)
+
+# qhasm:   *(uint32 *) (out + 120) = in14
+# asm 1: movl <in14=int32#3,120(<out=int32#6)
+# asm 2: movl <in14=%edx,120(<out=%edi)
+movl %edx,120(%edi)
+
+# qhasm:   *(uint32 *) (out + 124) = in15
+# asm 1: movl <in15=int32#4,124(<out=int32#6)
+# asm 2: movl <in15=%ebx,124(<out=%edi)
+movl %ebx,124(%edi)
+
+# qhasm:   in12 = z12
+# asm 1: movd   <z12=int6464#1,>in12=int32#1
+# asm 2: movd   <z12=%xmm0,>in12=%eax
+movd   %xmm0,%eax
+
+# qhasm:   in13 = z13
+# asm 1: movd   <z13=int6464#2,>in13=int32#2
+# asm 2: movd   <z13=%xmm1,>in13=%ecx
+movd   %xmm1,%ecx
+
+# qhasm:   in14 = z14
+# asm 1: movd   <z14=int6464#3,>in14=int32#3
+# asm 2: movd   <z14=%xmm2,>in14=%edx
+movd   %xmm2,%edx
+
+# qhasm:   in15 = z15
+# asm 1: movd   <z15=int6464#4,>in15=int32#4
+# asm 2: movd   <z15=%xmm3,>in15=%ebx
+movd   %xmm3,%ebx
+
+# qhasm:   z12 <<<= 96
+# asm 1: pshufd $0x39,<z12=int6464#1,<z12=int6464#1
+# asm 2: pshufd $0x39,<z12=%xmm0,<z12=%xmm0
+pshufd $0x39,%xmm0,%xmm0
+
+# qhasm:   z13 <<<= 96
+# asm 1: pshufd $0x39,<z13=int6464#2,<z13=int6464#2
+# asm 2: pshufd $0x39,<z13=%xmm1,<z13=%xmm1
+pshufd $0x39,%xmm1,%xmm1
+
+# qhasm:   z14 <<<= 96
+# asm 1: pshufd $0x39,<z14=int6464#3,<z14=int6464#3
+# asm 2: pshufd $0x39,<z14=%xmm2,<z14=%xmm2
+pshufd $0x39,%xmm2,%xmm2
+
+# qhasm:   z15 <<<= 96
+# asm 1: pshufd $0x39,<z15=int6464#4,<z15=int6464#4
+# asm 2: pshufd $0x39,<z15=%xmm3,<z15=%xmm3
+pshufd $0x39,%xmm3,%xmm3
+
+# qhasm:   in12 ^= *(uint32 *) (m + 176)
+# asm 1: xorl 176(<m=int32#5),<in12=int32#1
+# asm 2: xorl 176(<m=%esi),<in12=%eax
+xorl 176(%esi),%eax
+
+# qhasm:   in13 ^= *(uint32 *) (m + 180)
+# asm 1: xorl 180(<m=int32#5),<in13=int32#2
+# asm 2: xorl 180(<m=%esi),<in13=%ecx
+xorl 180(%esi),%ecx
+
+# qhasm:   in14 ^= *(uint32 *) (m + 184)
+# asm 1: xorl 184(<m=int32#5),<in14=int32#3
+# asm 2: xorl 184(<m=%esi),<in14=%edx
+xorl 184(%esi),%edx
+
+# qhasm:   in15 ^= *(uint32 *) (m + 188)
+# asm 1: xorl 188(<m=int32#5),<in15=int32#4
+# asm 2: xorl 188(<m=%esi),<in15=%ebx
+xorl 188(%esi),%ebx
+
+# qhasm:   *(uint32 *) (out + 176) = in12
+# asm 1: movl <in12=int32#1,176(<out=int32#6)
+# asm 2: movl <in12=%eax,176(<out=%edi)
+movl %eax,176(%edi)
+
+# qhasm:   *(uint32 *) (out + 180) = in13
+# asm 1: movl <in13=int32#2,180(<out=int32#6)
+# asm 2: movl <in13=%ecx,180(<out=%edi)
+movl %ecx,180(%edi)
+
+# qhasm:   *(uint32 *) (out + 184) = in14
+# asm 1: movl <in14=int32#3,184(<out=int32#6)
+# asm 2: movl <in14=%edx,184(<out=%edi)
+movl %edx,184(%edi)
+
+# qhasm:   *(uint32 *) (out + 188) = in15
+# asm 1: movl <in15=int32#4,188(<out=int32#6)
+# asm 2: movl <in15=%ebx,188(<out=%edi)
+movl %ebx,188(%edi)
+
+# qhasm:   in12 = z12
+# asm 1: movd   <z12=int6464#1,>in12=int32#1
+# asm 2: movd   <z12=%xmm0,>in12=%eax
+movd   %xmm0,%eax
+
+# qhasm:   in13 = z13
+# asm 1: movd   <z13=int6464#2,>in13=int32#2
+# asm 2: movd   <z13=%xmm1,>in13=%ecx
+movd   %xmm1,%ecx
+
+# qhasm:   in14 = z14
+# asm 1: movd   <z14=int6464#3,>in14=int32#3
+# asm 2: movd   <z14=%xmm2,>in14=%edx
+movd   %xmm2,%edx
+
+# qhasm:   in15 = z15
+# asm 1: movd   <z15=int6464#4,>in15=int32#4
+# asm 2: movd   <z15=%xmm3,>in15=%ebx
+movd   %xmm3,%ebx
+
+# qhasm:   in12 ^= *(uint32 *) (m + 240)
+# asm 1: xorl 240(<m=int32#5),<in12=int32#1
+# asm 2: xorl 240(<m=%esi),<in12=%eax
+xorl 240(%esi),%eax
+
+# qhasm:   in13 ^= *(uint32 *) (m + 244)
+# asm 1: xorl 244(<m=int32#5),<in13=int32#2
+# asm 2: xorl 244(<m=%esi),<in13=%ecx
+xorl 244(%esi),%ecx
+
+# qhasm:   in14 ^= *(uint32 *) (m + 248)
+# asm 1: xorl 248(<m=int32#5),<in14=int32#3
+# asm 2: xorl 248(<m=%esi),<in14=%edx
+xorl 248(%esi),%edx
+
+# qhasm:   in15 ^= *(uint32 *) (m + 252)
+# asm 1: xorl 252(<m=int32#5),<in15=int32#4
+# asm 2: xorl 252(<m=%esi),<in15=%ebx
+xorl 252(%esi),%ebx
+
+# qhasm:   *(uint32 *) (out + 240) = in12
+# asm 1: movl <in12=int32#1,240(<out=int32#6)
+# asm 2: movl <in12=%eax,240(<out=%edi)
+movl %eax,240(%edi)
+
+# qhasm:   *(uint32 *) (out + 244) = in13
+# asm 1: movl <in13=int32#2,244(<out=int32#6)
+# asm 2: movl <in13=%ecx,244(<out=%edi)
+movl %ecx,244(%edi)
+
+# qhasm:   *(uint32 *) (out + 248) = in14
+# asm 1: movl <in14=int32#3,248(<out=int32#6)
+# asm 2: movl <in14=%edx,248(<out=%edi)
+movl %edx,248(%edi)
+
+# qhasm:   *(uint32 *) (out + 252) = in15
+# asm 1: movl <in15=int32#4,252(<out=int32#6)
+# asm 2: movl <in15=%ebx,252(<out=%edi)
+movl %ebx,252(%edi)
+
+# qhasm:   bytes = bytes_stack
+# asm 1: movl <bytes_stack=stack32#7,>bytes=int32#1
+# asm 2: movl <bytes_stack=24(%esp),>bytes=%eax
+movl 24(%esp),%eax
+
+# qhasm:   bytes -= 256
+# asm 1: sub  $256,<bytes=int32#1
+# asm 2: sub  $256,<bytes=%eax
+sub  $256,%eax
+
+# qhasm:   m += 256
+# asm 1: add  $256,<m=int32#5
+# asm 2: add  $256,<m=%esi
+add  $256,%esi
+
+# qhasm:   out += 256
+# asm 1: add  $256,<out=int32#6
+# asm 2: add  $256,<out=%edi
+add  $256,%edi
+
+# qhasm:   out_stack = out
+# asm 1: movl <out=int32#6,>out_stack=stack32#6
+# asm 2: movl <out=%edi,>out_stack=20(%esp)
+movl %edi,20(%esp)
+
+# qhasm:                            unsigned<? bytes - 256
+# asm 1: cmp  $256,<bytes=int32#1
+# asm 2: cmp  $256,<bytes=%eax
+cmp  $256,%eax
+# comment:fp stack unchanged by jump
+
+# qhasm:   goto bytesatleast256 if !unsigned<
+jae ._bytesatleast256
+
+# qhasm:                 unsigned>? bytes - 0
+# asm 1: cmp  $0,<bytes=int32#1
+# asm 2: cmp  $0,<bytes=%eax
+cmp  $0,%eax
+# comment:fp stack unchanged by jump
+
+# qhasm:   goto done if !unsigned>
+jbe ._done
+# comment:fp stack unchanged by fallthrough
+
+# qhasm: bytesbetween1and255:
+._bytesbetween1and255:
+
+# qhasm:                   unsigned<? bytes - 64
+# asm 1: cmp  $64,<bytes=int32#1
+# asm 2: cmp  $64,<bytes=%eax
+cmp  $64,%eax
+# comment:fp stack unchanged by jump
+
+# qhasm:   goto nocopy if !unsigned<
+jae ._nocopy
+
+# qhasm:     ctarget = out
+# asm 1: movl <out=int32#6,>ctarget=stack32#6
+# asm 2: movl <out=%edi,>ctarget=20(%esp)
+movl %edi,20(%esp)
+
+# qhasm:     out = &tmp
+# asm 1: leal <tmp=stack512#1,>out=int32#6
+# asm 2: leal <tmp=640(%esp),>out=%edi
+leal 640(%esp),%edi
+
+# qhasm:     i = bytes
+# asm 1: mov  <bytes=int32#1,>i=int32#2
+# asm 2: mov  <bytes=%eax,>i=%ecx
+mov  %eax,%ecx
+
+# qhasm:     while (i) { *out++ = *m++; --i }
+rep movsb
+
+# qhasm:     out = &tmp
+# asm 1: leal <tmp=stack512#1,>out=int32#6
+# asm 2: leal <tmp=640(%esp),>out=%edi
+leal 640(%esp),%edi
+
+# qhasm:     m = &tmp
+# asm 1: leal <tmp=stack512#1,>m=int32#5
+# asm 2: leal <tmp=640(%esp),>m=%esi
+leal 640(%esp),%esi
+# comment:fp stack unchanged by fallthrough
+
+# qhasm:   nocopy:
+._nocopy:
+
+# qhasm:   bytes_stack = bytes
+# asm 1: movl <bytes=int32#1,>bytes_stack=stack32#7
+# asm 2: movl <bytes=%eax,>bytes_stack=24(%esp)
+movl %eax,24(%esp)
+
+# qhasm: diag0 = x0
+# asm 1: movdqa <x0=stack128#3,>diag0=int6464#1
+# asm 2: movdqa <x0=64(%esp),>diag0=%xmm0
+movdqa 64(%esp),%xmm0
+
+# qhasm: diag1 = x1
+# asm 1: movdqa <x1=stack128#2,>diag1=int6464#2
+# asm 2: movdqa <x1=48(%esp),>diag1=%xmm1
+movdqa 48(%esp),%xmm1
+
+# qhasm: diag2 = x2
+# asm 1: movdqa <x2=stack128#4,>diag2=int6464#3
+# asm 2: movdqa <x2=80(%esp),>diag2=%xmm2
+movdqa 80(%esp),%xmm2
+
+# qhasm: diag3 = x3
+# asm 1: movdqa <x3=stack128#1,>diag3=int6464#4
+# asm 2: movdqa <x3=32(%esp),>diag3=%xmm3
+movdqa 32(%esp),%xmm3
+
+# qhasm:                     a0 = diag1
+# asm 1: movdqa <diag1=int6464#2,>a0=int6464#5
+# asm 2: movdqa <diag1=%xmm1,>a0=%xmm4
+movdqa %xmm1,%xmm4
+
+# qhasm: i = 12
+# asm 1: mov  $12,>i=int32#1
+# asm 2: mov  $12,>i=%eax
+mov  $12,%eax
+
+# qhasm: mainloop2:
+._mainloop2:
+
+# qhasm: uint32323232        a0 += diag0
+# asm 1: paddd <diag0=int6464#1,<a0=int6464#5
+# asm 2: paddd <diag0=%xmm0,<a0=%xmm4
+paddd %xmm0,%xmm4
+
+# qhasm:                                 a1 = diag0
+# asm 1: movdqa <diag0=int6464#1,>a1=int6464#6
+# asm 2: movdqa <diag0=%xmm0,>a1=%xmm5
+movdqa %xmm0,%xmm5
+
+# qhasm:                     b0 = a0
+# asm 1: movdqa <a0=int6464#5,>b0=int6464#7
+# asm 2: movdqa <a0=%xmm4,>b0=%xmm6
+movdqa %xmm4,%xmm6
+
+# qhasm: uint32323232        a0 <<= 7
+# asm 1: pslld $7,<a0=int6464#5
+# asm 2: pslld $7,<a0=%xmm4
+pslld $7,%xmm4
+
+# qhasm: uint32323232        b0 >>= 25
+# asm 1: psrld $25,<b0=int6464#7
+# asm 2: psrld $25,<b0=%xmm6
+psrld $25,%xmm6
+
+# qhasm:                 diag3 ^= a0
+# asm 1: pxor  <a0=int6464#5,<diag3=int6464#4
+# asm 2: pxor  <a0=%xmm4,<diag3=%xmm3
+pxor  %xmm4,%xmm3
+
+# qhasm:                 diag3 ^= b0
+# asm 1: pxor  <b0=int6464#7,<diag3=int6464#4
+# asm 2: pxor  <b0=%xmm6,<diag3=%xmm3
+pxor  %xmm6,%xmm3
+
+# qhasm: uint32323232                        a1 += diag3
+# asm 1: paddd <diag3=int6464#4,<a1=int6464#6
+# asm 2: paddd <diag3=%xmm3,<a1=%xmm5
+paddd %xmm3,%xmm5
+
+# qhasm:                                                 a2 = diag3
+# asm 1: movdqa <diag3=int6464#4,>a2=int6464#5
+# asm 2: movdqa <diag3=%xmm3,>a2=%xmm4
+movdqa %xmm3,%xmm4
+
+# qhasm:                                     b1 = a1
+# asm 1: movdqa <a1=int6464#6,>b1=int6464#7
+# asm 2: movdqa <a1=%xmm5,>b1=%xmm6
+movdqa %xmm5,%xmm6
+
+# qhasm: uint32323232                        a1 <<= 9
+# asm 1: pslld $9,<a1=int6464#6
+# asm 2: pslld $9,<a1=%xmm5
+pslld $9,%xmm5
+
+# qhasm: uint32323232                        b1 >>= 23
+# asm 1: psrld $23,<b1=int6464#7
+# asm 2: psrld $23,<b1=%xmm6
+psrld $23,%xmm6
+
+# qhasm:                                 diag2 ^= a1
+# asm 1: pxor  <a1=int6464#6,<diag2=int6464#3
+# asm 2: pxor  <a1=%xmm5,<diag2=%xmm2
+pxor  %xmm5,%xmm2
+
+# qhasm:                 diag3 <<<= 32
+# asm 1: pshufd $0x93,<diag3=int6464#4,<diag3=int6464#4
+# asm 2: pshufd $0x93,<diag3=%xmm3,<diag3=%xmm3
+pshufd $0x93,%xmm3,%xmm3
+
+# qhasm:                                 diag2 ^= b1
+# asm 1: pxor  <b1=int6464#7,<diag2=int6464#3
+# asm 2: pxor  <b1=%xmm6,<diag2=%xmm2
+pxor  %xmm6,%xmm2
+
+# qhasm: uint32323232                                        a2 += diag2
+# asm 1: paddd <diag2=int6464#3,<a2=int6464#5
+# asm 2: paddd <diag2=%xmm2,<a2=%xmm4
+paddd %xmm2,%xmm4
+
+# qhasm:                                                                 a3 = diag2
+# asm 1: movdqa <diag2=int6464#3,>a3=int6464#6
+# asm 2: movdqa <diag2=%xmm2,>a3=%xmm5
+movdqa %xmm2,%xmm5
+
+# qhasm:                                                     b2 = a2
+# asm 1: movdqa <a2=int6464#5,>b2=int6464#7
+# asm 2: movdqa <a2=%xmm4,>b2=%xmm6
+movdqa %xmm4,%xmm6
+
+# qhasm: uint32323232                                        a2 <<= 13
+# asm 1: pslld $13,<a2=int6464#5
+# asm 2: pslld $13,<a2=%xmm4
+pslld $13,%xmm4
+
+# qhasm: uint32323232                                        b2 >>= 19
+# asm 1: psrld $19,<b2=int6464#7
+# asm 2: psrld $19,<b2=%xmm6
+psrld $19,%xmm6
+
+# qhasm:                                                 diag1 ^= a2
+# asm 1: pxor  <a2=int6464#5,<diag1=int6464#2
+# asm 2: pxor  <a2=%xmm4,<diag1=%xmm1
+pxor  %xmm4,%xmm1
+
+# qhasm:                                 diag2 <<<= 64
+# asm 1: pshufd $0x4e,<diag2=int6464#3,<diag2=int6464#3
+# asm 2: pshufd $0x4e,<diag2=%xmm2,<diag2=%xmm2
+pshufd $0x4e,%xmm2,%xmm2
+
+# qhasm:                                                 diag1 ^= b2
+# asm 1: pxor  <b2=int6464#7,<diag1=int6464#2
+# asm 2: pxor  <b2=%xmm6,<diag1=%xmm1
+pxor  %xmm6,%xmm1
+
+# qhasm: uint32323232                                                        a3 += diag1
+# asm 1: paddd <diag1=int6464#2,<a3=int6464#6
+# asm 2: paddd <diag1=%xmm1,<a3=%xmm5
+paddd %xmm1,%xmm5
+
+# qhasm:                 a4 = diag3
+# asm 1: movdqa <diag3=int6464#4,>a4=int6464#5
+# asm 2: movdqa <diag3=%xmm3,>a4=%xmm4
+movdqa %xmm3,%xmm4
+
+# qhasm:                                                                     b3 = a3
+# asm 1: movdqa <a3=int6464#6,>b3=int6464#7
+# asm 2: movdqa <a3=%xmm5,>b3=%xmm6
+movdqa %xmm5,%xmm6
+
+# qhasm: uint32323232                                                        a3 <<= 18
+# asm 1: pslld $18,<a3=int6464#6
+# asm 2: pslld $18,<a3=%xmm5
+pslld $18,%xmm5
+
+# qhasm: uint32323232                                                        b3 >>= 14
+# asm 1: psrld $14,<b3=int6464#7
+# asm 2: psrld $14,<b3=%xmm6
+psrld $14,%xmm6
+
+# qhasm:                                                                 diag0 ^= a3
+# asm 1: pxor  <a3=int6464#6,<diag0=int6464#1
+# asm 2: pxor  <a3=%xmm5,<diag0=%xmm0
+pxor  %xmm5,%xmm0
+
+# qhasm:                                                 diag1 <<<= 96
+# asm 1: pshufd $0x39,<diag1=int6464#2,<diag1=int6464#2
+# asm 2: pshufd $0x39,<diag1=%xmm1,<diag1=%xmm1
+pshufd $0x39,%xmm1,%xmm1
+
+# qhasm:                                                                 diag0 ^= b3
+# asm 1: pxor  <b3=int6464#7,<diag0=int6464#1
+# asm 2: pxor  <b3=%xmm6,<diag0=%xmm0
+pxor  %xmm6,%xmm0
+
+# qhasm: uint32323232        a4 += diag0
+# asm 1: paddd <diag0=int6464#1,<a4=int6464#5
+# asm 2: paddd <diag0=%xmm0,<a4=%xmm4
+paddd %xmm0,%xmm4
+
+# qhasm:                                 a5 = diag0
+# asm 1: movdqa <diag0=int6464#1,>a5=int6464#6
+# asm 2: movdqa <diag0=%xmm0,>a5=%xmm5
+movdqa %xmm0,%xmm5
+
+# qhasm:                     b4 = a4
+# asm 1: movdqa <a4=int6464#5,>b4=int6464#7
+# asm 2: movdqa <a4=%xmm4,>b4=%xmm6
+movdqa %xmm4,%xmm6
+
+# qhasm: uint32323232        a4 <<= 7
+# asm 1: pslld $7,<a4=int6464#5
+# asm 2: pslld $7,<a4=%xmm4
+pslld $7,%xmm4
+
+# qhasm: uint32323232        b4 >>= 25
+# asm 1: psrld $25,<b4=int6464#7
+# asm 2: psrld $25,<b4=%xmm6
+psrld $25,%xmm6
+
+# qhasm:                 diag1 ^= a4
+# asm 1: pxor  <a4=int6464#5,<diag1=int6464#2
+# asm 2: pxor  <a4=%xmm4,<diag1=%xmm1
+pxor  %xmm4,%xmm1
+
+# qhasm:                 diag1 ^= b4
+# asm 1: pxor  <b4=int6464#7,<diag1=int6464#2
+# asm 2: pxor  <b4=%xmm6,<diag1=%xmm1
+pxor  %xmm6,%xmm1
+
+# qhasm: uint32323232                        a5 += diag1
+# asm 1: paddd <diag1=int6464#2,<a5=int6464#6
+# asm 2: paddd <diag1=%xmm1,<a5=%xmm5
+paddd %xmm1,%xmm5
+
+# qhasm:                                                 a6 = diag1
+# asm 1: movdqa <diag1=int6464#2,>a6=int6464#5
+# asm 2: movdqa <diag1=%xmm1,>a6=%xmm4
+movdqa %xmm1,%xmm4
+
+# qhasm:                                     b5 = a5
+# asm 1: movdqa <a5=int6464#6,>b5=int6464#7
+# asm 2: movdqa <a5=%xmm5,>b5=%xmm6
+movdqa %xmm5,%xmm6
+
+# qhasm: uint32323232                        a5 <<= 9
+# asm 1: pslld $9,<a5=int6464#6
+# asm 2: pslld $9,<a5=%xmm5
+pslld $9,%xmm5
+
+# qhasm: uint32323232                        b5 >>= 23
+# asm 1: psrld $23,<b5=int6464#7
+# asm 2: psrld $23,<b5=%xmm6
+psrld $23,%xmm6
+
+# qhasm:                                 diag2 ^= a5
+# asm 1: pxor  <a5=int6464#6,<diag2=int6464#3
+# asm 2: pxor  <a5=%xmm5,<diag2=%xmm2
+pxor  %xmm5,%xmm2
+
+# qhasm:                 diag1 <<<= 32
+# asm 1: pshufd $0x93,<diag1=int6464#2,<diag1=int6464#2
+# asm 2: pshufd $0x93,<diag1=%xmm1,<diag1=%xmm1
+pshufd $0x93,%xmm1,%xmm1
+
+# qhasm:                                 diag2 ^= b5
+# asm 1: pxor  <b5=int6464#7,<diag2=int6464#3
+# asm 2: pxor  <b5=%xmm6,<diag2=%xmm2
+pxor  %xmm6,%xmm2
+
+# qhasm: uint32323232                                        a6 += diag2
+# asm 1: paddd <diag2=int6464#3,<a6=int6464#5
+# asm 2: paddd <diag2=%xmm2,<a6=%xmm4
+paddd %xmm2,%xmm4
+
+# qhasm:                                                                 a7 = diag2
+# asm 1: movdqa <diag2=int6464#3,>a7=int6464#6
+# asm 2: movdqa <diag2=%xmm2,>a7=%xmm5
+movdqa %xmm2,%xmm5
+
+# qhasm:                                                     b6 = a6
+# asm 1: movdqa <a6=int6464#5,>b6=int6464#7
+# asm 2: movdqa <a6=%xmm4,>b6=%xmm6
+movdqa %xmm4,%xmm6
+
+# qhasm: uint32323232                                        a6 <<= 13
+# asm 1: pslld $13,<a6=int6464#5
+# asm 2: pslld $13,<a6=%xmm4
+pslld $13,%xmm4
+
+# qhasm: uint32323232                                        b6 >>= 19
+# asm 1: psrld $19,<b6=int6464#7
+# asm 2: psrld $19,<b6=%xmm6
+psrld $19,%xmm6
+
+# qhasm:                                                 diag3 ^= a6
+# asm 1: pxor  <a6=int6464#5,<diag3=int6464#4
+# asm 2: pxor  <a6=%xmm4,<diag3=%xmm3
+pxor  %xmm4,%xmm3
+
+# qhasm:                                 diag2 <<<= 64
+# asm 1: pshufd $0x4e,<diag2=int6464#3,<diag2=int6464#3
+# asm 2: pshufd $0x4e,<diag2=%xmm2,<diag2=%xmm2
+pshufd $0x4e,%xmm2,%xmm2
+
+# qhasm:                                                 diag3 ^= b6
+# asm 1: pxor  <b6=int6464#7,<diag3=int6464#4
+# asm 2: pxor  <b6=%xmm6,<diag3=%xmm3
+pxor  %xmm6,%xmm3
+
+# qhasm: uint32323232                                                        a7 += diag3
+# asm 1: paddd <diag3=int6464#4,<a7=int6464#6
+# asm 2: paddd <diag3=%xmm3,<a7=%xmm5
+paddd %xmm3,%xmm5
+
+# qhasm:                 a0 = diag1
+# asm 1: movdqa <diag1=int6464#2,>a0=int6464#5
+# asm 2: movdqa <diag1=%xmm1,>a0=%xmm4
+movdqa %xmm1,%xmm4
+
+# qhasm:                                                                     b7 = a7
+# asm 1: movdqa <a7=int6464#6,>b7=int6464#7
+# asm 2: movdqa <a7=%xmm5,>b7=%xmm6
+movdqa %xmm5,%xmm6
+
+# qhasm: uint32323232                                                        a7 <<= 18
+# asm 1: pslld $18,<a7=int6464#6
+# asm 2: pslld $18,<a7=%xmm5
+pslld $18,%xmm5
+
+# qhasm: uint32323232                                                        b7 >>= 14
+# asm 1: psrld $14,<b7=int6464#7
+# asm 2: psrld $14,<b7=%xmm6
+psrld $14,%xmm6
+
+# qhasm:                                                                 diag0 ^= a7
+# asm 1: pxor  <a7=int6464#6,<diag0=int6464#1
+# asm 2: pxor  <a7=%xmm5,<diag0=%xmm0
+pxor  %xmm5,%xmm0
+
+# qhasm:                                                 diag3 <<<= 96
+# asm 1: pshufd $0x39,<diag3=int6464#4,<diag3=int6464#4
+# asm 2: pshufd $0x39,<diag3=%xmm3,<diag3=%xmm3
+pshufd $0x39,%xmm3,%xmm3
+
+# qhasm:                                                                 diag0 ^= b7
+# asm 1: pxor  <b7=int6464#7,<diag0=int6464#1
+# asm 2: pxor  <b7=%xmm6,<diag0=%xmm0
+pxor  %xmm6,%xmm0
+
+# qhasm: uint32323232        a0 += diag0
+# asm 1: paddd <diag0=int6464#1,<a0=int6464#5
+# asm 2: paddd <diag0=%xmm0,<a0=%xmm4
+paddd %xmm0,%xmm4
+
+# qhasm:                                 a1 = diag0
+# asm 1: movdqa <diag0=int6464#1,>a1=int6464#6
+# asm 2: movdqa <diag0=%xmm0,>a1=%xmm5
+movdqa %xmm0,%xmm5
+
+# qhasm:                     b0 = a0
+# asm 1: movdqa <a0=int6464#5,>b0=int6464#7
+# asm 2: movdqa <a0=%xmm4,>b0=%xmm6
+movdqa %xmm4,%xmm6
+
+# qhasm: uint32323232        a0 <<= 7
+# asm 1: pslld $7,<a0=int6464#5
+# asm 2: pslld $7,<a0=%xmm4
+pslld $7,%xmm4
+
+# qhasm: uint32323232        b0 >>= 25
+# asm 1: psrld $25,<b0=int6464#7
+# asm 2: psrld $25,<b0=%xmm6
+psrld $25,%xmm6
+
+# qhasm:                 diag3 ^= a0
+# asm 1: pxor  <a0=int6464#5,<diag3=int6464#4
+# asm 2: pxor  <a0=%xmm4,<diag3=%xmm3
+pxor  %xmm4,%xmm3
+
+# qhasm:                 diag3 ^= b0
+# asm 1: pxor  <b0=int6464#7,<diag3=int6464#4
+# asm 2: pxor  <b0=%xmm6,<diag3=%xmm3
+pxor  %xmm6,%xmm3
+
+# qhasm: uint32323232                        a1 += diag3
+# asm 1: paddd <diag3=int6464#4,<a1=int6464#6
+# asm 2: paddd <diag3=%xmm3,<a1=%xmm5
+paddd %xmm3,%xmm5
+
+# qhasm:                                                 a2 = diag3
+# asm 1: movdqa <diag3=int6464#4,>a2=int6464#5
+# asm 2: movdqa <diag3=%xmm3,>a2=%xmm4
+movdqa %xmm3,%xmm4
+
+# qhasm:                                     b1 = a1
+# asm 1: movdqa <a1=int6464#6,>b1=int6464#7
+# asm 2: movdqa <a1=%xmm5,>b1=%xmm6
+movdqa %xmm5,%xmm6
+
+# qhasm: uint32323232                        a1 <<= 9
+# asm 1: pslld $9,<a1=int6464#6
+# asm 2: pslld $9,<a1=%xmm5
+pslld $9,%xmm5
+
+# qhasm: uint32323232                        b1 >>= 23
+# asm 1: psrld $23,<b1=int6464#7
+# asm 2: psrld $23,<b1=%xmm6
+psrld $23,%xmm6
+
+# qhasm:                                 diag2 ^= a1
+# asm 1: pxor  <a1=int6464#6,<diag2=int6464#3
+# asm 2: pxor  <a1=%xmm5,<diag2=%xmm2
+pxor  %xmm5,%xmm2
+
+# qhasm:                 diag3 <<<= 32
+# asm 1: pshufd $0x93,<diag3=int6464#4,<diag3=int6464#4
+# asm 2: pshufd $0x93,<diag3=%xmm3,<diag3=%xmm3
+pshufd $0x93,%xmm3,%xmm3
+
+# qhasm:                                 diag2 ^= b1
+# asm 1: pxor  <b1=int6464#7,<diag2=int6464#3
+# asm 2: pxor  <b1=%xmm6,<diag2=%xmm2
+pxor  %xmm6,%xmm2
+
+# qhasm: uint32323232                                        a2 += diag2
+# asm 1: paddd <diag2=int6464#3,<a2=int6464#5
+# asm 2: paddd <diag2=%xmm2,<a2=%xmm4
+paddd %xmm2,%xmm4
+
+# qhasm:                                                                 a3 = diag2
+# asm 1: movdqa <diag2=int6464#3,>a3=int6464#6
+# asm 2: movdqa <diag2=%xmm2,>a3=%xmm5
+movdqa %xmm2,%xmm5
+
+# qhasm:                                                     b2 = a2
+# asm 1: movdqa <a2=int6464#5,>b2=int6464#7
+# asm 2: movdqa <a2=%xmm4,>b2=%xmm6
+movdqa %xmm4,%xmm6
+
+# qhasm: uint32323232                                        a2 <<= 13
+# asm 1: pslld $13,<a2=int6464#5
+# asm 2: pslld $13,<a2=%xmm4
+pslld $13,%xmm4
+
+# qhasm: uint32323232                                        b2 >>= 19
+# asm 1: psrld $19,<b2=int6464#7
+# asm 2: psrld $19,<b2=%xmm6
+psrld $19,%xmm6
+
+# qhasm:                                                 diag1 ^= a2
+# asm 1: pxor  <a2=int6464#5,<diag1=int6464#2
+# asm 2: pxor  <a2=%xmm4,<diag1=%xmm1
+pxor  %xmm4,%xmm1
+
+# qhasm:                                 diag2 <<<= 64
+# asm 1: pshufd $0x4e,<diag2=int6464#3,<diag2=int6464#3
+# asm 2: pshufd $0x4e,<diag2=%xmm2,<diag2=%xmm2
+pshufd $0x4e,%xmm2,%xmm2
+
+# qhasm:                                                 diag1 ^= b2
+# asm 1: pxor  <b2=int6464#7,<diag1=int6464#2
+# asm 2: pxor  <b2=%xmm6,<diag1=%xmm1
+pxor  %xmm6,%xmm1
+
+# qhasm: uint32323232                                                        a3 += diag1
+# asm 1: paddd <diag1=int6464#2,<a3=int6464#6
+# asm 2: paddd <diag1=%xmm1,<a3=%xmm5
+paddd %xmm1,%xmm5
+
+# qhasm:                 a4 = diag3
+# asm 1: movdqa <diag3=int6464#4,>a4=int6464#5
+# asm 2: movdqa <diag3=%xmm3,>a4=%xmm4
+movdqa %xmm3,%xmm4
+
+# qhasm:                                                                     b3 = a3
+# asm 1: movdqa <a3=int6464#6,>b3=int6464#7
+# asm 2: movdqa <a3=%xmm5,>b3=%xmm6
+movdqa %xmm5,%xmm6
+
+# qhasm: uint32323232                                                        a3 <<= 18
+# asm 1: pslld $18,<a3=int6464#6
+# asm 2: pslld $18,<a3=%xmm5
+pslld $18,%xmm5
+
+# qhasm: uint32323232                                                        b3 >>= 14
+# asm 1: psrld $14,<b3=int6464#7
+# asm 2: psrld $14,<b3=%xmm6
+psrld $14,%xmm6
+
+# qhasm:                                                                 diag0 ^= a3
+# asm 1: pxor  <a3=int6464#6,<diag0=int6464#1
+# asm 2: pxor  <a3=%xmm5,<diag0=%xmm0
+pxor  %xmm5,%xmm0
+
+# qhasm:                                                 diag1 <<<= 96
+# asm 1: pshufd $0x39,<diag1=int6464#2,<diag1=int6464#2
+# asm 2: pshufd $0x39,<diag1=%xmm1,<diag1=%xmm1
+pshufd $0x39,%xmm1,%xmm1
+
+# qhasm:                                                                 diag0 ^= b3
+# asm 1: pxor  <b3=int6464#7,<diag0=int6464#1
+# asm 2: pxor  <b3=%xmm6,<diag0=%xmm0
+pxor  %xmm6,%xmm0
+
+# qhasm: uint32323232        a4 += diag0
+# asm 1: paddd <diag0=int6464#1,<a4=int6464#5
+# asm 2: paddd <diag0=%xmm0,<a4=%xmm4
+paddd %xmm0,%xmm4
+
+# qhasm:                                 a5 = diag0
+# asm 1: movdqa <diag0=int6464#1,>a5=int6464#6
+# asm 2: movdqa <diag0=%xmm0,>a5=%xmm5
+movdqa %xmm0,%xmm5
+
+# qhasm:                     b4 = a4
+# asm 1: movdqa <a4=int6464#5,>b4=int6464#7
+# asm 2: movdqa <a4=%xmm4,>b4=%xmm6
+movdqa %xmm4,%xmm6
+
+# qhasm: uint32323232        a4 <<= 7
+# asm 1: pslld $7,<a4=int6464#5
+# asm 2: pslld $7,<a4=%xmm4
+pslld $7,%xmm4
+
+# qhasm: uint32323232        b4 >>= 25
+# asm 1: psrld $25,<b4=int6464#7
+# asm 2: psrld $25,<b4=%xmm6
+psrld $25,%xmm6
+
+# qhasm:                 diag1 ^= a4
+# asm 1: pxor  <a4=int6464#5,<diag1=int6464#2
+# asm 2: pxor  <a4=%xmm4,<diag1=%xmm1
+pxor  %xmm4,%xmm1
+
+# qhasm:                 diag1 ^= b4
+# asm 1: pxor  <b4=int6464#7,<diag1=int6464#2
+# asm 2: pxor  <b4=%xmm6,<diag1=%xmm1
+pxor  %xmm6,%xmm1
+
+# qhasm: uint32323232                        a5 += diag1
+# asm 1: paddd <diag1=int6464#2,<a5=int6464#6
+# asm 2: paddd <diag1=%xmm1,<a5=%xmm5
+paddd %xmm1,%xmm5
+
+# qhasm:                                                 a6 = diag1
+# asm 1: movdqa <diag1=int6464#2,>a6=int6464#5
+# asm 2: movdqa <diag1=%xmm1,>a6=%xmm4
+movdqa %xmm1,%xmm4
+
+# qhasm:                                     b5 = a5
+# asm 1: movdqa <a5=int6464#6,>b5=int6464#7
+# asm 2: movdqa <a5=%xmm5,>b5=%xmm6
+movdqa %xmm5,%xmm6
+
+# qhasm: uint32323232                        a5 <<= 9
+# asm 1: pslld $9,<a5=int6464#6
+# asm 2: pslld $9,<a5=%xmm5
+pslld $9,%xmm5
+
+# qhasm: uint32323232                        b5 >>= 23
+# asm 1: psrld $23,<b5=int6464#7
+# asm 2: psrld $23,<b5=%xmm6
+psrld $23,%xmm6
+
+# qhasm:                                 diag2 ^= a5
+# asm 1: pxor  <a5=int6464#6,<diag2=int6464#3
+# asm 2: pxor  <a5=%xmm5,<diag2=%xmm2
+pxor  %xmm5,%xmm2
+
+# qhasm:                 diag1 <<<= 32
+# asm 1: pshufd $0x93,<diag1=int6464#2,<diag1=int6464#2
+# asm 2: pshufd $0x93,<diag1=%xmm1,<diag1=%xmm1
+pshufd $0x93,%xmm1,%xmm1
+
+# qhasm:                                 diag2 ^= b5
+# asm 1: pxor  <b5=int6464#7,<diag2=int6464#3
+# asm 2: pxor  <b5=%xmm6,<diag2=%xmm2
+pxor  %xmm6,%xmm2
+
+# qhasm: uint32323232                                        a6 += diag2
+# asm 1: paddd <diag2=int6464#3,<a6=int6464#5
+# asm 2: paddd <diag2=%xmm2,<a6=%xmm4
+paddd %xmm2,%xmm4
+
+# qhasm:                                                                 a7 = diag2
+# asm 1: movdqa <diag2=int6464#3,>a7=int6464#6
+# asm 2: movdqa <diag2=%xmm2,>a7=%xmm5
+movdqa %xmm2,%xmm5
+
+# qhasm:                                                     b6 = a6
+# asm 1: movdqa <a6=int6464#5,>b6=int6464#7
+# asm 2: movdqa <a6=%xmm4,>b6=%xmm6
+movdqa %xmm4,%xmm6
+
+# qhasm: uint32323232                                        a6 <<= 13
+# asm 1: pslld $13,<a6=int6464#5
+# asm 2: pslld $13,<a6=%xmm4
+pslld $13,%xmm4
+
+# qhasm: uint32323232                                        b6 >>= 19
+# asm 1: psrld $19,<b6=int6464#7
+# asm 2: psrld $19,<b6=%xmm6
+psrld $19,%xmm6
+
+# qhasm:                                                 diag3 ^= a6
+# asm 1: pxor  <a6=int6464#5,<diag3=int6464#4
+# asm 2: pxor  <a6=%xmm4,<diag3=%xmm3
+pxor  %xmm4,%xmm3
+
+# qhasm:                                 diag2 <<<= 64
+# asm 1: pshufd $0x4e,<diag2=int6464#3,<diag2=int6464#3
+# asm 2: pshufd $0x4e,<diag2=%xmm2,<diag2=%xmm2
+pshufd $0x4e,%xmm2,%xmm2
+
+# qhasm:                                                 diag3 ^= b6
+# asm 1: pxor  <b6=int6464#7,<diag3=int6464#4
+# asm 2: pxor  <b6=%xmm6,<diag3=%xmm3
+pxor  %xmm6,%xmm3
+
+# qhasm:                  unsigned>? i -= 4
+# asm 1: sub  $4,<i=int32#1
+# asm 2: sub  $4,<i=%eax
+sub  $4,%eax
+
+# qhasm: uint32323232                                                        a7 += diag3
+# asm 1: paddd <diag3=int6464#4,<a7=int6464#6
+# asm 2: paddd <diag3=%xmm3,<a7=%xmm5
+paddd %xmm3,%xmm5
+
+# qhasm:                 a0 = diag1
+# asm 1: movdqa <diag1=int6464#2,>a0=int6464#5
+# asm 2: movdqa <diag1=%xmm1,>a0=%xmm4
+movdqa %xmm1,%xmm4
+
+# qhasm:                                                                     b7 = a7
+# asm 1: movdqa <a7=int6464#6,>b7=int6464#7
+# asm 2: movdqa <a7=%xmm5,>b7=%xmm6
+movdqa %xmm5,%xmm6
+
+# qhasm: uint32323232                                                        a7 <<= 18
+# asm 1: pslld $18,<a7=int6464#6
+# asm 2: pslld $18,<a7=%xmm5
+pslld $18,%xmm5
+
+# qhasm:                 b0 = 0
+# asm 1: pxor   >b0=int6464#8,>b0=int6464#8
+# asm 2: pxor   >b0=%xmm7,>b0=%xmm7
+pxor   %xmm7,%xmm7
+
+# qhasm: uint32323232                                                        b7 >>= 14
+# asm 1: psrld $14,<b7=int6464#7
+# asm 2: psrld $14,<b7=%xmm6
+psrld $14,%xmm6
+
+# qhasm:                                                                 diag0 ^= a7
+# asm 1: pxor  <a7=int6464#6,<diag0=int6464#1
+# asm 2: pxor  <a7=%xmm5,<diag0=%xmm0
+pxor  %xmm5,%xmm0
+
+# qhasm:                                                 diag3 <<<= 96
+# asm 1: pshufd $0x39,<diag3=int6464#4,<diag3=int6464#4
+# asm 2: pshufd $0x39,<diag3=%xmm3,<diag3=%xmm3
+pshufd $0x39,%xmm3,%xmm3
+
+# qhasm:                                                                 diag0 ^= b7
+# asm 1: pxor  <b7=int6464#7,<diag0=int6464#1
+# asm 2: pxor  <b7=%xmm6,<diag0=%xmm0
+pxor  %xmm6,%xmm0
+# comment:fp stack unchanged by jump
+
+# qhasm: goto mainloop2 if unsigned>
+ja ._mainloop2
+
+# qhasm: uint32323232 diag0 += x0
+# asm 1: paddd <x0=stack128#3,<diag0=int6464#1
+# asm 2: paddd <x0=64(%esp),<diag0=%xmm0
+paddd 64(%esp),%xmm0
+
+# qhasm: uint32323232 diag1 += x1
+# asm 1: paddd <x1=stack128#2,<diag1=int6464#2
+# asm 2: paddd <x1=48(%esp),<diag1=%xmm1
+paddd 48(%esp),%xmm1
+
+# qhasm: uint32323232 diag2 += x2
+# asm 1: paddd <x2=stack128#4,<diag2=int6464#3
+# asm 2: paddd <x2=80(%esp),<diag2=%xmm2
+paddd 80(%esp),%xmm2
+
+# qhasm: uint32323232 diag3 += x3
+# asm 1: paddd <x3=stack128#1,<diag3=int6464#4
+# asm 2: paddd <x3=32(%esp),<diag3=%xmm3
+paddd 32(%esp),%xmm3
+
+# qhasm: in0 = diag0
+# asm 1: movd   <diag0=int6464#1,>in0=int32#1
+# asm 2: movd   <diag0=%xmm0,>in0=%eax
+movd   %xmm0,%eax
+
+# qhasm: in12 = diag1
+# asm 1: movd   <diag1=int6464#2,>in12=int32#2
+# asm 2: movd   <diag1=%xmm1,>in12=%ecx
+movd   %xmm1,%ecx
+
+# qhasm: in8 = diag2
+# asm 1: movd   <diag2=int6464#3,>in8=int32#3
+# asm 2: movd   <diag2=%xmm2,>in8=%edx
+movd   %xmm2,%edx
+
+# qhasm: in4 = diag3
+# asm 1: movd   <diag3=int6464#4,>in4=int32#4
+# asm 2: movd   <diag3=%xmm3,>in4=%ebx
+movd   %xmm3,%ebx
+
+# qhasm: diag0 <<<= 96
+# asm 1: pshufd $0x39,<diag0=int6464#1,<diag0=int6464#1
+# asm 2: pshufd $0x39,<diag0=%xmm0,<diag0=%xmm0
+pshufd $0x39,%xmm0,%xmm0
+
+# qhasm: diag1 <<<= 96
+# asm 1: pshufd $0x39,<diag1=int6464#2,<diag1=int6464#2
+# asm 2: pshufd $0x39,<diag1=%xmm1,<diag1=%xmm1
+pshufd $0x39,%xmm1,%xmm1
+
+# qhasm: diag2 <<<= 96
+# asm 1: pshufd $0x39,<diag2=int6464#3,<diag2=int6464#3
+# asm 2: pshufd $0x39,<diag2=%xmm2,<diag2=%xmm2
+pshufd $0x39,%xmm2,%xmm2
+
+# qhasm: diag3 <<<= 96
+# asm 1: pshufd $0x39,<diag3=int6464#4,<diag3=int6464#4
+# asm 2: pshufd $0x39,<diag3=%xmm3,<diag3=%xmm3
+pshufd $0x39,%xmm3,%xmm3
+
+# qhasm: in0 ^= *(uint32 *) (m + 0)
+# asm 1: xorl 0(<m=int32#5),<in0=int32#1
+# asm 2: xorl 0(<m=%esi),<in0=%eax
+xorl 0(%esi),%eax
+
+# qhasm: in12 ^= *(uint32 *) (m + 48)
+# asm 1: xorl 48(<m=int32#5),<in12=int32#2
+# asm 2: xorl 48(<m=%esi),<in12=%ecx
+xorl 48(%esi),%ecx
+
+# qhasm: in8 ^= *(uint32 *) (m + 32)
+# asm 1: xorl 32(<m=int32#5),<in8=int32#3
+# asm 2: xorl 32(<m=%esi),<in8=%edx
+xorl 32(%esi),%edx
+
+# qhasm: in4 ^= *(uint32 *) (m + 16)
+# asm 1: xorl 16(<m=int32#5),<in4=int32#4
+# asm 2: xorl 16(<m=%esi),<in4=%ebx
+xorl 16(%esi),%ebx
+
+# qhasm: *(uint32 *) (out + 0) = in0
+# asm 1: movl <in0=int32#1,0(<out=int32#6)
+# asm 2: movl <in0=%eax,0(<out=%edi)
+movl %eax,0(%edi)
+
+# qhasm: *(uint32 *) (out + 48) = in12
+# asm 1: movl <in12=int32#2,48(<out=int32#6)
+# asm 2: movl <in12=%ecx,48(<out=%edi)
+movl %ecx,48(%edi)
+
+# qhasm: *(uint32 *) (out + 32) = in8
+# asm 1: movl <in8=int32#3,32(<out=int32#6)
+# asm 2: movl <in8=%edx,32(<out=%edi)
+movl %edx,32(%edi)
+
+# qhasm: *(uint32 *) (out + 16) = in4
+# asm 1: movl <in4=int32#4,16(<out=int32#6)
+# asm 2: movl <in4=%ebx,16(<out=%edi)
+movl %ebx,16(%edi)
+
+# qhasm: in5 = diag0
+# asm 1: movd   <diag0=int6464#1,>in5=int32#1
+# asm 2: movd   <diag0=%xmm0,>in5=%eax
+movd   %xmm0,%eax
+
+# qhasm: in1 = diag1
+# asm 1: movd   <diag1=int6464#2,>in1=int32#2
+# asm 2: movd   <diag1=%xmm1,>in1=%ecx
+movd   %xmm1,%ecx
+
+# qhasm: in13 = diag2
+# asm 1: movd   <diag2=int6464#3,>in13=int32#3
+# asm 2: movd   <diag2=%xmm2,>in13=%edx
+movd   %xmm2,%edx
+
+# qhasm: in9 = diag3
+# asm 1: movd   <diag3=int6464#4,>in9=int32#4
+# asm 2: movd   <diag3=%xmm3,>in9=%ebx
+movd   %xmm3,%ebx
+
+# qhasm: diag0 <<<= 96
+# asm 1: pshufd $0x39,<diag0=int6464#1,<diag0=int6464#1
+# asm 2: pshufd $0x39,<diag0=%xmm0,<diag0=%xmm0
+pshufd $0x39,%xmm0,%xmm0
+
+# qhasm: diag1 <<<= 96
+# asm 1: pshufd $0x39,<diag1=int6464#2,<diag1=int6464#2
+# asm 2: pshufd $0x39,<diag1=%xmm1,<diag1=%xmm1
+pshufd $0x39,%xmm1,%xmm1
+
+# qhasm: diag2 <<<= 96
+# asm 1: pshufd $0x39,<diag2=int6464#3,<diag2=int6464#3
+# asm 2: pshufd $0x39,<diag2=%xmm2,<diag2=%xmm2
+pshufd $0x39,%xmm2,%xmm2
+
+# qhasm: diag3 <<<= 96
+# asm 1: pshufd $0x39,<diag3=int6464#4,<diag3=int6464#4
+# asm 2: pshufd $0x39,<diag3=%xmm3,<diag3=%xmm3
+pshufd $0x39,%xmm3,%xmm3
+
+# qhasm: in5 ^= *(uint32 *) (m + 20)
+# asm 1: xorl 20(<m=int32#5),<in5=int32#1
+# asm 2: xorl 20(<m=%esi),<in5=%eax
+xorl 20(%esi),%eax
+
+# qhasm: in1 ^= *(uint32 *) (m + 4)
+# asm 1: xorl 4(<m=int32#5),<in1=int32#2
+# asm 2: xorl 4(<m=%esi),<in1=%ecx
+xorl 4(%esi),%ecx
+
+# qhasm: in13 ^= *(uint32 *) (m + 52)
+# asm 1: xorl 52(<m=int32#5),<in13=int32#3
+# asm 2: xorl 52(<m=%esi),<in13=%edx
+xorl 52(%esi),%edx
+
+# qhasm: in9 ^= *(uint32 *) (m + 36)
+# asm 1: xorl 36(<m=int32#5),<in9=int32#4
+# asm 2: xorl 36(<m=%esi),<in9=%ebx
+xorl 36(%esi),%ebx
+
+# qhasm: *(uint32 *) (out + 20) = in5
+# asm 1: movl <in5=int32#1,20(<out=int32#6)
+# asm 2: movl <in5=%eax,20(<out=%edi)
+movl %eax,20(%edi)
+
+# qhasm: *(uint32 *) (out + 4) = in1
+# asm 1: movl <in1=int32#2,4(<out=int32#6)
+# asm 2: movl <in1=%ecx,4(<out=%edi)
+movl %ecx,4(%edi)
+
+# qhasm: *(uint32 *) (out + 52) = in13
+# asm 1: movl <in13=int32#3,52(<out=int32#6)
+# asm 2: movl <in13=%edx,52(<out=%edi)
+movl %edx,52(%edi)
+
+# qhasm: *(uint32 *) (out + 36) = in9
+# asm 1: movl <in9=int32#4,36(<out=int32#6)
+# asm 2: movl <in9=%ebx,36(<out=%edi)
+movl %ebx,36(%edi)
+
+# qhasm: in10 = diag0
+# asm 1: movd   <diag0=int6464#1,>in10=int32#1
+# asm 2: movd   <diag0=%xmm0,>in10=%eax
+movd   %xmm0,%eax
+
+# qhasm: in6 = diag1
+# asm 1: movd   <diag1=int6464#2,>in6=int32#2
+# asm 2: movd   <diag1=%xmm1,>in6=%ecx
+movd   %xmm1,%ecx
+
+# qhasm: in2 = diag2
+# asm 1: movd   <diag2=int6464#3,>in2=int32#3
+# asm 2: movd   <diag2=%xmm2,>in2=%edx
+movd   %xmm2,%edx
+
+# qhasm: in14 = diag3
+# asm 1: movd   <diag3=int6464#4,>in14=int32#4
+# asm 2: movd   <diag3=%xmm3,>in14=%ebx
+movd   %xmm3,%ebx
+
+# qhasm: diag0 <<<= 96
+# asm 1: pshufd $0x39,<diag0=int6464#1,<diag0=int6464#1
+# asm 2: pshufd $0x39,<diag0=%xmm0,<diag0=%xmm0
+pshufd $0x39,%xmm0,%xmm0
+
+# qhasm: diag1 <<<= 96
+# asm 1: pshufd $0x39,<diag1=int6464#2,<diag1=int6464#2
+# asm 2: pshufd $0x39,<diag1=%xmm1,<diag1=%xmm1
+pshufd $0x39,%xmm1,%xmm1
+
+# qhasm: diag2 <<<= 96
+# asm 1: pshufd $0x39,<diag2=int6464#3,<diag2=int6464#3
+# asm 2: pshufd $0x39,<diag2=%xmm2,<diag2=%xmm2
+pshufd $0x39,%xmm2,%xmm2
+
+# qhasm: diag3 <<<= 96
+# asm 1: pshufd $0x39,<diag3=int6464#4,<diag3=int6464#4
+# asm 2: pshufd $0x39,<diag3=%xmm3,<diag3=%xmm3
+pshufd $0x39,%xmm3,%xmm3
+
+# qhasm: in10 ^= *(uint32 *) (m + 40)
+# asm 1: xorl 40(<m=int32#5),<in10=int32#1
+# asm 2: xorl 40(<m=%esi),<in10=%eax
+xorl 40(%esi),%eax
+
+# qhasm: in6 ^= *(uint32 *) (m + 24)
+# asm 1: xorl 24(<m=int32#5),<in6=int32#2
+# asm 2: xorl 24(<m=%esi),<in6=%ecx
+xorl 24(%esi),%ecx
+
+# qhasm: in2 ^= *(uint32 *) (m + 8)
+# asm 1: xorl 8(<m=int32#5),<in2=int32#3
+# asm 2: xorl 8(<m=%esi),<in2=%edx
+xorl 8(%esi),%edx
+
+# qhasm: in14 ^= *(uint32 *) (m + 56)
+# asm 1: xorl 56(<m=int32#5),<in14=int32#4
+# asm 2: xorl 56(<m=%esi),<in14=%ebx
+xorl 56(%esi),%ebx
+
+# qhasm: *(uint32 *) (out + 40) = in10
+# asm 1: movl <in10=int32#1,40(<out=int32#6)
+# asm 2: movl <in10=%eax,40(<out=%edi)
+movl %eax,40(%edi)
+
+# qhasm: *(uint32 *) (out + 24) = in6
+# asm 1: movl <in6=int32#2,24(<out=int32#6)
+# asm 2: movl <in6=%ecx,24(<out=%edi)
+movl %ecx,24(%edi)
+
+# qhasm: *(uint32 *) (out + 8) = in2
+# asm 1: movl <in2=int32#3,8(<out=int32#6)
+# asm 2: movl <in2=%edx,8(<out=%edi)
+movl %edx,8(%edi)
+
+# qhasm: *(uint32 *) (out + 56) = in14
+# asm 1: movl <in14=int32#4,56(<out=int32#6)
+# asm 2: movl <in14=%ebx,56(<out=%edi)
+movl %ebx,56(%edi)
+
+# qhasm: in15 = diag0
+# asm 1: movd   <diag0=int6464#1,>in15=int32#1
+# asm 2: movd   <diag0=%xmm0,>in15=%eax
+movd   %xmm0,%eax
+
+# qhasm: in11 = diag1
+# asm 1: movd   <diag1=int6464#2,>in11=int32#2
+# asm 2: movd   <diag1=%xmm1,>in11=%ecx
+movd   %xmm1,%ecx
+
+# qhasm: in7 = diag2
+# asm 1: movd   <diag2=int6464#3,>in7=int32#3
+# asm 2: movd   <diag2=%xmm2,>in7=%edx
+movd   %xmm2,%edx
+
+# qhasm: in3 = diag3
+# asm 1: movd   <diag3=int6464#4,>in3=int32#4
+# asm 2: movd   <diag3=%xmm3,>in3=%ebx
+movd   %xmm3,%ebx
+
+# qhasm: in15 ^= *(uint32 *) (m + 60)
+# asm 1: xorl 60(<m=int32#5),<in15=int32#1
+# asm 2: xorl 60(<m=%esi),<in15=%eax
+xorl 60(%esi),%eax
+
+# qhasm: in11 ^= *(uint32 *) (m + 44)
+# asm 1: xorl 44(<m=int32#5),<in11=int32#2
+# asm 2: xorl 44(<m=%esi),<in11=%ecx
+xorl 44(%esi),%ecx
+
+# qhasm: in7 ^= *(uint32 *) (m + 28)
+# asm 1: xorl 28(<m=int32#5),<in7=int32#3
+# asm 2: xorl 28(<m=%esi),<in7=%edx
+xorl 28(%esi),%edx
+
+# qhasm: in3 ^= *(uint32 *) (m + 12)
+# asm 1: xorl 12(<m=int32#5),<in3=int32#4
+# asm 2: xorl 12(<m=%esi),<in3=%ebx
+xorl 12(%esi),%ebx
+
+# qhasm: *(uint32 *) (out + 60) = in15
+# asm 1: movl <in15=int32#1,60(<out=int32#6)
+# asm 2: movl <in15=%eax,60(<out=%edi)
+movl %eax,60(%edi)
+
+# qhasm: *(uint32 *) (out + 44) = in11
+# asm 1: movl <in11=int32#2,44(<out=int32#6)
+# asm 2: movl <in11=%ecx,44(<out=%edi)
+movl %ecx,44(%edi)
+
+# qhasm: *(uint32 *) (out + 28) = in7
+# asm 1: movl <in7=int32#3,28(<out=int32#6)
+# asm 2: movl <in7=%edx,28(<out=%edi)
+movl %edx,28(%edi)
+
+# qhasm: *(uint32 *) (out + 12) = in3
+# asm 1: movl <in3=int32#4,12(<out=int32#6)
+# asm 2: movl <in3=%ebx,12(<out=%edi)
+movl %ebx,12(%edi)
+
+# qhasm:   bytes = bytes_stack
+# asm 1: movl <bytes_stack=stack32#7,>bytes=int32#1
+# asm 2: movl <bytes_stack=24(%esp),>bytes=%eax
+movl 24(%esp),%eax
+
+# qhasm:   in8 = ((uint32 *)&x2)[0]
+# asm 1: movl <x2=stack128#4,>in8=int32#2
+# asm 2: movl <x2=80(%esp),>in8=%ecx
+movl 80(%esp),%ecx
+
+# qhasm:   in9 = ((uint32 *)&x3)[1]
+# asm 1: movl 4+<x3=stack128#1,>in9=int32#3
+# asm 2: movl 4+<x3=32(%esp),>in9=%edx
+movl 4+32(%esp),%edx
+
+# qhasm:   carry? in8 += 1
+# asm 1: add  $1,<in8=int32#2
+# asm 2: add  $1,<in8=%ecx
+add  $1,%ecx
+
+# qhasm:   in9 += 0 + carry
+# asm 1: adc $0,<in9=int32#3
+# asm 2: adc $0,<in9=%edx
+adc $0,%edx
+
+# qhasm:   ((uint32 *)&x2)[0] = in8
+# asm 1: movl <in8=int32#2,>x2=stack128#4
+# asm 2: movl <in8=%ecx,>x2=80(%esp)
+movl %ecx,80(%esp)
+
+# qhasm:   ((uint32 *)&x3)[1] = in9
+# asm 1: movl <in9=int32#3,4+<x3=stack128#1
+# asm 2: movl <in9=%edx,4+<x3=32(%esp)
+movl %edx,4+32(%esp)
+
+# qhasm:                          unsigned>? unsigned<? bytes - 64
+# asm 1: cmp  $64,<bytes=int32#1
+# asm 2: cmp  $64,<bytes=%eax
+cmp  $64,%eax
+# comment:fp stack unchanged by jump
+
+# qhasm:   goto bytesatleast65 if unsigned>
+ja ._bytesatleast65
+# comment:fp stack unchanged by jump
+
+# qhasm:     goto bytesatleast64 if !unsigned<
+jae ._bytesatleast64
+
+# qhasm:       m = out
+# asm 1: mov  <out=int32#6,>m=int32#5
+# asm 2: mov  <out=%edi,>m=%esi
+mov  %edi,%esi
+
+# qhasm:       out = ctarget
+# asm 1: movl <ctarget=stack32#6,>out=int32#6
+# asm 2: movl <ctarget=20(%esp),>out=%edi
+movl 20(%esp),%edi
+
+# qhasm:       i = bytes
+# asm 1: mov  <bytes=int32#1,>i=int32#2
+# asm 2: mov  <bytes=%eax,>i=%ecx
+mov  %eax,%ecx
+
+# qhasm:       while (i) { *out++ = *m++; --i }
+rep movsb
+# comment:fp stack unchanged by fallthrough
+
+# qhasm:     bytesatleast64:
+._bytesatleast64:
+# comment:fp stack unchanged by fallthrough
+
+# qhasm:     done:
+._done:
+
+# qhasm:     eax = eax_stack
+# asm 1: movl <eax_stack=stack32#1,>eax=int32#1
+# asm 2: movl <eax_stack=0(%esp),>eax=%eax
+movl 0(%esp),%eax
+
+# qhasm:     ebx = ebx_stack
+# asm 1: movl <ebx_stack=stack32#2,>ebx=int32#4
+# asm 2: movl <ebx_stack=4(%esp),>ebx=%ebx
+movl 4(%esp),%ebx
+
+# qhasm:     esi = esi_stack
+# asm 1: movl <esi_stack=stack32#3,>esi=int32#5
+# asm 2: movl <esi_stack=8(%esp),>esi=%esi
+movl 8(%esp),%esi
+
+# qhasm:     edi = edi_stack
+# asm 1: movl <edi_stack=stack32#4,>edi=int32#6
+# asm 2: movl <edi_stack=12(%esp),>edi=%edi
+movl 12(%esp),%edi
+
+# qhasm:     ebp = ebp_stack
+# asm 1: movl <ebp_stack=stack32#5,>ebp=int32#7
+# asm 2: movl <ebp_stack=16(%esp),>ebp=%ebp
+movl 16(%esp),%ebp
+
+# qhasm:     leave
+add %eax,%esp
+xor %eax,%eax
+ret
+
+# qhasm:   bytesatleast65:
+._bytesatleast65:
+
+# qhasm:   bytes -= 64
+# asm 1: sub  $64,<bytes=int32#1
+# asm 2: sub  $64,<bytes=%eax
+sub  $64,%eax
+
+# qhasm:   out += 64
+# asm 1: add  $64,<out=int32#6
+# asm 2: add  $64,<out=%edi
+add  $64,%edi
+
+# qhasm:   m += 64
+# asm 1: add  $64,<m=int32#5
+# asm 2: add  $64,<m=%esi
+add  $64,%esi
+# comment:fp stack unchanged by jump
+
+# qhasm: goto bytesbetween1and255
+jmp ._bytesbetween1and255
diff --git a/src/crypto/cipher/salsa2012/xmm/salsa2012_xmm.c b/src/crypto/cipher/salsa2012/xmm/salsa2012_xmm.c
new file mode 100644
index 0000000..1bed11e
--- /dev/null
+++ b/src/crypto/cipher/salsa2012/xmm/salsa2012_xmm.c
@@ -0,0 +1,84 @@
+/*
+  Copyright (c) 2012-2013, Matthias Schiffer <mschiffer@universe-factory.net>
+  All rights reserved.
+
+  Redistribution and use in source and binary forms, with or without
+  modification, are permitted provided that the following conditions are met:
+
+    1. Redistributions of source code must retain the above copyright notice,
+       this list of conditions and the following disclaimer.
+    2. Redistributions in binary form must reproduce the above copyright notice,
+       this list of conditions and the following disclaimer in the documentation
+       and/or other materials provided with the distribution.
+
+  THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS "AS IS"
+  AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
+  IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE ARE
+  DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT HOLDER OR CONTRIBUTORS BE LIABLE
+  FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL
+  DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR
+  SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER
+  CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY,
+  OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE
+  OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
+*/
+
+/*
+  The assembly implementations were written by D. J. Bernstein and are
+  Public Domain. For more information see http://cr.yp.to/snuffle.html
+*/
+
+#include "../../../../crypto.h"
+#include "../../../../cpuid.h"
+
+
+#define KEYBYTES 32
+
+
+#ifdef __x86_64__
+#define crypto_stream_salsa2012_xor crypto_stream_salsa2012_amd64_xmm6_xor
+#endif
+
+#ifdef __i386__
+#define crypto_stream_salsa2012_xor crypto_stream_salsa2012_x86_xmm5_xor
+#endif
+
+
+int crypto_stream_salsa2012_xor(unsigned char *c, const unsigned char *m, unsigned long long mlen, const unsigned char *n, const unsigned char *k);
+
+
+struct fastd_cipher_state {
+	uint8_t key[KEYBYTES];
+};
+
+
+static bool salsa2012_available(void) {
+	return fastd_cpuid() & CPUID_SSE2;
+}
+
+static fastd_cipher_state_t* salsa2012_init(const uint8_t *key) {
+	fastd_cipher_state_t *state = malloc(sizeof(fastd_cipher_state_t));
+	memcpy(state->key, key, KEYBYTES);
+
+	return state;
+}
+
+static bool salsa2012_crypt(const fastd_cipher_state_t *state, fastd_block128_t *out, const fastd_block128_t *in, size_t len, const uint8_t *iv) {
+	crypto_stream_salsa2012_xor(out->b, in->b, len, iv, state->key);
+	return true;
+}
+
+static void salsa2012_free(fastd_cipher_state_t *state) {
+	if (state) {
+		secure_memzero(state, sizeof(*state));
+		free(state);
+	}
+}
+
+const fastd_cipher_t fastd_cipher_salsa2012_xmm = {
+	.available = salsa2012_available,
+
+	.init = salsa2012_init,
+	.crypt = salsa2012_crypt,
+	.free = salsa2012_free,
+};
-- 
cgit v1.2.3