In visitSTORE, always use FindBetterChain, rather than only when UseAA is enabled.

Retrying after fixing overly aggressive load-store forwarding optimization.

Simplify Consecutive Merge Store Candidate Search

Now that address aliasing is much less conservative, push through
simplified store merging search which only checks for parallel stores
through the chain subgraph. This is cleaner as the separation of
non-interfering loads/stores from the store-merging logic.

Whem merging stores, search up the chain through a single load, and
finds all possible stores by looking down from through a load and a
TokenFactor to all stores visited. This improves the quality of the
output SelectionDAG and generally the output CodeGen (with some
exceptions).

Additional Minor Changes:

   1. Finishes removing unused AliasLoad code
   2. Unifies the the chain aggregation in the merged stores across
      code paths
   3. Re-add the Store node to the worklist after calling
      SimplifyDemandedBits.
   4. Increase GatherAllAliasesMaxDepth from 6 to 18. That number is
      arbitrary, but seemed sufficient to not cause regressions in
      tests.

This finishes the change Matt Arsenault started in r246307 and
jyknight's original patch.

Many tests required some changes as memory operations are now
reorderable. Some tests relying on the order were changed to use
volatile memory operations

Noteworthy tests:

    CodeGen/AArch64/argument-blocks.ll -
      It's not entirely clear what the test_varargs_stackalign test is
      supposed to be asserting, but the new code looks right.

    CodeGen/AArch64/arm64-memset-inline.lli -
    CodeGen/AArch64/arm64-stur.ll -
    CodeGen/ARM/memset-inline.ll -

      The backend now generates *worse* code due to store merging
      succeeding, as we do do a 16-byte constant-zero store efficiently.

    CodeGen/AArch64/merge-store.ll -
      Improved, but there still seems to be an extraneous vector insert
      from an element to itself?

    CodeGen/PowerPC/ppc64-align-long-double.ll -
      Worse code emitted in this case, due to the improved store->load
      forwarding.

    CodeGen/X86/dag-merge-fast-accesses.ll -
    CodeGen/X86/MergeConsecutiveStores.ll -
    CodeGen/X86/stores-merging.ll -
    CodeGen/Mips/load-store-left-right.ll -
      Restored correct merging of non-aligned stores

    CodeGen/AMDGPU/promote-alloca-stored-pointer-value.ll -
      Improved. Correctly merges buffer_store_dword calls

    CodeGen/AMDGPU/si-triv-disjoint-mem-access.ll -
      Improved. Sidesteps loading a stored value and
      merges two stores

    CodeGen/X86/pr18023.ll -
      This test has been removed, as it was asserting incorrect
      behavior. Non-volatile stores *CAN* be moved past volatile loads,
      and now are.

    CodeGen/X86/vector-idiv.ll -
    CodeGen/X86/vector-lzcnt-128.ll -
      It's basically impossible to tell what these tests are actually
      testing. But, looks like the code got better due to the memory
      operations being recognized as non-aliasing.

    CodeGen/X86/win32-eh.ll -
      Both loads of the securitycookie are now merged.

Reviewers: arsenm, hfinkel, tstellarAMD, jyknight, nhaehnle

Subscribers: wdng, nhaehnle, nemanjai, arsenm, weimingz, niravd, RKSimon, aemerson, qcolombet, dsanders, resistor, tstellarAMD, t.p.northover, spatel

Differential Revision: https://reviews.llvm.org/D14834

llvm-svn: 289221

25 files changed, 1054 insertions, 1255 deletions

diff --git a/llvm/test/CodeGen/X86/2010-09-17-SideEffectsInChain.ll b/llvm/test/CodeGen/X86/2010-09-17-SideEffectsInChain.ll
index b7380196bd9..54a7763eb69 100644
--- a/llvm/test/CodeGen/X86/2010-09-17-SideEffectsInChain.ll
+++ b/llvm/test/CodeGen/X86/2010-09-17-SideEffectsInChain.ll
@@ -1,4 +1,4 @@
-; RUN: llc < %s -combiner-alias-analysis -march=x86-64 -mcpu=core2 | FileCheck %s
+; RUN: llc < %s -march=x86-64 -mcpu=core2 | FileCheck %s
 
 target datalayout = "e-p:64:64:64-i1:8:8-i8:8:8-i16:16:16-i32:32:32-i64:64:64-f32:32:32-f64:64:64-v64:64:64-v128:128:128-a0:0:64-s0:64:64-f80:128:128-n8:16:32:64"
 target triple = "x86_64-apple-darwin10.4"
diff --git a/llvm/test/CodeGen/X86/2012-11-28-merge-store-alias.ll b/llvm/test/CodeGen/X86/2012-11-28-merge-store-alias.ll
index c16deeff3d9..2e8206a7591 100644
--- a/llvm/test/CodeGen/X86/2012-11-28-merge-store-alias.ll
+++ b/llvm/test/CodeGen/X86/2012-11-28-merge-store-alias.ll
@@ -3,8 +3,8 @@
 ; CHECK: merge_stores_can
 ; CHECK: callq foo
 ; CHECK: xorps %xmm0, %xmm0
-; CHECK-NEXT: movl 36(%rsp), %ebp
 ; CHECK-NEXT: movups  %xmm0
+; CHECK-NEXT: movl 36(%rsp), %ebp
 ; CHECK: callq foo
 ; CHECK: ret
 declare i32 @foo([10 x i32]* )
diff --git a/llvm/test/CodeGen/X86/MergeConsecutiveStores.ll b/llvm/test/CodeGen/X86/MergeConsecutiveStores.ll
index b50253bf2b0..ee7ae24e288 100644
--- a/llvm/test/CodeGen/X86/MergeConsecutiveStores.ll
+++ b/llvm/test/CodeGen/X86/MergeConsecutiveStores.ll
@@ -292,16 +292,12 @@ block4:                                       ; preds = %4, %.lr.ph
   ret void
 }
 
-;; On x86, even unaligned copies should be merged to vector ops.
-;; TODO: however, this cannot happen at the moment, due to brokenness
-;; in MergeConsecutiveStores. See UseAA FIXME in DAGCombiner.cpp
-;; visitSTORE.
-
+;; On x86, even unaligned copies can be merged to vector ops.
 ; CHECK-LABEL: merge_loads_no_align:
 ;  load:
-; CHECK-NOT: vmovups ;; TODO
+; CHECK: vmovups
 ;  store:
-; CHECK-NOT: vmovups ;; TODO
+; CHECK: vmovups
 ; CHECK: ret
 define void @merge_loads_no_align(i32 %count, %struct.B* noalias nocapture %q, %struct.B* noalias nocapture %p) nounwind uwtable noinline ssp {
   %a1 = icmp sgt i32 %count, 0
@@ -549,8 +545,8 @@ define void @merge_vec_element_and_scalar_load([6 x i64]* %array) {
 
 ; CHECK-LABEL: merge_vec_element_and_scalar_load
 ; CHECK:      movq	(%rdi), %rax
+; CHECK-NEXT: movq	8(%rdi), %rcx
 ; CHECK-NEXT: movq	%rax, 32(%rdi)
-; CHECK-NEXT: movq	8(%rdi), %rax
-; CHECK-NEXT: movq	%rax, 40(%rdi)
+; CHECK-NEXT: movq	%rcx, 40(%rdi)
 ; CHECK-NEXT: retq
 }
diff --git a/llvm/test/CodeGen/X86/avx512-mask-op.ll b/llvm/test/CodeGen/X86/avx512-mask-op.ll
index 4bfdd711deb..6ea64774528 100644
--- a/llvm/test/CodeGen/X86/avx512-mask-op.ll
+++ b/llvm/test/CodeGen/X86/avx512-mask-op.ll
@@ -1173,10 +1173,6 @@ define void @ktest_2(<32 x float> %in, float * %base) {
 ; KNL-NEXT:    kmovw %k0, %eax
 ; KNL-NEXT:    vpinsrb $15, %eax, %xmm2, %xmm2
 ; KNL-NEXT:    vinserti128 $1, %xmm3, %ymm2, %ymm2
-; KNL-NEXT:    vpsllw $7, %ymm2, %ymm2
-; KNL-NEXT:    vpand {{.*}}(%rip), %ymm2, %ymm2
-; KNL-NEXT:    vpxor %ymm3, %ymm3, %ymm3
-; KNL-NEXT:    vpcmpgtb %ymm2, %ymm3, %ymm2
 ; KNL-NEXT:    vmovups 4(%rdi), %zmm3 {%k2} {z}
 ; KNL-NEXT:    vmovups 68(%rdi), %zmm4 {%k1} {z}
 ; KNL-NEXT:    vcmpltps %zmm4, %zmm1, %k0
diff --git a/llvm/test/CodeGen/X86/chain_order.ll b/llvm/test/CodeGen/X86/chain_order.ll
index 8c3aa6e1515..cc48e5b6149 100644
--- a/llvm/test/CodeGen/X86/chain_order.ll
+++ b/llvm/test/CodeGen/X86/chain_order.ll
@@ -11,9 +11,9 @@ define void @cftx020(double* nocapture %a) {
 ; CHECK-NEXT:    vmovhpd {{.*#+}} xmm1 = xmm1[0],mem[0]
 ; CHECK-NEXT:    vaddpd %xmm1, %xmm0, %xmm0
 ; CHECK-NEXT:    vmovupd (%rdi), %xmm1
-; CHECK-NEXT:    vsubpd 16(%rdi), %xmm1, %xmm1
 ; CHECK-NEXT:    vmovupd %xmm0, (%rdi)
-; CHECK-NEXT:    vmovupd %xmm1, 16(%rdi)
+; CHECK-NEXT:    vsubpd 16(%rdi), %xmm1, %xmm0
+; CHECK-NEXT:    vmovupd %xmm0, 16(%rdi)
 ; CHECK-NEXT:    retq
 entry:
   %0 = load double, double* %a, align 8
diff --git a/llvm/test/CodeGen/X86/clear_upper_vector_element_bits.ll b/llvm/test/CodeGen/X86/clear_upper_vector_element_bits.ll
index 693bf2e17d5..f8833b733af 100644
--- a/llvm/test/CodeGen/X86/clear_upper_vector_element_bits.ll
+++ b/llvm/test/CodeGen/X86/clear_upper_vector_element_bits.ll
@@ -151,47 +151,47 @@ define <16 x i8> @_clearupper16xi8a(<16 x i8>) nounwind {
 ; SSE-NEXT:    movaps %xmm0, -{{[0-9]+}}(%rsp)
 ; SSE-NEXT:    movzbl -{{[0-9]+}}(%rsp), %eax
 ; SSE-NEXT:    movd %eax, %xmm0
-; SSE-NEXT:    movzbl -{{[0-9]+}}(%rsp), %r9d
-; SSE-NEXT:    movzbl -{{[0-9]+}}(%rsp), %edx
-; SSE-NEXT:    movzbl -{{[0-9]+}}(%rsp), %esi
-; SSE-NEXT:    movzbl -{{[0-9]+}}(%rsp), %r8d
-; SSE-NEXT:    movzbl -{{[0-9]+}}(%rsp), %edi
 ; SSE-NEXT:    movzbl -{{[0-9]+}}(%rsp), %eax
 ; SSE-NEXT:    movd %eax, %xmm1
 ; SSE-NEXT:    punpcklbw {{.*#+}} xmm1 = xmm1[0],xmm0[0],xmm1[1],xmm0[1],xmm1[2],xmm0[2],xmm1[3],xmm0[3],xmm1[4],xmm0[4],xmm1[5],xmm0[5],xmm1[6],xmm0[6],xmm1[7],xmm0[7]
-; SSE-NEXT:    movd %esi, %xmm0
 ; SSE-NEXT:    movzbl -{{[0-9]+}}(%rsp), %eax
-; SSE-NEXT:    movzbl -{{[0-9]+}}(%rsp), %esi
-; SSE-NEXT:    movzbl -{{[0-9]+}}(%rsp), %ecx
-; SSE-NEXT:    movd %ecx, %xmm2
+; SSE-NEXT:    movd %eax, %xmm0
+; SSE-NEXT:    movzbl -{{[0-9]+}}(%rsp), %eax
+; SSE-NEXT:    movd %eax, %xmm2
 ; SSE-NEXT:    punpcklbw {{.*#+}} xmm2 = xmm2[0],xmm0[0],xmm2[1],xmm0[1],xmm2[2],xmm0[2],xmm2[3],xmm0[3],xmm2[4],xmm0[4],xmm2[5],xmm0[5],xmm2[6],xmm0[6],xmm2[7],xmm0[7]
 ; SSE-NEXT:    punpcklbw {{.*#+}} xmm2 = xmm2[0],xmm1[0],xmm2[1],xmm1[1],xmm2[2],xmm1[2],xmm2[3],xmm1[3],xmm2[4],xmm1[4],xmm2[5],xmm1[5],xmm2[6],xmm1[6],xmm2[7],xmm1[7]
-; SSE-NEXT:    movd %edx, %xmm0
-; SSE-NEXT:    movd %esi, %xmm1
-; SSE-NEXT:    punpcklbw {{.*#+}} xmm1 = xmm1[0],xmm0[0],xmm1[1],xmm0[1],xmm1[2],xmm0[2],xmm1[3],xmm0[3],xmm1[4],xmm0[4],xmm1[5],xmm0[5],xmm1[6],xmm0[6],xmm1[7],xmm0[7]
-; SSE-NEXT:    movd %edi, %xmm0
-; SSE-NEXT:    movzbl -{{[0-9]+}}(%rsp), %ecx
-; SSE-NEXT:    movzbl -{{[0-9]+}}(%rsp), %edx
-; SSE-NEXT:    movd %edx, %xmm3
+; SSE-NEXT:    movzbl -{{[0-9]+}}(%rsp), %eax
+; SSE-NEXT:    movd %eax, %xmm0
+; SSE-NEXT:    movzbl -{{[0-9]+}}(%rsp), %eax
+; SSE-NEXT:    movd %eax, %xmm3
 ; SSE-NEXT:    punpcklbw {{.*#+}} xmm3 = xmm3[0],xmm0[0],xmm3[1],xmm0[1],xmm3[2],xmm0[2],xmm3[3],xmm0[3],xmm3[4],xmm0[4],xmm3[5],xmm0[5],xmm3[6],xmm0[6],xmm3[7],xmm0[7]
-; SSE-NEXT:    punpcklbw {{.*#+}} xmm3 = xmm3[0],xmm1[0],xmm3[1],xmm1[1],xmm3[2],xmm1[2],xmm3[3],xmm1[3],xmm3[4],xmm1[4],xmm3[5],xmm1[5],xmm3[6],xmm1[6],xmm3[7],xmm1[7]
-; SSE-NEXT:    punpcklbw {{.*#+}} xmm3 = xmm3[0],xmm2[0],xmm3[1],xmm2[1],xmm3[2],xmm2[2],xmm3[3],xmm2[3],xmm3[4],xmm2[4],xmm3[5],xmm2[5],xmm3[6],xmm2[6],xmm3[7],xmm2[7]
-; SSE-NEXT:    movd %r9d, %xmm0
+; SSE-NEXT:    movzbl -{{[0-9]+}}(%rsp), %eax
+; SSE-NEXT:    movd %eax, %xmm0
+; SSE-NEXT:    movzbl -{{[0-9]+}}(%rsp), %eax
 ; SSE-NEXT:    movd %eax, %xmm1
 ; SSE-NEXT:    punpcklbw {{.*#+}} xmm1 = xmm1[0],xmm0[0],xmm1[1],xmm0[1],xmm1[2],xmm0[2],xmm1[3],xmm0[3],xmm1[4],xmm0[4],xmm1[5],xmm0[5],xmm1[6],xmm0[6],xmm1[7],xmm0[7]
-; SSE-NEXT:    movd %r8d, %xmm0
-; SSE-NEXT:    movd %ecx, %xmm2
-; SSE-NEXT:    punpcklbw {{.*#+}} xmm2 = xmm2[0],xmm0[0],xmm2[1],xmm0[1],xmm2[2],xmm0[2],xmm2[3],xmm0[3],xmm2[4],xmm0[4],xmm2[5],xmm0[5],xmm2[6],xmm0[6],xmm2[7],xmm0[7]
-; SSE-NEXT:    punpcklbw {{.*#+}} xmm2 = xmm2[0],xmm1[0],xmm2[1],xmm1[1],xmm2[2],xmm1[2],xmm2[3],xmm1[3],xmm2[4],xmm1[4],xmm2[5],xmm1[5],xmm2[6],xmm1[6],xmm2[7],xmm1[7]
+; SSE-NEXT:    punpcklbw {{.*#+}} xmm1 = xmm1[0],xmm3[0],xmm1[1],xmm3[1],xmm1[2],xmm3[2],xmm1[3],xmm3[3],xmm1[4],xmm3[4],xmm1[5],xmm3[5],xmm1[6],xmm3[6],xmm1[7],xmm3[7]
+; SSE-NEXT:    punpcklbw {{.*#+}} xmm1 = xmm1[0],xmm2[0],xmm1[1],xmm2[1],xmm1[2],xmm2[2],xmm1[3],xmm2[3],xmm1[4],xmm2[4],xmm1[5],xmm2[5],xmm1[6],xmm2[6],xmm1[7],xmm2[7]
 ; SSE-NEXT:    movd {{.*#+}} xmm0 = mem[0],zero,zero,zero
-; SSE-NEXT:    movd {{.*#+}} xmm1 = mem[0],zero,zero,zero
-; SSE-NEXT:    punpcklbw {{.*#+}} xmm1 = xmm1[0],xmm0[0],xmm1[1],xmm0[1],xmm1[2],xmm0[2],xmm1[3],xmm0[3],xmm1[4],xmm0[4],xmm1[5],xmm0[5],xmm1[6],xmm0[6],xmm1[7],xmm0[7]
-; SSE-NEXT:    movd {{.*#+}} xmm4 = mem[0],zero,zero,zero
+; SSE-NEXT:    movd {{.*#+}} xmm2 = mem[0],zero,zero,zero
+; SSE-NEXT:    punpcklbw {{.*#+}} xmm2 = xmm2[0],xmm0[0],xmm2[1],xmm0[1],xmm2[2],xmm0[2],xmm2[3],xmm0[3],xmm2[4],xmm0[4],xmm2[5],xmm0[5],xmm2[6],xmm0[6],xmm2[7],xmm0[7]
+; SSE-NEXT:    movd {{.*#+}} xmm3 = mem[0],zero,zero,zero
 ; SSE-NEXT:    movd {{.*#+}} xmm0 = mem[0],zero,zero,zero
+; SSE-NEXT:    punpcklbw {{.*#+}} xmm0 = xmm0[0],xmm3[0],xmm0[1],xmm3[1],xmm0[2],xmm3[2],xmm0[3],xmm3[3],xmm0[4],xmm3[4],xmm0[5],xmm3[5],xmm0[6],xmm3[6],xmm0[7],xmm3[7]
+; SSE-NEXT:    punpcklbw {{.*#+}} xmm0 = xmm0[0],xmm2[0],xmm0[1],xmm2[1],xmm0[2],xmm2[2],xmm0[3],xmm2[3],xmm0[4],xmm2[4],xmm0[5],xmm2[5],xmm0[6],xmm2[6],xmm0[7],xmm2[7]
+; SSE-NEXT:    movzbl -{{[0-9]+}}(%rsp), %eax
+; SSE-NEXT:    movd %eax, %xmm2
+; SSE-NEXT:    movzbl -{{[0-9]+}}(%rsp), %eax
+; SSE-NEXT:    movd %eax, %xmm3
+; SSE-NEXT:    punpcklbw {{.*#+}} xmm3 = xmm3[0],xmm2[0],xmm3[1],xmm2[1],xmm3[2],xmm2[2],xmm3[3],xmm2[3],xmm3[4],xmm2[4],xmm3[5],xmm2[5],xmm3[6],xmm2[6],xmm3[7],xmm2[7]
+; SSE-NEXT:    movzbl -{{[0-9]+}}(%rsp), %eax
+; SSE-NEXT:    movd %eax, %xmm2
+; SSE-NEXT:    movzbl -{{[0-9]+}}(%rsp), %eax
+; SSE-NEXT:    movd %eax, %xmm4
+; SSE-NEXT:    punpcklbw {{.*#+}} xmm4 = xmm4[0],xmm2[0],xmm4[1],xmm2[1],xmm4[2],xmm2[2],xmm4[3],xmm2[3],xmm4[4],xmm2[4],xmm4[5],xmm2[5],xmm4[6],xmm2[6],xmm4[7],xmm2[7]
+; SSE-NEXT:    punpcklbw {{.*#+}} xmm4 = xmm4[0],xmm3[0],xmm4[1],xmm3[1],xmm4[2],xmm3[2],xmm4[3],xmm3[3],xmm4[4],xmm3[4],xmm4[5],xmm3[5],xmm4[6],xmm3[6],xmm4[7],xmm3[7]
 ; SSE-NEXT:    punpcklbw {{.*#+}} xmm0 = xmm0[0],xmm4[0],xmm0[1],xmm4[1],xmm0[2],xmm4[2],xmm0[3],xmm4[3],xmm0[4],xmm4[4],xmm0[5],xmm4[5],xmm0[6],xmm4[6],xmm0[7],xmm4[7]
 ; SSE-NEXT:    punpcklbw {{.*#+}} xmm0 = xmm0[0],xmm1[0],xmm0[1],xmm1[1],xmm0[2],xmm1[2],xmm0[3],xmm1[3],xmm0[4],xmm1[4],xmm0[5],xmm1[5],xmm0[6],xmm1[6],xmm0[7],xmm1[7]
-; SSE-NEXT:    punpcklbw {{.*#+}} xmm0 = xmm0[0],xmm2[0],xmm0[1],xmm2[1],xmm0[2],xmm2[2],xmm0[3],xmm2[3],xmm0[4],xmm2[4],xmm0[5],xmm2[5],xmm0[6],xmm2[6],xmm0[7],xmm2[7]
-; SSE-NEXT:    punpcklbw {{.*#+}} xmm0 = xmm0[0],xmm3[0],xmm0[1],xmm3[1],xmm0[2],xmm3[2],xmm0[3],xmm3[3],xmm0[4],xmm3[4],xmm0[5],xmm3[5],xmm0[6],xmm3[6],xmm0[7],xmm3[7]
 ; SSE-NEXT:    pand {{.*}}(%rip), %xmm0
 ; SSE-NEXT:    retq
 ;
diff --git a/llvm/test/CodeGen/X86/combiner-aa-0.ll b/llvm/test/CodeGen/X86/combiner-aa-0.ll
deleted file mode 100644
index 403059d90ab..00000000000
--- a/llvm/test/CodeGen/X86/combiner-aa-0.ll
+++ /dev/null
@@ -1,20 +0,0 @@
-; RUN: llc < %s -march=x86-64 -combiner-global-alias-analysis -combiner-alias-analysis
-
-target datalayout = "e-p:64:64:64-i1:8:8-i8:8:8-i16:16:16-i32:32:32-i64:64:64-f32:32:32-f64:64:64-v64:64:64-v128:128:128-a0:0:64-s0:64:64-f80:128:128"
-	%struct.Hash_Key = type { [4 x i32], i32 }
-@g_flipV_hashkey = external global %struct.Hash_Key, align 16		; <%struct.Hash_Key*> [#uses=1]
-
-define void @foo() nounwind {
-	%t0 = load i32, i32* undef, align 16		; <i32> [#uses=1]
-	%t1 = load i32, i32* null, align 4		; <i32> [#uses=1]
-	%t2 = srem i32 %t0, 32		; <i32> [#uses=1]
-	%t3 = shl i32 1, %t2		; <i32> [#uses=1]
-	%t4 = xor i32 %t3, %t1		; <i32> [#uses=1]
-	store i32 %t4, i32* null, align 4
-	%t5 = getelementptr %struct.Hash_Key, %struct.Hash_Key* @g_flipV_hashkey, i64 0, i32 0, i64 0		; <i32*> [#uses=2]
-	%t6 = load i32, i32* %t5, align 4		; <i32> [#uses=1]
-	%t7 = shl i32 1, undef		; <i32> [#uses=1]
-	%t8 = xor i32 %t7, %t6		; <i32> [#uses=1]
-	store i32 %t8, i32* %t5, align 4
-	unreachable
-}
diff --git a/llvm/test/CodeGen/X86/combiner-aa-1.ll b/llvm/test/CodeGen/X86/combiner-aa-1.ll
deleted file mode 100644
index cc3e5ca1260..00000000000
--- a/llvm/test/CodeGen/X86/combiner-aa-1.ll
+++ /dev/null
@@ -1,23 +0,0 @@
-; RUN: llc < %s --combiner-alias-analysis --combiner-global-alias-analysis
-; PR4880
-
-target datalayout = "e-p:32:32:32-i1:8:8-i8:8:8-i16:16:16-i32:32:32-i64:32:64-f32:32:32-f64:32:64-v64:64:64-v128:128:128-a0:0:64-f80:32:32"
-target triple = "i386-pc-linux-gnu"
-
-%struct.alst_node = type { %struct.node }
-%struct.arg_node = type { %struct.node, i8*, %struct.alst_node* }
-%struct.arglst_node = type { %struct.alst_node, %struct.arg_node*, %struct.arglst_node* }
-%struct.lam_node = type { %struct.alst_node, %struct.arg_node*, %struct.alst_node* }
-%struct.node = type { i32 (...)**, %struct.node* }
-
-define i32 @._ZN8lam_node18resolve_name_clashEP8arg_nodeP9alst_node._ZNK8lam_nodeeqERK8exp_node._ZN11arglst_nodeD0Ev(%struct.lam_node* %this.this, %struct.arg_node* %outer_arg, %struct.alst_node* %env.cmp, %struct.arglst_node* %this, i32 %functionID) {
-comb_entry:
-  %.SV59 = alloca %struct.node*                   ; <%struct.node**> [#uses=1]
-  %0 = load i32 (...)**, i32 (...)*** null, align 4            ; <i32 (...)**> [#uses=1]
-  %1 = getelementptr inbounds i32 (...)*, i32 (...)** %0, i32 3 ; <i32 (...)**> [#uses=1]
-  %2 = load i32 (...)*, i32 (...)** %1, align 4               ; <i32 (...)*> [#uses=1]
-  store %struct.node* undef, %struct.node** %.SV59
-  %3 = bitcast i32 (...)* %2 to i32 (%struct.node*)* ; <i32 (%struct.node*)*> [#uses=1]
-  %4 = tail call i32 %3(%struct.node* undef)      ; <i32> [#uses=0]
-  unreachable
-}
diff --git a/llvm/test/CodeGen/X86/copy-eflags.ll b/llvm/test/CodeGen/X86/copy-eflags.ll
index 796c1ecd8c7..d98d8a7839b 100644
--- a/llvm/test/CodeGen/X86/copy-eflags.ll
+++ b/llvm/test/CodeGen/X86/copy-eflags.ll
@@ -9,19 +9,22 @@ target triple = "i686-unknown-linux-gnu"
 @.str = private unnamed_addr constant [4 x i8] c"%d\0A\00", align 1
 
 ; CHECK-LABEL: func:
-; This tests whether eax is properly saved/restored around the lahf/sahf
-; instruction sequences.
+; This tests whether eax is properly saved/restored around the
+; lahf/sahf instruction sequences. We make mem op volatile to prevent
+; their reordering to avoid spills.
+
+
 define i32 @func() {
 entry:
   %bval = load i8, i8* @b
   %inc = add i8 %bval, 1
-  store i8 %inc, i8* @b
-  %cval = load i32, i32* @c
+  store volatile i8 %inc, i8* @b
+  %cval = load volatile i32, i32* @c
   %inc1 = add nsw i32 %cval, 1
-  store i32 %inc1, i32* @c
-  %aval = load i8, i8* @a
+  store volatile i32 %inc1, i32* @c
+  %aval = load volatile i8, i8* @a
   %inc2 = add i8 %aval, 1
-  store i8 %inc2, i8* @a
+  store volatile i8 %inc2, i8* @a
 ; Copy flags produced by the incb of %inc1 to a register, need to save+restore
 ; eax around it. The flags will be reused by %tobool.
 ; CHECK: pushl %eax
diff --git a/llvm/test/CodeGen/X86/dag-merge-fast-accesses.ll b/llvm/test/CodeGen/X86/dag-merge-fast-accesses.ll
index 867881d83d3..e5dfccb278c 100644
--- a/llvm/test/CodeGen/X86/dag-merge-fast-accesses.ll
+++ b/llvm/test/CodeGen/X86/dag-merge-fast-accesses.ll
@@ -51,19 +51,11 @@ define void @merge_vec_element_store(<4 x double> %v, double* %ptr) {
 }
 
 
-;; TODO: FAST *should* be:
-;;    movups (%rdi), %xmm0
-;;    movups %xmm0, 40(%rdi)
-;; ..but is not currently. See the UseAA FIXME in DAGCombiner.cpp
-;; visitSTORE.
-
 define void @merge_vec_load_and_stores(i64 *%ptr) {
 ; FAST-LABEL: merge_vec_load_and_stores:
 ; FAST:       # BB#0:
-; FAST-NEXT:    movq (%rdi), %rax
-; FAST-NEXT:    movq 8(%rdi), %rcx
-; FAST-NEXT:    movq %rax, 40(%rdi)
-; FAST-NEXT:    movq %rcx, 48(%rdi)
+; FAST-NEXT:    movups (%rdi), %xmm0
+; FAST-NEXT:    movups %xmm0, 40(%rdi)
 ; FAST-NEXT:    retq
 ;
 ; SLOW-LABEL: merge_vec_load_and_stores:
diff --git a/llvm/test/CodeGen/X86/dont-trunc-store-double-to-float.ll b/llvm/test/CodeGen/X86/dont-trunc-store-double-to-float.ll
index 8a334d21631..05245d0d9e1 100644
--- a/llvm/test/CodeGen/X86/dont-trunc-store-double-to-float.ll
+++ b/llvm/test/CodeGen/X86/dont-trunc-store-double-to-float.ll
@@ -1,9 +1,9 @@
 ; RUN: llc -march=x86 < %s | FileCheck %s
 
 ; CHECK-LABEL: @bar
-; CHECK: movl $1074339512,
-; CHECK: movl $1374389535,
-; CHECK: movl $1078523331,
+; CHECK-DAG: movl $1074339512,
+; CHECK-DAG: movl $1374389535,
+; CHECK-DAG: movl $1078523331,
 define void @bar() unnamed_addr {
 entry-block:
   %a = alloca double
diff --git a/llvm/test/CodeGen/X86/extractelement-legalization-store-ordering.ll b/llvm/test/CodeGen/X86/extractelement-legalization-store-ordering.ll
index 946516c8a46..d472cdad0ea 100644
--- a/llvm/test/CodeGen/X86/extractelement-legalization-store-ordering.ll
+++ b/llvm/test/CodeGen/X86/extractelement-legalization-store-ordering.ll
@@ -18,13 +18,13 @@ target datalayout = "e-m:o-p:32:32-f64:32:64-f80:128-n8:16:32-S128"
 ; CHECK-NEXT: movdqa %xmm0, (%edx)
 ; CHECK-NEXT: shll $4, %ecx
 ; CHECK-NEXT: movl (%ecx,%edx), %esi
-; CHECK-NEXT: movl 12(%ecx,%edx), %edi
+; CHECK-NEXT: movl 4(%ecx,%edx), %edi
 ; CHECK-NEXT: movl 8(%ecx,%edx), %ebx
-; CHECK-NEXT: movl 4(%ecx,%edx), %edx
+; CHECK-NEXT: movl 12(%ecx,%edx), %edx
 ; CHECK-NEXT: movl %esi, 12(%eax,%ecx)
-; CHECK-NEXT: movl %edx, (%eax,%ecx)
+; CHECK-NEXT: movl %edi, (%eax,%ecx)
 ; CHECK-NEXT: movl %ebx, 8(%eax,%ecx)
-; CHECK-NEXT: movl %edi, 4(%eax,%ecx)
+; CHECK-NEXT: movl %edx, 4(%eax,%ecx)
 ; CHECK-NEXT: popl %esi
 ; CHECK-NEXT: popl %edi
 ; CHECK-NEXT: popl %ebx
diff --git a/llvm/test/CodeGen/X86/i256-add.ll b/llvm/test/CodeGen/X86/i256-add.ll
index 6164d898ca1..d2821dc3dce 100644
--- a/llvm/test/CodeGen/X86/i256-add.ll
+++ b/llvm/test/CodeGen/X86/i256-add.ll
@@ -2,17 +2,17 @@
 ; RUN: grep adcl %t | count 7
 ; RUN: grep sbbl %t | count 7
 
-define void @add(i256* %p, i256* %q) nounwind {
+define void @add(i256* %p, i256* %q, i256* %r) nounwind {
   %a = load i256, i256* %p
   %b = load i256, i256* %q
   %c = add i256 %a, %b
-  store i256 %c, i256* %p
+  store i256 %c, i256* %r
   ret void
 }
-define void @sub(i256* %p, i256* %q) nounwind {
+define void @sub(i256* %p, i256* %q, i256* %r) nounwind {
   %a = load i256, i256* %p
   %b = load i256, i256* %q
   %c = sub i256 %a, %b
-  store i256 %c, i256* %p
+  store i256 %c, i256* %r
   ret void
 }
diff --git a/llvm/test/CodeGen/X86/i386-shrink-wrapping.ll b/llvm/test/CodeGen/X86/i386-shrink-wrapping.ll
index 2c3e384b70a..d4e099ac655 100644
--- a/llvm/test/CodeGen/X86/i386-shrink-wrapping.ll
+++ b/llvm/test/CodeGen/X86/i386-shrink-wrapping.ll
@@ -55,8 +55,7 @@ target triple = "i386-apple-macosx10.5"
 ;
 ; CHECK-NEXT: L_e$non_lazy_ptr, [[E:%[a-z]+]]
 ; CHECK-NEXT: movb [[D]], ([[E]])
-; CHECK-NEXT: L_f$non_lazy_ptr, [[F:%[a-z]+]]
-; CHECK-NEXT: movsbl ([[F]]), [[CONV:%[a-z]+]]
+; CHECK-NEXT: movsbl ([[E]]), [[CONV:%[a-z]+]]
 ; CHECK-NEXT: movl $6, [[CONV:%[a-z]+]]
 ; The eflags is used in the next instruction.
 ; If that instruction disappear, we are not exercising the bug
@@ -96,7 +95,7 @@ for.end:                                          ; preds = %for.cond.preheader
   %.b3 = load i1, i1* @d, align 1
   %tmp2 = select i1 %.b3, i8 0, i8 6
   store i8 %tmp2, i8* @e, align 1
-  %tmp3 = load i8, i8* @f, align 1
+  %tmp3 = load i8, i8* @e, align 1
   %conv = sext i8 %tmp3 to i32
   %add = add nsw i32 %conv, 1
   %rem = srem i32 %tmp1, %add
diff --git a/llvm/test/CodeGen/X86/live-range-nosubreg.ll b/llvm/test/CodeGen/X86/live-range-nosubreg.ll
index f28d59237b4..899a375221c 100644
--- a/llvm/test/CodeGen/X86/live-range-nosubreg.ll
+++ b/llvm/test/CodeGen/X86/live-range-nosubreg.ll
@@ -1,7 +1,6 @@
-; RUN: llc -march=x86-64 < %s | FileCheck %s
+; RUN: llc -march=x86-64 < %s
 
-; Check for a sane output. This testcase used to crash. See PR29132.
-; CHECK: leal -1
+; This testcase used to crash. See PR29132.
 
 target triple = "x86_64-unknown-linux-gnu"
 
diff --git a/llvm/test/CodeGen/X86/merge-consecutive-loads-128.ll b/llvm/test/CodeGen/X86/merge-consecutive-loads-128.ll
index 1f2bd4bb0dc..5f07aa80b45 100644
--- a/llvm/test/CodeGen/X86/merge-consecutive-loads-128.ll
+++ b/llvm/test/CodeGen/X86/merge-consecutive-loads-128.ll
@@ -1037,12 +1037,12 @@ define <2 x i64> @merge_2i64_i64_12_volatile(i64* %ptr) nounwind uwtable noinlin
 define <4 x float> @merge_4f32_f32_2345_volatile(float* %ptr) nounwind uwtable noinline ssp {
 ; SSE2-LABEL: merge_4f32_f32_2345_volatile:
 ; SSE2:       # BB#0:
-; SSE2-NEXT:    movss {{.*#+}} xmm0 = mem[0],zero,zero,zero
 ; SSE2-NEXT:    movss {{.*#+}} xmm1 = mem[0],zero,zero,zero
+; SSE2-NEXT:    movss {{.*#+}} xmm0 = mem[0],zero,zero,zero
 ; SSE2-NEXT:    movss {{.*#+}} xmm2 = mem[0],zero,zero,zero
-; SSE2-NEXT:    movss {{.*#+}} xmm3 = mem[0],zero,zero,zero
 ; SSE2-NEXT:    unpcklps {{.*#+}} xmm2 = xmm2[0],xmm1[0],xmm2[1],xmm1[1]
-; SSE2-NEXT:    unpcklps {{.*#+}} xmm0 = xmm0[0],xmm3[0],xmm0[1],xmm3[1]
+; SSE2-NEXT:    movss {{.*#+}} xmm1 = mem[0],zero,zero,zero
+; SSE2-NEXT:    unpcklps {{.*#+}} xmm0 = xmm0[0],xmm1[0],xmm0[1],xmm1[1]
 ; SSE2-NEXT:    unpcklps {{.*#+}} xmm0 = xmm0[0],xmm2[0],xmm0[1],xmm2[1]
 ; SSE2-NEXT:    retq
 ;
@@ -1065,13 +1065,13 @@ define <4 x float> @merge_4f32_f32_2345_volatile(float* %ptr) nounwind uwtable n
 ; X32-SSE1-LABEL: merge_4f32_f32_2345_volatile:
 ; X32-SSE1:       # BB#0:
 ; X32-SSE1-NEXT:    movl {{[0-9]+}}(%esp), %eax
-; X32-SSE1-NEXT:    movss {{.*#+}} xmm0 = mem[0],zero,zero,zero
-; X32-SSE1-NEXT:    movss {{.*#+}} xmm1 = mem[0],zero,zero,zero
-; X32-SSE1-NEXT:    movss {{.*#+}} xmm2 = mem[0],zero,zero,zero
-; X32-SSE1-NEXT:    movss {{.*#+}} xmm3 = mem[0],zero,zero,zero
-; X32-SSE1-NEXT:    unpcklps {{.*#+}} xmm2 = xmm2[0],xmm1[0],xmm2[1],xmm1[1]
-; X32-SSE1-NEXT:    unpcklps {{.*#+}} xmm0 = xmm0[0],xmm3[0],xmm0[1],xmm3[1]
-; X32-SSE1-NEXT:    unpcklps {{.*#+}} xmm0 = xmm0[0],xmm2[0],xmm0[1],xmm2[1]
+; X32-SSE1-DAG:    movss 8(%eax), %[[R0:xmm[0-3]]]  # [[R0]] = mem[0],zero,zero,zero
+; X32-SSE1-DAG:    movss 12(%eax), %[[R1:xmm[0-3]]] # [[R1]] = mem[0],zero,zero,zero
+; X32-SSE1-DAG:    movss 16(%eax), %[[R2:xmm[0-3]]] # [[R2]] = mem[0],zero,zero,zero
+; X32-SSE1-DAG:    movss 20(%eax), %[[R3:xmm[0-3]]] # [[R3]] = mem[0],zero,zero,zero
+; X32-SSE1-DAG:    unpcklps %[[R2]], %[[R0]] # [[R0]] = [[R0]][0],[[R2]][0],[[R0]][1],[[R2]][1]
+; X32-SSE1-DAG:    unpcklps %[[R3]], %[[R1]] # [[R1]] = [[R1]][0],[[R3]][0],[[R1]][1],[[R3]][1]
+; X32-SSE1-DAG:    unpcklps %[[R1]], %[[R0]] # [[R0]] = [[R0]][0],[[R1]][0],[[R0]][1],[[R1]][1]
 ; X32-SSE1-NEXT:    retl
 ;
 ; X32-SSE41-LABEL: merge_4f32_f32_2345_volatile:
diff --git a/llvm/test/CodeGen/X86/merge-consecutive-loads-256.ll b/llvm/test/CodeGen/X86/merge-consecutive-loads-256.ll
index 48a277a9c95..79bce603d12 100644
--- a/llvm/test/CodeGen/X86/merge-consecutive-loads-256.ll
+++ b/llvm/test/CodeGen/X86/merge-consecutive-loads-256.ll
@@ -682,10 +682,10 @@ define <16 x i16> @merge_16i16_i16_0uu3zzuuuuuzCuEF_volatile(i16* %ptr) nounwind
 ; AVX1:       # BB#0:
 ; AVX1-NEXT:    vpxor %xmm0, %xmm0, %xmm0
 ; AVX1-NEXT:    vpinsrw $0, (%rdi), %xmm0, %xmm1
-; AVX1-NEXT:    vpinsrw $3, 6(%rdi), %xmm1, %xmm1
 ; AVX1-NEXT:    vpinsrw $4, 24(%rdi), %xmm0, %xmm0
 ; AVX1-NEXT:    vpinsrw $6, 28(%rdi), %xmm0, %xmm0
 ; AVX1-NEXT:    vpinsrw $7, 30(%rdi), %xmm0, %xmm0
+; AVX1-NEXT:    vpinsrw $3, 6(%rdi), %xmm1, %xmm1
 ; AVX1-NEXT:    vinsertf128 $1, %xmm0, %ymm1, %ymm0
 ; AVX1-NEXT:    retq
 ;
@@ -693,10 +693,10 @@ define <16 x i16> @merge_16i16_i16_0uu3zzuuuuuzCuEF_volatile(i16* %ptr) nounwind
 ; AVX2:       # BB#0:
 ; AVX2-NEXT:    vpxor %xmm0, %xmm0, %xmm0
 ; AVX2-NEXT:    vpinsrw $0, (%rdi), %xmm0, %xmm1
-; AVX2-NEXT:    vpinsrw $3, 6(%rdi), %xmm1, %xmm1
 ; AVX2-NEXT:    vpinsrw $4, 24(%rdi), %xmm0, %xmm0
 ; AVX2-NEXT:    vpinsrw $6, 28(%rdi), %xmm0, %xmm0
 ; AVX2-NEXT:    vpinsrw $7, 30(%rdi), %xmm0, %xmm0
+; AVX2-NEXT:    vpinsrw $3, 6(%rdi), %xmm1, %xmm1
 ; AVX2-NEXT:    vinserti128 $1, %xmm0, %ymm1, %ymm0
 ; AVX2-NEXT:    retq
 ;
@@ -704,10 +704,10 @@ define <16 x i16> @merge_16i16_i16_0uu3zzuuuuuzCuEF_volatile(i16* %ptr) nounwind
 ; AVX512F:       # BB#0:
 ; AVX512F-NEXT:    vpxor %xmm0, %xmm0, %xmm0
 ; AVX512F-NEXT:    vpinsrw $0, (%rdi), %xmm0, %xmm1
-; AVX512F-NEXT:    vpinsrw $3, 6(%rdi), %xmm1, %xmm1
 ; AVX512F-NEXT:    vpinsrw $4, 24(%rdi), %xmm0, %xmm0
 ; AVX512F-NEXT:    vpinsrw $6, 28(%rdi), %xmm0, %xmm0
 ; AVX512F-NEXT:    vpinsrw $7, 30(%rdi), %xmm0, %xmm0
+; AVX512F-NEXT:    vpinsrw $3, 6(%rdi), %xmm1, %xmm1
 ; AVX512F-NEXT:    vinserti128 $1, %xmm0, %ymm1, %ymm0
 ; AVX512F-NEXT:    retq
 ;
@@ -716,10 +716,10 @@ define <16 x i16> @merge_16i16_i16_0uu3zzuuuuuzCuEF_volatile(i16* %ptr) nounwind
 ; X32-AVX-NEXT:    movl {{[0-9]+}}(%esp), %eax
 ; X32-AVX-NEXT:    vpxor %xmm0, %xmm0, %xmm0
 ; X32-AVX-NEXT:    vpinsrw $0, (%eax), %xmm0, %xmm1
-; X32-AVX-NEXT:    vpinsrw $3, 6(%eax), %xmm1, %xmm1
 ; X32-AVX-NEXT:    vpinsrw $4, 24(%eax), %xmm0, %xmm0
 ; X32-AVX-NEXT:    vpinsrw $6, 28(%eax), %xmm0, %xmm0
 ; X32-AVX-NEXT:    vpinsrw $7, 30(%eax), %xmm0, %xmm0
+; X32-AVX-NEXT:    vpinsrw $3, 6(%eax), %xmm1, %xmm1
 ; X32-AVX-NEXT:    vinsertf128 $1, %xmm0, %ymm1, %ymm0
 ; X32-AVX-NEXT:    retl
   %ptr0 = getelementptr inbounds i16, i16* %ptr, i64 0
diff --git a/llvm/test/CodeGen/X86/merge-store-partially-alias-loads.ll b/llvm/test/CodeGen/X86/merge-store-partially-alias-loads.ll
index 735e64a076d..5a180229bb9 100644
--- a/llvm/test/CodeGen/X86/merge-store-partially-alias-loads.ll
+++ b/llvm/test/CodeGen/X86/merge-store-partially-alias-loads.ll
@@ -21,11 +21,11 @@
 ; DBGDAG-DAG: [[LD2:t[0-9]+]]: i16,ch = load<LD2[%tmp81](align=1)> [[ENTRYTOKEN]], [[BASEPTR]], undef:i64
 ; DBGDAG-DAG: [[LD1:t[0-9]+]]: i8,ch = load<LD1[%tmp12]> [[ENTRYTOKEN]], [[ADDPTR]], undef:i64
 
-; DBGDAG: [[LOADTOKEN:t[0-9]+]]: ch = TokenFactor [[LD2]]:1, [[LD1]]:1
+; DBGDAG-DAG: [[ST1:t[0-9]+]]: ch = store<ST1[%tmp14]> [[ENTRYTOKEN]], [[LD1]], t{{[0-9]+}}, undef:i64
+; DBGDAG-DAG: [[LOADTOKEN:t[0-9]+]]: ch = TokenFactor [[LD2]]:1, [[LD1]]:1
+; DBGDAG: [[ST2:t[0-9]+]]: ch = store<ST2[%tmp10](align=1)> [[LOADTOKEN]], [[LD2]], t{{[0-9]+}}, undef:i64
 
-; DBGDAG-DAG: [[ST2:t[0-9]+]]: ch = store<ST2[%tmp10](align=1)> [[LOADTOKEN]], [[LD2]], t{{[0-9]+}}, undef:i64
-; DBGDAG-DAG: [[ST1:t[0-9]+]]: ch = store<ST1[%tmp14]> [[ST2]], [[LD1]], t{{[0-9]+}}, undef:i64
-; DBGDAG: X86ISD::RET_FLAG [[ST1]],
+; DBGDAG: X86ISD::RET_FLAG t{{[0-9]+}},
 
 ; DBGDAG: Type-legalized selection DAG: BB#0 'merge_store_partial_overlap_load:'
 define void @merge_store_partial_overlap_load([4 x i8]* %tmp) {
diff --git a/llvm/test/CodeGen/X86/pr18023.ll b/llvm/test/CodeGen/X86/pr18023.ll
deleted file mode 100644
index c7ea20c281b..00000000000
--- a/llvm/test/CodeGen/X86/pr18023.ll
+++ /dev/null
@@ -1,31 +0,0 @@
-; RUN: llc < %s -mtriple x86_64-apple-macosx10.9.0 | FileCheck %s
-; PR18023
-
-; CHECK: movabsq $4294967296, %rcx
-; CHECK: movq  %rcx, (%rax)
-; CHECK: movl  $1, 4(%rax)
-; CHECK: movl  $0, 4(%rax)
-; CHECK: movq  $1, 4(%rax)
-
-@c = common global i32 0, align 4
-@a = common global [3 x i32] zeroinitializer, align 4
-@b = common global i32 0, align 4
-@.str = private unnamed_addr constant [4 x i8] c"%d\0A\00", align 1
-
-define void @func() {
-  store i32 1, i32* getelementptr inbounds ([3 x i32], [3 x i32]* @a, i64 0, i64 1), align 4
-  store i32 0, i32* getelementptr inbounds ([3 x i32], [3 x i32]* @a, i64 0, i64 0), align 4
-  %1 = load volatile i32, i32* @b, align 4
-  store i32 1, i32* getelementptr inbounds ([3 x i32], [3 x i32]* @a, i64 0, i64 1), align 4
-  store i32 0, i32* getelementptr inbounds ([3 x i32], [3 x i32]* @a, i64 0, i64 1), align 4
-  %2 = load volatile i32, i32* @b, align 4
-  store i32 1, i32* getelementptr inbounds ([3 x i32], [3 x i32]* @a, i64 0, i64 1), align 4
-  store i32 0, i32* getelementptr inbounds ([3 x i32], [3 x i32]* @a, i64 0, i64 2), align 4
-  %3 = load volatile i32, i32* @b, align 4
-  store i32 3, i32* @c, align 4
-  %4 = load i32, i32* getelementptr inbounds ([3 x i32], [3 x i32]* @a, i64 0, i64 1), align 4
-  %call = call i32 (i8*, ...) @printf(i8* getelementptr inbounds ([4 x i8], [4 x i8]* @.str, i64 0, i64 0), i32 %4)
-  ret void
-}
-
-declare i32 @printf(i8*, ...)
diff --git a/llvm/test/CodeGen/X86/split-store.ll b/llvm/test/CodeGen/X86/split-store.ll
index 8a276f8d2f7..1d3fba69ca7 100644
--- a/llvm/test/CodeGen/X86/split-store.ll
+++ b/llvm/test/CodeGen/X86/split-store.ll
@@ -1,8 +1,8 @@
 ; RUN: llc -mtriple=x86_64-unknown-unknown < %s | FileCheck %s
 
 ; CHECK-LABEL: int32_float_pair
-; CHECK: movl %edi, (%rsi)
-; CHECK: movss %xmm0, 4(%rsi)
+; CHECK-DAG: movl %edi, (%rsi)
+; CHECK-DAG: movss %xmm0, 4(%rsi)
 define void @int32_float_pair(i32 %tmp1, float %tmp2, i64* %ref.tmp) {
 entry:
   %t0 = bitcast float %tmp2 to i32
@@ -15,8 +15,8 @@ entry:
 }
 
 ; CHECK-LABEL: float_int32_pair
-; CHECK: movss %xmm0, (%rsi)
-; CHECK: movl %edi, 4(%rsi)
+; CHECK-DAG: movss %xmm0, (%rsi)
+; CHECK-DAG: movl %edi, 4(%rsi)
 define void @float_int32_pair(float %tmp1, i32 %tmp2, i64* %ref.tmp) {
 entry:
   %t0 = bitcast float %tmp1 to i32
@@ -29,9 +29,9 @@ entry:
 }
 
 ; CHECK-LABEL: int16_float_pair
-; CHECK: movzwl	%di, %eax
-; CHECK: movl %eax, (%rsi)
-; CHECK: movss %xmm0, 4(%rsi)
+; CHECK-DAG: movzwl	%di, %eax
+; CHECK-DAG: movl %eax, (%rsi)
+; CHECK-DAG: movss %xmm0, 4(%rsi)
 define void @int16_float_pair(i16 signext %tmp1, float %tmp2, i64* %ref.tmp) {
 entry:
   %t0 = bitcast float %tmp2 to i32
@@ -44,9 +44,9 @@ entry:
 }
 
 ; CHECK-LABEL: int8_float_pair
-; CHECK: movzbl	%dil, %eax
-; CHECK: movl %eax, (%rsi)
-; CHECK: movss %xmm0, 4(%rsi)
+; CHECK-DAG: movzbl	%dil, %eax
+; CHECK-DAG: movl %eax, (%rsi)
+; CHECK-DAG: movss %xmm0, 4(%rsi)
 define void @int8_float_pair(i8 signext %tmp1, float %tmp2, i64* %ref.tmp) {
 entry:
   %t0 = bitcast float %tmp2 to i32
diff --git a/llvm/test/CodeGen/X86/stores-merging.ll b/llvm/test/CodeGen/X86/stores-merging.ll
index 9e479bd71b9..dbfb06881d8 100644
--- a/llvm/test/CodeGen/X86/stores-merging.ll
+++ b/llvm/test/CodeGen/X86/stores-merging.ll
@@ -13,9 +13,9 @@ target triple = "x86_64-unknown-linux-gnu"
 ;; the same result in memory in the end.
 
 ; CHECK-LABEL: redundant_stores_merging:
-; CHECK:   movl    $123, e+8(%rip)
-; CHECK:   movabsq $1958505086977, %rax
+; CHECK:   movabsq $528280977409, %rax
 ; CHECK:   movq    %rax, e+4(%rip)
+; CHECK:   movl    $456, e+8(%rip)
 define void @redundant_stores_merging() {
 entry:
   store i32 1, i32* getelementptr inbounds (%structTy, %structTy* @e, i64 0, i32 1), align 4
@@ -26,9 +26,9 @@ entry:
 
 ;; This variant tests PR25154.
 ; CHECK-LABEL: redundant_stores_merging_reverse:
-; CHECK:   movl    $123, e+8(%rip)
-; CHECK:   movabsq $1958505086977, %rax
+; CHECK:   movabsq $528280977409, %rax
 ; CHECK:   movq    %rax, e+4(%rip)
+; CHECK:   movl    $456, e+8(%rip)
 define void @redundant_stores_merging_reverse() {
 entry:
   store i32 123, i32* getelementptr inbounds (%structTy, %structTy* @e, i64 0, i32 2), align 4
@@ -45,9 +45,8 @@ entry:
 ;; a movl, after the store to 3).
 
 ;; CHECK-LABEL: overlapping_stores_merging:
-;; CHECK:  movw    $0, b+2(%rip)
+;; CHECK:  movl    $1, b(%rip)
 ;; CHECK:  movw    $2, b+3(%rip)
-;; CHECK:  movw    $1, b(%rip)
 define void @overlapping_stores_merging() {
 entry:
   store i16 0, i16* bitcast (i8* getelementptr inbounds ([8 x i8], [8 x i8]* @b, i64 0, i64 2) to i16*), align 2
diff --git a/llvm/test/CodeGen/X86/vector-compare-results.ll b/llvm/test/CodeGen/X86/vector-compare-results.ll
index 81112d25b32..213a3ec1bc8 100644
--- a/llvm/test/CodeGen/X86/vector-compare-results.ll
+++ b/llvm/test/CodeGen/X86/vector-compare-results.ll
@@ -327,98 +327,98 @@ define <32 x i1> @test_cmp_v32i8(<32 x i8> %a0, <32 x i8> %a1) nounwind {
 ; SSE2-NEXT:    movb -{{[0-9]+}}(%rsp), %al
 ; SSE2-NEXT:    andb $1, %al
 ; SSE2-NEXT:    movb %al, 2(%rdi)
-; SSE2-NEXT:    movdqa %xmm0, -{{[0-9]+}}(%rsp)
-; SSE2-NEXT:    movb -{{[0-9]+}}(%rsp), %al
-; SSE2-NEXT:    andb $1, %al
-; SSE2-NEXT:    movb %al, (%rdi)
 ; SSE2-NEXT:    movb -{{[0-9]+}}(%rsp), %al
 ; SSE2-NEXT:    andb $1, %al
 ; SSE2-NEXT:    movb %al, 2(%rdi)
 ; SSE2-NEXT:    movb -{{[0-9]+}}(%rsp), %al
 ; SSE2-NEXT:    andb $1, %al
-; SSE2-NEXT:    movb %al, (%rdi)
-; SSE2-NEXT:    movb -{{[0-9]+}}(%rsp), %al
-; SSE2-NEXT:    andb $1, %al
 ; SSE2-NEXT:    movb %al, 2(%rdi)
 ; SSE2-NEXT:    movb -{{[0-9]+}}(%rsp), %al
 ; SSE2-NEXT:    andb $1, %al
-; SSE2-NEXT:    movb %al, (%rdi)
+; SSE2-NEXT:    movb %al, 2(%rdi)
 ; SSE2-NEXT:    movb -{{[0-9]+}}(%rsp), %al
 ; SSE2-NEXT:    andb $1, %al
 ; SSE2-NEXT:    movb %al, 2(%rdi)
 ; SSE2-NEXT:    movb -{{[0-9]+}}(%rsp), %al
 ; SSE2-NEXT:    andb $1, %al
-; SSE2-NEXT:    movb %al, (%rdi)
+; SSE2-NEXT:    movb %al, 2(%rdi)
 ; SSE2-NEXT:    movb -{{[0-9]+}}(%rsp), %al
 ; SSE2-NEXT:    andb $1, %al
 ; SSE2-NEXT:    movb %al, 2(%rdi)
 ; SSE2-NEXT:    movb -{{[0-9]+}}(%rsp), %al
 ; SSE2-NEXT:    andb $1, %al
-; SSE2-NEXT:    movb %al, (%rdi)
+; SSE2-NEXT:    movb %al, 2(%rdi)
 ; SSE2-NEXT:    movb -{{[0-9]+}}(%rsp), %al
 ; SSE2-NEXT:    andb $1, %al
 ; SSE2-NEXT:    movb %al, 2(%rdi)
 ; SSE2-NEXT:    movb -{{[0-9]+}}(%rsp), %al
 ; SSE2-NEXT:    andb $1, %al
-; SSE2-NEXT:    movb %al, (%rdi)
+; SSE2-NEXT:    movb %al, 2(%rdi)
 ; SSE2-NEXT:    movb -{{[0-9]+}}(%rsp), %al
 ; SSE2-NEXT:    andb $1, %al
 ; SSE2-NEXT:    movb %al, 2(%rdi)
 ; SSE2-NEXT:    movb -{{[0-9]+}}(%rsp), %al
 ; SSE2-NEXT:    andb $1, %al
-; SSE2-NEXT:    movb %al, (%rdi)
+; SSE2-NEXT:    movb %al, 2(%rdi)
 ; SSE2-NEXT:    movb -{{[0-9]+}}(%rsp), %al
 ; SSE2-NEXT:    andb $1, %al
 ; SSE2-NEXT:    movb %al, 2(%rdi)
 ; SSE2-NEXT:    movb -{{[0-9]+}}(%rsp), %al
 ; SSE2-NEXT:    andb $1, %al
-; SSE2-NEXT:    movb %al, (%rdi)
+; SSE2-NEXT:    movb %al, 2(%rdi)
 ; SSE2-NEXT:    movb -{{[0-9]+}}(%rsp), %al
+; SSE2-NEXT:    movb -{{[0-9]+}}(%rsp), %cl
+; SSE2-NEXT:    andb $1, %cl
+; SSE2-NEXT:    movb %cl, 2(%rdi)
 ; SSE2-NEXT:    andb $1, %al
 ; SSE2-NEXT:    movb %al, 2(%rdi)
+; SSE2-NEXT:    movdqa %xmm0, -{{[0-9]+}}(%rsp)
 ; SSE2-NEXT:    movb -{{[0-9]+}}(%rsp), %al
 ; SSE2-NEXT:    andb $1, %al
 ; SSE2-NEXT:    movb %al, (%rdi)
 ; SSE2-NEXT:    movb -{{[0-9]+}}(%rsp), %al
 ; SSE2-NEXT:    andb $1, %al
-; SSE2-NEXT:    movb %al, 2(%rdi)
+; SSE2-NEXT:    movb %al, (%rdi)
 ; SSE2-NEXT:    movb -{{[0-9]+}}(%rsp), %al
 ; SSE2-NEXT:    andb $1, %al
 ; SSE2-NEXT:    movb %al, (%rdi)
 ; SSE2-NEXT:    movb -{{[0-9]+}}(%rsp), %al
 ; SSE2-NEXT:    andb $1, %al
-; SSE2-NEXT:    movb %al, 2(%rdi)
+; SSE2-NEXT:    movb %al, (%rdi)
 ; SSE2-NEXT:    movb -{{[0-9]+}}(%rsp), %al
 ; SSE2-NEXT:    andb $1, %al
 ; SSE2-NEXT:    movb %al, (%rdi)
 ; SSE2-NEXT:    movb -{{[0-9]+}}(%rsp), %al
 ; SSE2-NEXT:    andb $1, %al
-; SSE2-NEXT:    movb %al, 2(%rdi)
+; SSE2-NEXT:    movb %al, (%rdi)
 ; SSE2-NEXT:    movb -{{[0-9]+}}(%rsp), %al
 ; SSE2-NEXT:    andb $1, %al
 ; SSE2-NEXT:    movb %al, (%rdi)
 ; SSE2-NEXT:    movb -{{[0-9]+}}(%rsp), %al
 ; SSE2-NEXT:    andb $1, %al
-; SSE2-NEXT:    movb %al, 2(%rdi)
+; SSE2-NEXT:    movb %al, (%rdi)
 ; SSE2-NEXT:    movb -{{[0-9]+}}(%rsp), %al
 ; SSE2-NEXT:    andb $1, %al
 ; SSE2-NEXT:    movb %al, (%rdi)
 ; SSE2-NEXT:    movb -{{[0-9]+}}(%rsp), %al
 ; SSE2-NEXT:    andb $1, %al
-; SSE2-NEXT:    movb %al, 2(%rdi)
+; SSE2-NEXT:    movb %al, (%rdi)
 ; SSE2-NEXT:    movb -{{[0-9]+}}(%rsp), %al
 ; SSE2-NEXT:    andb $1, %al
 ; SSE2-NEXT:    movb %al, (%rdi)
 ; SSE2-NEXT:    movb -{{[0-9]+}}(%rsp), %al
 ; SSE2-NEXT:    andb $1, %al
-; SSE2-NEXT:    movb %al, 2(%rdi)
+; SSE2-NEXT:    movb %al, (%rdi)
 ; SSE2-NEXT:    movb -{{[0-9]+}}(%rsp), %al
 ; SSE2-NEXT:    andb $1, %al
 ; SSE2-NEXT:    movb %al, (%rdi)
 ; SSE2-NEXT:    movb -{{[0-9]+}}(%rsp), %al
 ; SSE2-NEXT:    andb $1, %al
-; SSE2-NEXT:    movb %al, 2(%rdi)
+; SSE2-NEXT:    movb %al, (%rdi)
 ; SSE2-NEXT:    movb -{{[0-9]+}}(%rsp), %al
+; SSE2-NEXT:    movb -{{[0-9]+}}(%rsp), %cl
+; SSE2-NEXT:    andb $1, %cl
+; SSE2-NEXT:    movb %cl, (%rdi)
 ; SSE2-NEXT:    andb $1, %al
 ; SSE2-NEXT:    movb %al, (%rdi)
 ; SSE2-NEXT:    movq %rdi, %rax
@@ -813,98 +813,98 @@ define <32 x i1> @test_cmp_v32i16(<32 x i16> %a0, <32 x i16> %a1) nounwind {
 ; SSE2-NEXT:    movb -{{[0-9]+}}(%rsp), %al
 ; SSE2-NEXT:    andb $1, %al
 ; SSE2-NEXT:    movb %al, 2(%rdi)
-; SSE2-NEXT:    movdqa %xmm0, -{{[0-9]+}}(%rsp)
-; SSE2-NEXT:    movb -{{[0-9]+}}(%rsp), %al
-; SSE2-NEXT:    andb $1, %al
-; SSE2-NEXT:    movb %al, (%rdi)
 ; SSE2-NEXT:    movb -{{[0-9]+}}(%rsp), %al
 ; SSE2-NEXT:    andb $1, %al
 ; SSE2-NEXT:    movb %al, 2(%rdi)
 ; SSE2-NEXT:    movb -{{[0-9]+}}(%rsp), %al
 ; SSE2-NEXT:    andb $1, %al
-; SSE2-NEXT:    movb %al, (%rdi)
-; SSE2-NEXT:    movb -{{[0-9]+}}(%rsp), %al
-; SSE2-NEXT:    andb $1, %al
 ; SSE2-NEXT:    movb %al, 2(%rdi)
 ; SSE2-NEXT:    movb -{{[0-9]+}}(%rsp), %al
 ; SSE2-NEXT:    andb $1, %al
-; SSE2-NEXT:    movb %al, (%rdi)
+; SSE2-NEXT:    movb %al, 2(%rdi)
 ; SSE2-NEXT:    movb -{{[0-9]+}}(%rsp), %al
 ; SSE2-NEXT:    andb $1, %al
 ; SSE2-NEXT:    movb %al, 2(%rdi)
 ; SSE2-NEXT:    movb -{{[0-9]+}}(%rsp), %al
 ; SSE2-NEXT:    andb $1, %al
-; SSE2-NEXT:    movb %al, (%rdi)
+; SSE2-NEXT:    movb %al, 2(%rdi)
 ; SSE2-NEXT:    movb -{{[0-9]+}}(%rsp), %al
 ; SSE2-NEXT:    andb $1, %al
 ; SSE2-NEXT:    movb %al, 2(%rdi)
 ; SSE2-NEXT:    movb -{{[0-9]+}}(%rsp), %al
 ; SSE2-NEXT:    andb $1, %al
-; SSE2-NEXT:    movb %al, (%rdi)
+; SSE2-NEXT:    movb %al, 2(%rdi)
 ; SSE2-NEXT:    movb -{{[0-9]+}}(%rsp), %al
 ; SSE2-NEXT:    andb $1, %al
 ; SSE2-NEXT:    movb %al, 2(%rdi)
 ; SSE2-NEXT:    movb -{{[0-9]+}}(%rsp), %al
 ; SSE2-NEXT:    andb $1, %al
-; SSE2-NEXT:    movb %al, (%rdi)
+; SSE2-NEXT:    movb %al, 2(%rdi)
 ; SSE2-NEXT:    movb -{{[0-9]+}}(%rsp), %al
 ; SSE2-NEXT:    andb $1, %al
 ; SSE2-NEXT:    movb %al, 2(%rdi)
 ; SSE2-NEXT:    movb -{{[0-9]+}}(%rsp), %al
 ; SSE2-NEXT:    andb $1, %al
-; SSE2-NEXT:    movb %al, (%rdi)
+; SSE2-NEXT:    movb %al, 2(%rdi)
 ; SSE2-NEXT:    movb -{{[0-9]+}}(%rsp), %al
 ; SSE2-NEXT:    andb $1, %al
 ; SSE2-NEXT:    movb %al, 2(%rdi)
 ; SSE2-NEXT:    movb -{{[0-9]+}}(%rsp), %al
 ; SSE2-NEXT:    andb $1, %al
-; SSE2-NEXT:    movb %al, (%rdi)
+; SSE2-NEXT:    movb %al, 2(%rdi)
 ; SSE2-NEXT:    movb -{{[0-9]+}}(%rsp), %al
+; SSE2-NEXT:    movb -{{[0-9]+}}(%rsp), %cl
+; SSE2-NEXT:    andb $1, %cl
+; SSE2-NEXT:    movb %cl, 2(%rdi)
 ; SSE2-NEXT:    andb $1, %al
 ; SSE2-NEXT:    movb %al, 2(%rdi)
+; SSE2-NEXT:    movdqa %xmm0, -{{[0-9]+}}(%rsp)
 ; SSE2-NEXT:    movb -{{[0-9]+}}(%rsp), %al
 ; SSE2-NEXT:    andb $1, %al
 ; SSE2-NEXT:    movb %al, (%rdi)
 ; SSE2-NEXT:    movb -{{[0-9]+}}(%rsp), %al
 ; SSE2-NEXT:    andb $1, %al
-; SSE2-NEXT:    movb %al, 2(%rdi)
+; SSE2-NEXT:    movb %al, (%rdi)
 ; SSE2-NEXT:    movb -{{[0-9]+}}(%rsp), %al
 ; SSE2-NEXT:    andb $1, %al
 ; SSE2-NEXT:    movb %al, (%rdi)
 ; SSE2-NEXT:    movb -{{[0-9]+}}(%rsp), %al
 ; SSE2-NEXT:    andb $1, %al
-; SSE2-NEXT:    movb %al, 2(%rdi)
+; SSE2-NEXT:    movb %al, (%rdi)
 ; SSE2-NEXT:    movb -{{[0-9]+}}(%rsp), %al
 ; SSE2-NEXT:    andb $1, %al
 ; SSE2-NEXT:    movb %al, (%rdi)
 ; SSE2-NEXT:    movb -{{[0-9]+}}(%rsp), %al
 ; SSE2-NEXT:    andb $1, %al
-; SSE2-NEXT:    movb %al, 2(%rdi)
+; SSE2-NEXT:    movb %al, (%rdi)
 ; SSE2-NEXT:    movb -{{[0-9]+}}(%rsp), %al
 ; SSE2-NEXT:    andb $1, %al
 ; SSE2-NEXT:    movb %al, (%rdi)
 ; SSE2-NEXT:    movb -{{[0-9]+}}(%rsp), %al
 ; SSE2-NEXT:    andb $1, %al
-; SSE2-NEXT:    movb %al, 2(%rdi)
+; SSE2-NEXT:    movb %al, (%rdi)
 ; SSE2-NEXT:    movb -{{[0-9]+}}(%rsp), %al
 ; SSE2-NEXT:    andb $1, %al
 ; SSE2-NEXT:    movb %al, (%rdi)
 ; SSE2-NEXT:    movb -{{[0-9]+}}(%rsp), %al
 ; SSE2-NEXT:    andb $1, %al
-; SSE2-NEXT:    movb %al, 2(%rdi)
+; SSE2-NEXT:    movb %al, (%rdi)
 ; SSE2-NEXT:    movb -{{[0-9]+}}(%rsp), %al
 ; SSE2-NEXT:    andb $1, %al
 ; SSE2-NEXT:    movb %al, (%rdi)
 ; SSE2-NEXT:    movb -{{[0-9]+}}(%rsp), %al
 ; SSE2-NEXT:    andb $1, %al
-; SSE2-NEXT:    movb %al, 2(%rdi)
+; SSE2-NEXT:    movb %al, (%rdi)
 ; SSE2-NEXT:    movb -{{[0-9]+}}(%rsp), %al
 ; SSE2-NEXT:    andb $1, %al
 ; SSE2-NEXT:    movb %al, (%rdi)
 ; SSE2-NEXT:    movb -{{[0-9]+}}(%rsp), %al
 ; SSE2-NEXT:    andb $1, %al
-; SSE2-NEXT:    movb %al, 2(%rdi)
+; SSE2-NEXT:    movb %al, (%rdi)
 ; SSE2-NEXT:    movb -{{[0-9]+}}(%rsp), %al
+; SSE2-NEXT:    movb -{{[0-9]+}}(%rsp), %cl
+; SSE2-NEXT:    andb $1, %cl
+; SSE2-NEXT:    movb %cl, (%rdi)
 ; SSE2-NEXT:    andb $1, %al
 ; SSE2-NEXT:    movb %al, (%rdi)
 ; SSE2-NEXT:    movq %rdi, %rax
@@ -1070,196 +1070,196 @@ define <64 x i1> @test_cmp_v64i8(<64 x i8> %a0, <64 x i8> %a1) nounwind {
 ; SSE2-NEXT:    movb -{{[0-9]+}}(%rsp), %al
 ; SSE2-NEXT:    andb $1, %al
 ; SSE2-NEXT:    movb %al, 6(%rdi)
-; SSE2-NEXT:    movdqa %xmm2, -{{[0-9]+}}(%rsp)
-; SSE2-NEXT:    movb -{{[0-9]+}}(%rsp), %al
-; SSE2-NEXT:    andb $1, %al
-; SSE2-NEXT:    movb %al, 4(%rdi)
-; SSE2-NEXT:    movdqa %xmm1, -{{[0-9]+}}(%rsp)
-; SSE2-NEXT:    movb -{{[0-9]+}}(%rsp), %al
-; SSE2-NEXT:    andb $1, %al
-; SSE2-NEXT:    movb %al, 2(%rdi)
-; SSE2-NEXT:    movdqa %xmm0, -{{[0-9]+}}(%rsp)
 ; SSE2-NEXT:    movb -{{[0-9]+}}(%rsp), %al
 ; SSE2-NEXT:    andb $1, %al
-; SSE2-NEXT:    movb %al, (%rdi)
+; SSE2-NEXT:    movb %al, 6(%rdi)
 ; SSE2-NEXT:    movb -{{[0-9]+}}(%rsp), %al
 ; SSE2-NEXT:    andb $1, %al
 ; SSE2-NEXT:    movb %al, 6(%rdi)
 ; SSE2-NEXT:    movb -{{[0-9]+}}(%rsp), %al
 ; SSE2-NEXT:    andb $1, %al
-; SSE2-NEXT:    movb %al, 4(%rdi)
+; SSE2-NEXT:    movb %al, 6(%rdi)
 ; SSE2-NEXT:    movb -{{[0-9]+}}(%rsp), %al
 ; SSE2-NEXT:    andb $1, %al
-; SSE2-NEXT:    movb %al, 2(%rdi)
+; SSE2-NEXT:    movb %al, 6(%rdi)
 ; SSE2-NEXT:    movb -{{[0-9]+}}(%rsp), %al
 ; SSE2-NEXT:    andb $1, %al
-; SSE2-NEXT:    movb %al, (%rdi)
+; SSE2-NEXT:    movb %al, 6(%rdi)
 ; SSE2-NEXT:    movb -{{[0-9]+}}(%rsp), %al
 ; SSE2-NEXT:    andb $1, %al
 ; SSE2-NEXT:    movb %al, 6(%rdi)
 ; SSE2-NEXT:    movb -{{[0-9]+}}(%rsp), %al
 ; SSE2-NEXT:    andb $1, %al
-; SSE2-NEXT:    movb %al, 4(%rdi)
+; SSE2-NEXT:    movb %al, 6(%rdi)
 ; SSE2-NEXT:    movb -{{[0-9]+}}(%rsp), %al
 ; SSE2-NEXT:    andb $1, %al
-; SSE2-NEXT:    movb %al, 2(%rdi)
+; SSE2-NEXT:    movb %al, 6(%rdi)
 ; SSE2-NEXT:    movb -{{[0-9]+}}(%rsp), %al
 ; SSE2-NEXT:    andb $1, %al
-; SSE2-NEXT:    movb %al, (%rdi)
+; SSE2-NEXT:    movb %al, 6(%rdi)
 ; SSE2-NEXT:    movb -{{[0-9]+}}(%rsp), %al
 ; SSE2-NEXT:    andb $1, %al
 ; SSE2-NEXT:    movb %al, 6(%rdi)
 ; SSE2-NEXT:    movb -{{[0-9]+}}(%rsp), %al
 ; SSE2-NEXT:    andb $1, %al
-; SSE2-NEXT:    movb %al, 4(%rdi)
+; SSE2-NEXT:    movb %al, 6(%rdi)
 ; SSE2-NEXT:    movb -{{[0-9]+}}(%rsp), %al
 ; SSE2-NEXT:    andb $1, %al
-; SSE2-NEXT:    movb %al, 2(%rdi)
+; SSE2-NEXT:    movb %al, 6(%rdi)
 ; SSE2-NEXT:    movb -{{[0-9]+}}(%rsp), %al
 ; SSE2-NEXT:    andb $1, %al
-; SSE2-NEXT:    movb %al, (%rdi)
+; SSE2-NEXT:    movb %al, 6(%rdi)
 ; SSE2-NEXT:    movb -{{[0-9]+}}(%rsp), %al
+; SSE2-NEXT:    movb -{{[0-9]+}}(%rsp), %cl
+; SSE2-NEXT:    andb $1, %cl
+; SSE2-NEXT:    movb %cl, 6(%rdi)
 ; SSE2-NEXT:    andb $1, %al
 ; SSE2-NEXT:    movb %al, 6(%rdi)
+; SSE2-NEXT:    movdqa %xmm2, -{{[0-9]+}}(%rsp)
 ; SSE2-NEXT:    movb -{{[0-9]+}}(%rsp), %al
 ; SSE2-NEXT:    andb $1, %al
 ; SSE2-NEXT:    movb %al, 4(%rdi)
 ; SSE2-NEXT:    movb -{{[0-9]+}}(%rsp), %al
 ; SSE2-NEXT:    andb $1, %al
-; SSE2-NEXT:    movb %al, 2(%rdi)
-; SSE2-NEXT:    movb -{{[0-9]+}}(%rsp), %al
-; SSE2-NEXT:    andb $1, %al
-; SSE2-NEXT:    movb %al, (%rdi)
-; SSE2-NEXT:    movb -{{[0-9]+}}(%rsp), %al
-; SSE2-NEXT:    andb $1, %al
-; SSE2-NEXT:    movb %al, 6(%rdi)
+; SSE2-NEXT:    movb %al, 4(%rdi)
 ; SSE2-NEXT:    movb -{{[0-9]+}}(%rsp), %al
 ; SSE2-NEXT:    andb $1, %al
 ; SSE2-NEXT:    movb %al, 4(%rdi)
 ; SSE2-NEXT:    movb -{{[0-9]+}}(%rsp), %al
 ; SSE2-NEXT:    andb $1, %al
-; SSE2-NEXT:    movb %al, 2(%rdi)
+; SSE2-NEXT:    movb %al, 4(%rdi)
 ; SSE2-NEXT:    movb -{{[0-9]+}}(%rsp), %al
 ; SSE2-NEXT:    andb $1, %al
-; SSE2-NEXT:    movb %al, (%rdi)
+; SSE2-NEXT:    movb %al, 4(%rdi)
 ; SSE2-NEXT:    movb -{{[0-9]+}}(%rsp), %al
 ; SSE2-NEXT:    andb $1, %al
-; SSE2-NEXT:    movb %al, 6(%rdi)
+; SSE2-NEXT:    movb %al, 4(%rdi)
 ; SSE2-NEXT:    movb -{{[0-9]+}}(%rsp), %al
 ; SSE2-NEXT:    andb $1, %al
 ; SSE2-NEXT:    movb %al, 4(%rdi)
 ; SSE2-NEXT:    movb -{{[0-9]+}}(%rsp), %al
 ; SSE2-NEXT:    andb $1, %al
-; SSE2-NEXT:    movb %al, 2(%rdi)
+; SSE2-NEXT:    movb %al, 4(%rdi)
 ; SSE2-NEXT:    movb -{{[0-9]+}}(%rsp), %al
 ; SSE2-NEXT:    andb $1, %al
-; SSE2-NEXT:    movb %al, (%rdi)
+; SSE2-NEXT:    movb %al, 4(%rdi)
 ; SSE2-NEXT:    movb -{{[0-9]+}}(%rsp), %al
 ; SSE2-NEXT:    andb $1, %al
-; SSE2-NEXT:    movb %al, 6(%rdi)
+; SSE2-NEXT:    movb %al, 4(%rdi)
 ; SSE2-NEXT:    movb -{{[0-9]+}}(%rsp), %al
 ; SSE2-NEXT:    andb $1, %al
 ; SSE2-NEXT:    movb %al, 4(%rdi)
 ; SSE2-NEXT:    movb -{{[0-9]+}}(%rsp), %al
 ; SSE2-NEXT:    andb $1, %al
-; SSE2-NEXT:    movb %al, 2(%rdi)
+; SSE2-NEXT:    movb %al, 4(%rdi)
 ; SSE2-NEXT:    movb -{{[0-9]+}}(%rsp), %al
 ; SSE2-NEXT:    andb $1, %al
-; SSE2-NEXT:    movb %al, (%rdi)
+; SSE2-NEXT:    movb %al, 4(%rdi)
 ; SSE2-NEXT:    movb -{{[0-9]+}}(%rsp), %al
 ; SSE2-NEXT:    andb $1, %al
-; SSE2-NEXT:    movb %al, 6(%rdi)
+; SSE2-NEXT:    movb %al, 4(%rdi)
 ; SSE2-NEXT:    movb -{{[0-9]+}}(%rsp), %al
+; SSE2-NEXT:    movb -{{[0-9]+}}(%rsp), %cl
+; SSE2-NEXT:    andb $1, %cl
+; SSE2-NEXT:    movb %cl, 4(%rdi)
 ; SSE2-NEXT:    andb $1, %al
 ; SSE2-NEXT:    movb %al, 4(%rdi)
+; SSE2-NEXT:    movdqa %xmm1, -{{[0-9]+}}(%rsp)
 ; SSE2-NEXT:    movb -{{[0-9]+}}(%rsp), %al
 ; SSE2-NEXT:    andb $1, %al
 ; SSE2-NEXT:    movb %al, 2(%rdi)
 ; SSE2-NEXT:    movb -{{[0-9]+}}(%rsp), %al
 ; SSE2-NEXT:    andb $1, %al
-; SSE2-NEXT:    movb %al, (%rdi)
+; SSE2-NEXT:    movb %al, 2(%rdi)
 ; SSE2-NEXT:    movb -{{[0-9]+}}(%rsp), %al
 ; SSE2-NEXT:    andb $1, %al
-; SSE2-NEXT:    movb %al, 6(%rdi)
+; SSE2-NEXT:    movb %al, 2(%rdi)
 ; SSE2-NEXT:    movb -{{[0-9]+}}(%rsp), %al
 ; SSE2-NEXT:    andb $1, %al
-; SSE2-NEXT:    movb %al, 4(%rdi)
+; SSE2-NEXT:    movb %al, 2(%rdi)
 ; SSE2-NEXT:    movb -{{[0-9]+}}(%rsp), %al
 ; SSE2-NEXT:    andb $1, %al
 ; SSE2-NEXT:    movb %al, 2(%rdi)
 ; SSE2-NEXT:    movb -{{[0-9]+}}(%rsp), %al
 ; SSE2-NEXT:    andb $1, %al
-; SSE2-NEXT:    movb %al, (%rdi)
+; SSE2-NEXT:    movb %al, 2(%rdi)
 ; SSE2-NEXT:    movb -{{[0-9]+}}(%rsp), %al
 ; SSE2-NEXT:    andb $1, %al
-; SSE2-NEXT:    movb %al, 6(%rdi)
+; SSE2-NEXT:    movb %al, 2(%rdi)
 ; SSE2-NEXT:    movb -{{[0-9]+}}(%rsp), %al
 ; SSE2-NEXT:    andb $1, %al
-; SSE2-NEXT:    movb %al, 4(%rdi)
+; SSE2-NEXT:    movb %al, 2(%rdi)
 ; SSE2-NEXT:    movb -{{[0-9]+}}(%rsp), %al
 ; SSE2-NEXT:    andb $1, %al
 ; SSE2-NEXT:    movb %al, 2(%rdi)
 ; SSE2-NEXT:    movb -{{[0-9]+}}(%rsp), %al
 ; SSE2-NEXT:    andb $1, %al
-; SSE2-NEXT:    movb %al, (%rdi)
+; SSE2-NEXT:    movb %al, 2(%rdi)
 ; SSE2-NEXT:    movb -{{[0-9]+}}(%rsp), %al
 ; SSE2-NEXT:    andb $1, %al
-; SSE2-NEXT:    movb %al, 6(%rdi)
+; SSE2-NEXT:    movb %al, 2(%rdi)
 ; SSE2-NEXT:    movb -{{[0-9]+}}(%rsp), %al
 ; SSE2-NEXT:    andb $1, %al
-; SSE2-NEXT:    movb %al, 4(%rdi)
+; SSE2-NEXT:    movb %al, 2(%rdi)
 ; SSE2-NEXT:    movb -{{[0-9]+}}(%rsp), %al
 ; SSE2-NEXT:    andb $1, %al
 ; SSE2-NEXT:    movb %al, 2(%rdi)
 ; SSE2-NEXT:    movb -{{[0-9]+}}(%rsp), %al
 ; SSE2-NEXT:    andb $1, %al
-; SSE2-NEXT:    movb %al, (%rdi)
+; SSE2-NEXT:    movb %al, 2(%rdi)
 ; SSE2-NEXT:    movb -{{[0-9]+}}(%rsp), %al
+; SSE2-NEXT:    movb -{{[0-9]+}}(%rsp), %cl
+; SSE2-NEXT:    andb $1, %cl
+; SSE2-NEXT:    movb %cl, 2(%rdi)
 ; SSE2-NEXT:    andb $1, %al
-; SSE2-NEXT:    movb %al, 6(%rdi)
+; SSE2-NEXT:    movb %al, 2(%rdi)
+; SSE2-NEXT:    movdqa %xmm0, -{{[0-9]+}}(%rsp)
 ; SSE2-NEXT:    movb -{{[0-9]+}}(%rsp), %al
 ; SSE2-NEXT:    andb $1, %al
-; SSE2-NEXT:    movb %al, 4(%rdi)
+; SSE2-NEXT:    movb %al, (%rdi)
 ; SSE2-NEXT:    movb -{{[0-9]+}}(%rsp), %al
 ; SSE2-NEXT:    andb $1, %al
-; SSE2-NEXT:    movb %al, 2(%rdi)
+; SSE2-NEXT:    movb %al, (%rdi)
 ; SSE2-NEXT:    movb -{{[0-9]+}}(%rsp), %al
 ; SSE2-NEXT:    andb $1, %al
 ; SSE2-NEXT:    movb %al, (%rdi)
 ; SSE2-NEXT:    movb -{{[0-9]+}}(%rsp), %al
 ; SSE2-NEXT:    andb $1, %al
-; SSE2-NEXT:    movb %al, 6(%rdi)
+; SSE2-NEXT:    movb %al, (%rdi)
 ; SSE2-NEXT:    movb -{{[0-9]+}}(%rsp), %al
 ; SSE2-NEXT:    andb $1, %al
-; SSE2-NEXT:    movb %al, 4(%rdi)
+; SSE2-NEXT:    movb %al, (%rdi)
 ; SSE2-NEXT:    movb -{{[0-9]+}}(%rsp), %al
 ; SSE2-NEXT:    andb $1, %al
-; SSE2-NEXT:    movb %al, 2(%rdi)
+; SSE2-NEXT:    movb %al, (%rdi)
 ; SSE2-NEXT:    movb -{{[0-9]+}}(%rsp), %al
 ; SSE2-NEXT:    andb $1, %al
 ; SSE2-NEXT:    movb %al, (%rdi)
 ; SSE2-NEXT:    movb -{{[0-9]+}}(%rsp), %al
 ; SSE2-NEXT:    andb $1, %al
-; SSE2-NEXT:    movb %al, 6(%rdi)
+; SSE2-NEXT:    movb %al, (%rdi)
 ; SSE2-NEXT:    movb -{{[0-9]+}}(%rsp), %al
 ; SSE2-NEXT:    andb $1, %al
-; SSE2-NEXT:    movb %al, 4(%rdi)
+; SSE2-NEXT:    movb %al, (%rdi)
 ; SSE2-NEXT:    movb -{{[0-9]+}}(%rsp), %al
 ; SSE2-NEXT:    andb $1, %al
-; SSE2-NEXT:    movb %al, 2(%rdi)
+; SSE2-NEXT:    movb %al, (%rdi)
 ; SSE2-NEXT:    movb -{{[0-9]+}}(%rsp), %al
 ; SSE2-NEXT:    andb $1, %al
 ; SSE2-NEXT:    movb %al, (%rdi)
 ; SSE2-NEXT:    movb -{{[0-9]+}}(%rsp), %al
 ; SSE2-NEXT:    andb $1, %al
-; SSE2-NEXT:    movb %al, 6(%rdi)
+; SSE2-NEXT:    movb %al, (%rdi)
 ; SSE2-NEXT:    movb -{{[0-9]+}}(%rsp), %al
 ; SSE2-NEXT:    andb $1, %al
-; SSE2-NEXT:    movb %al, 4(%rdi)
+; SSE2-NEXT:    movb %al, (%rdi)
 ; SSE2-NEXT:    movb -{{[0-9]+}}(%rsp), %al
 ; SSE2-NEXT:    andb $1, %al
-; SSE2-NEXT:    movb %al, 2(%rdi)
+; SSE2-NEXT:    movb %al, (%rdi)
 ; SSE2-NEXT:    movb -{{[0-9]+}}(%rsp), %al
+; SSE2-NEXT:    movb -{{[0-9]+}}(%rsp), %cl
+; SSE2-NEXT:    andb $1, %cl
+; SSE2-NEXT:    movb %cl, (%rdi)
 ; SSE2-NEXT:    andb $1, %al
 ; SSE2-NEXT:    movb %al, (%rdi)
 ; SSE2-NEXT:    movq %rdi, %rax
@@ -2153,98 +2153,98 @@ define <32 x i1> @test_cmp_v32f32(<32 x float> %a0, <32 x float> %a1) nounwind {
 ; SSE2-NEXT:    movb -{{[0-9]+}}(%rsp), %al
 ; SSE2-NEXT:    andb $1, %al
 ; SSE2-NEXT:    movb %al, 2(%rdi)
-; SSE2-NEXT:    movdqa %xmm0, -{{[0-9]+}}(%rsp)
-; SSE2-NEXT:    movb -{{[0-9]+}}(%rsp), %al
-; SSE2-NEXT:    andb $1, %al
-; SSE2-NEXT:    movb %al, (%rdi)
 ; SSE2-NEXT:    movb -{{[0-9]+}}(%rsp), %al
 ; SSE2-NEXT:    andb $1, %al
 ; SSE2-NEXT:    movb %al, 2(%rdi)
 ; SSE2-NEXT:    movb -{{[0-9]+}}(%rsp), %al
 ; SSE2-NEXT:    andb $1, %al
-; SSE2-NEXT:    movb %al, (%rdi)
-; SSE2-NEXT:    movb -{{[0-9]+}}(%rsp), %al
-; SSE2-NEXT:    andb $1, %al
 ; SSE2-NEXT:    movb %al, 2(%rdi)
 ; SSE2-NEXT:    movb -{{[0-9]+}}(%rsp), %al
 ; SSE2-NEXT:    andb $1, %al
-; SSE2-NEXT:    movb %al, (%rdi)
+; SSE2-NEXT:    movb %al, 2(%rdi)
 ; SSE2-NEXT:    movb -{{[0-9]+}}(%rsp), %al
 ; SSE2-NEXT:    andb $1, %al
 ; SSE2-NEXT:    movb %al, 2(%rdi)
 ; SSE2-NEXT:    movb -{{[0-9]+}}(%rsp), %al
 ; SSE2-NEXT:    andb $1, %al
-; SSE2-NEXT:    movb %al, (%rdi)
+; SSE2-NEXT:    movb %al, 2(%rdi)
 ; SSE2-NEXT:    movb -{{[0-9]+}}(%rsp), %al
 ; SSE2-NEXT:    andb $1, %al
 ; SSE2-NEXT:    movb %al, 2(%rdi)
 ; SSE2-NEXT:    movb -{{[0-9]+}}(%rsp), %al
 ; SSE2-NEXT:    andb $1, %al
-; SSE2-NEXT:    movb %al, (%rdi)
+; SSE2-NEXT:    movb %al, 2(%rdi)
 ; SSE2-NEXT:    movb -{{[0-9]+}}(%rsp), %al
 ; SSE2-NEXT:    andb $1, %al
 ; SSE2-NEXT:    movb %al, 2(%rdi)
 ; SSE2-NEXT:    movb -{{[0-9]+}}(%rsp), %al
 ; SSE2-NEXT:    andb $1, %al
-; SSE2-NEXT:    movb %al, (%rdi)
+; SSE2-NEXT:    movb %al, 2(%rdi)
 ; SSE2-NEXT:    movb -{{[0-9]+}}(%rsp), %al
 ; SSE2-NEXT:    andb $1, %al
 ; SSE2-NEXT:    movb %al, 2(%rdi)
 ; SSE2-NEXT:    movb -{{[0-9]+}}(%rsp), %al
 ; SSE2-NEXT:    andb $1, %al
-; SSE2-NEXT:    movb %al, (%rdi)
+; SSE2-NEXT:    movb %al, 2(%rdi)
 ; SSE2-NEXT:    movb -{{[0-9]+}}(%rsp), %al
 ; SSE2-NEXT:    andb $1, %al
 ; SSE2-NEXT:    movb %al, 2(%rdi)
 ; SSE2-NEXT:    movb -{{[0-9]+}}(%rsp), %al
 ; SSE2-NEXT:    andb $1, %al
-; SSE2-NEXT:    movb %al, (%rdi)
+; SSE2-NEXT:    movb %al, 2(%rdi)
 ; SSE2-NEXT:    movb -{{[0-9]+}}(%rsp), %al
+; SSE2-NEXT:    movb -{{[0-9]+}}(%rsp), %cl
+; SSE2-NEXT:    andb $1, %cl
+; SSE2-NEXT:    movb %cl, 2(%rdi)
 ; SSE2-NEXT:    andb $1, %al
 ; SSE2-NEXT:    movb %al, 2(%rdi)
+; SSE2-NEXT:    movdqa %xmm0, -{{[0-9]+}}(%rsp)
 ; SSE2-NEXT:    movb -{{[0-9]+}}(%rsp), %al
 ; SSE2-NEXT:    andb $1, %al
 ; SSE2-NEXT:    movb %al, (%rdi)
 ; SSE2-NEXT:    movb -{{[0-9]+}}(%rsp), %al
 ; SSE2-NEXT:    andb $1, %al
-; SSE2-NEXT:    movb %al, 2(%rdi)
+; SSE2-NEXT:    movb %al, (%rdi)
 ; SSE2-NEXT:    movb -{{[0-9]+}}(%rsp), %al
 ; SSE2-NEXT:    andb $1, %al
 ; SSE2-NEXT:    movb %al, (%rdi)
 ; SSE2-NEXT:    movb -{{[0-9]+}}(%rsp), %al
 ; SSE2-NEXT:    andb $1, %al
-; SSE2-NEXT:    movb %al, 2(%rdi)
+; SSE2-NEXT:    movb %al, (%rdi)
 ; SSE2-NEXT:    movb -{{[0-9]+}}(%rsp), %al
 ; SSE2-NEXT:    andb $1, %al
 ; SSE2-NEXT:    movb %al, (%rdi)
 ; SSE2-NEXT:    movb -{{[0-9]+}}(%rsp), %al
 ; SSE2-NEXT:    andb $1, %al
-; SSE2-NEXT:    movb %al, 2(%rdi)
+; SSE2-NEXT:    movb %al, (%rdi)
 ; SSE2-NEXT:    movb -{{[0-9]+}}(%rsp), %al
 ; SSE2-NEXT:    andb $1, %al
 ; SSE2-NEXT:    movb %al, (%rdi)
 ; SSE2-NEXT:    movb -{{[0-9]+}}(%rsp), %al
 ; SSE2-NEXT:    andb $1, %al
-; SSE2-NEXT:    movb %al, 2(%rdi)
+; SSE2-NEXT:    movb %al, (%rdi)
 ; SSE2-NEXT:    movb -{{[0-9]+}}(%rsp), %al
 ; SSE2-NEXT:    andb $1, %al
 ; SSE2-NEXT:    movb %al, (%rdi)
 ; SSE2-NEXT:    movb -{{[0-9]+}}(%rsp), %al
 ; SSE2-NEXT:    andb $1, %al
-; SSE2-NEXT:    movb %al, 2(%rdi)
+; SSE2-NEXT:    movb %al, (%rdi)
 ; SSE2-NEXT:    movb -{{[0-9]+}}(%rsp), %al
 ; SSE2-NEXT:    andb $1, %al
 ; SSE2-NEXT:    movb %al, (%rdi)
 ; SSE2-NEXT:    movb -{{[0-9]+}}(%rsp), %al
 ; SSE2-NEXT:    andb $1, %al
-; SSE2-NEXT:    movb %al, 2(%rdi)
+; SSE2-NEXT:    movb %al, (%rdi)
 ; SSE2-NEXT:    movb -{{[0-9]+}}(%rsp), %al
 ; SSE2-NEXT:    andb $1, %al
 ; SSE2-NEXT:    movb %al, (%rdi)
 ; SSE2-NEXT:    movb -{{[0-9]+}}(%rsp), %al
 ; SSE2-NEXT:    andb $1, %al
-; SSE2-NEXT:    movb %al, 2(%rdi)
+; SSE2-NEXT:    movb %al, (%rdi)
 ; SSE2-NEXT:    movb -{{[0-9]+}}(%rsp), %al
+; SSE2-NEXT:    movb -{{[0-9]+}}(%rsp), %cl
+; SSE2-NEXT:    andb $1, %cl
+; SSE2-NEXT:    movb %cl, (%rdi)
 ; SSE2-NEXT:    andb $1, %al
 ; SSE2-NEXT:    movb %al, (%rdi)
 ; SSE2-NEXT:    movq %rdi, %rax
@@ -2989,98 +2989,98 @@ define <32 x i1> @test_cmp_v32i32(<32 x i32> %a0, <32 x i32> %a1) nounwind {
 ; SSE2-NEXT:    movb -{{[0-9]+}}(%rsp), %al
 ; SSE2-NEXT:    andb $1, %al
 ; SSE2-NEXT:    movb %al, 2(%rdi)
-; SSE2-NEXT:    movdqa %xmm0, -{{[0-9]+}}(%rsp)
-; SSE2-NEXT:    movb -{{[0-9]+}}(%rsp), %al
-; SSE2-NEXT:    andb $1, %al
-; SSE2-NEXT:    movb %al, (%rdi)
 ; SSE2-NEXT:    movb -{{[0-9]+}}(%rsp), %al
 ; SSE2-NEXT:    andb $1, %al
 ; SSE2-NEXT:    movb %al, 2(%rdi)
 ; SSE2-NEXT:    movb -{{[0-9]+}}(%rsp), %al
 ; SSE2-NEXT:    andb $1, %al
-; SSE2-NEXT:    movb %al, (%rdi)
-; SSE2-NEXT:    movb -{{[0-9]+}}(%rsp), %al
-; SSE2-NEXT:    andb $1, %al
 ; SSE2-NEXT:    movb %al, 2(%rdi)
 ; SSE2-NEXT:    movb -{{[0-9]+}}(%rsp), %al
 ; SSE2-NEXT:    andb $1, %al
-; SSE2-NEXT:    movb %al, (%rdi)
+; SSE2-NEXT:    movb %al, 2(%rdi)
 ; SSE2-NEXT:    movb -{{[0-9]+}}(%rsp), %al
 ; SSE2-NEXT:    andb $1, %al
 ; SSE2-NEXT:    movb %al, 2(%rdi)
 ; SSE2-NEXT:    movb -{{[0-9]+}}(%rsp), %al
 ; SSE2-NEXT:    andb $1, %al
-; SSE2-NEXT:    movb %al, (%rdi)
+; SSE2-NEXT:    movb %al, 2(%rdi)
 ; SSE2-NEXT:    movb -{{[0-9]+}}(%rsp), %al
 ; SSE2-NEXT:    andb $1, %al
 ; SSE2-NEXT:    movb %al, 2(%rdi)
 ; SSE2-NEXT:    movb -{{[0-9]+}}(%rsp), %al
 ; SSE2-NEXT:    andb $1, %al
-; SSE2-NEXT:    movb %al, (%rdi)
+; SSE2-NEXT:    movb %al, 2(%rdi)
 ; SSE2-NEXT:    movb -{{[0-9]+}}(%rsp), %al
 ; SSE2-NEXT:    andb $1, %al
 ; SSE2-NEXT:    movb %al, 2(%rdi)
 ; SSE2-NEXT:    movb -{{[0-9]+}}(%rsp), %al
 ; SSE2-NEXT:    andb $1, %al
-; SSE2-NEXT:    movb %al, (%rdi)
+; SSE2-NEXT:    movb %al, 2(%rdi)
 ; SSE2-NEXT:    movb -{{[0-9]+}}(%rsp), %al
 ; SSE2-NEXT:    andb $1, %al
 ; SSE2-NEXT:    movb %al, 2(%rdi)
 ; SSE2-NEXT:    movb -{{[0-9]+}}(%rsp), %al
 ; SSE2-NEXT:    andb $1, %al
-; SSE2-NEXT:    movb %al, (%rdi)
+; SSE2-NEXT:    movb %al, 2(%rdi)
 ; SSE2-NEXT:    movb -{{[0-9]+}}(%rsp), %al
 ; SSE2-NEXT:    andb $1, %al
 ; SSE2-NEXT:    movb %al, 2(%rdi)
 ; SSE2-NEXT:    movb -{{[0-9]+}}(%rsp), %al
 ; SSE2-NEXT:    andb $1, %al
-; SSE2-NEXT:    movb %al, (%rdi)
+; SSE2-NEXT:    movb %al, 2(%rdi)
 ; SSE2-NEXT:    movb -{{[0-9]+}}(%rsp), %al
+; SSE2-NEXT:    movb -{{[0-9]+}}(%rsp), %cl
+; SSE2-NEXT:    andb $1, %cl
+; SSE2-NEXT:    movb %cl, 2(%rdi)
 ; SSE2-NEXT:    andb $1, %al
 ; SSE2-NEXT:    movb %al, 2(%rdi)
+; SSE2-NEXT:    movdqa %xmm0, -{{[0-9]+}}(%rsp)
 ; SSE2-NEXT:    movb -{{[0-9]+}}(%rsp), %al
 ; SSE2-NEXT:    andb $1, %al
 ; SSE2-NEXT:    movb %al, (%rdi)
 ; SSE2-NEXT:    movb -{{[0-9]+}}(%rsp), %al
 ; SSE2-NEXT:    andb $1, %al
-; SSE2-NEXT:    movb %al, 2(%rdi)
+; SSE2-NEXT:    movb %al, (%rdi)
 ; SSE2-NEXT:    movb -{{[0-9]+}}(%rsp), %al
 ; SSE2-NEXT:    andb $1, %al
 ; SSE2-NEXT:    movb %al, (%rdi)
 ; SSE2-NEXT:    movb -{{[0-9]+}}(%rsp), %al
 ; SSE2-NEXT:    andb $1, %al
-; SSE2-NEXT:    movb %al, 2(%rdi)
+; SSE2-NEXT:    movb %al, (%rdi)
 ; SSE2-NEXT:    movb -{{[0-9]+}}(%rsp), %al
 ; SSE2-NEXT:    andb $1, %al
 ; SSE2-NEXT:    movb %al, (%rdi)
 ; SSE2-NEXT:    movb -{{[0-9]+}}(%rsp), %al
 ; SSE2-NEXT:    andb $1, %al
-; SSE2-NEXT:    movb %al, 2(%rdi)
+; SSE2-NEXT:    movb %al, (%rdi)
 ; SSE2-NEXT:    movb -{{[0-9]+}}(%rsp), %al
 ; SSE2-NEXT:    andb $1, %al
 ; SSE2-NEXT:    movb %al, (%rdi)
 ; SSE2-NEXT:    movb -{{[0-9]+}}(%rsp), %al
 ; SSE2-NEXT:    andb $1, %al
-; SSE2-NEXT:    movb %al, 2(%rdi)
+; SSE2-NEXT:    movb %al, (%rdi)
 ; SSE2-NEXT:    movb -{{[0-9]+}}(%rsp), %al
 ; SSE2-NEXT:    andb $1, %al
 ; SSE2-NEXT:    movb %al, (%rdi)
 ; SSE2-NEXT:    movb -{{[0-9]+}}(%rsp), %al
 ; SSE2-NEXT:    andb $1, %al
-; SSE2-NEXT:    movb %al, 2(%rdi)
+; SSE2-NEXT:    movb %al, (%rdi)
 ; SSE2-NEXT:    movb -{{[0-9]+}}(%rsp), %al
 ; SSE2-NEXT:    andb $1, %al
 ; SSE2-NEXT:    movb %al, (%rdi)
 ; SSE2-NEXT:    movb -{{[0-9]+}}(%rsp), %al
 ; SSE2-NEXT:    andb $1, %al
-; SSE2-NEXT:    movb %al, 2(%rdi)
+; SSE2-NEXT:    movb %al, (%rdi)
 ; SSE2-NEXT:    movb -{{[0-9]+}}(%rsp), %al
 ; SSE2-NEXT:    andb $1, %al
 ; SSE2-NEXT:    movb %al, (%rdi)
 ; SSE2-NEXT:    movb -{{[0-9]+}}(%rsp), %al
 ; SSE2-NEXT:    andb $1, %al
-; SSE2-NEXT:    movb %al, 2(%rdi)
+; SSE2-NEXT:    movb %al, (%rdi)
 ; SSE2-NEXT:    movb -{{[0-9]+}}(%rsp), %al
+; SSE2-NEXT:    movb -{{[0-9]+}}(%rsp), %cl
+; SSE2-NEXT:    andb $1, %cl
+; SSE2-NEXT:    movb %cl, (%rdi)
 ; SSE2-NEXT:    andb $1, %al
 ; SSE2-NEXT:    movb %al, (%rdi)
 ; SSE2-NEXT:    movq %rdi, %rax
@@ -3513,196 +3513,196 @@ define <64 x i1> @test_cmp_v64i16(<64 x i16> %a0, <64 x i16> %a1) nounwind {
 ; SSE2-NEXT:    movb -{{[0-9]+}}(%rsp), %al
 ; SSE2-NEXT:    andb $1, %al
 ; SSE2-NEXT:    movb %al, 6(%rdi)
-; SSE2-NEXT:    movdqa %xmm4, -{{[0-9]+}}(%rsp)
-; SSE2-NEXT:    movb -{{[0-9]+}}(%rsp), %al
-; SSE2-NEXT:    andb $1, %al
-; SSE2-NEXT:    movb %al, 4(%rdi)
-; SSE2-NEXT:    movdqa %xmm2, -{{[0-9]+}}(%rsp)
 ; SSE2-NEXT:    movb -{{[0-9]+}}(%rsp), %al
 ; SSE2-NEXT:    andb $1, %al
-; SSE2-NEXT:    movb %al, 2(%rdi)
-; SSE2-NEXT:    movdqa %xmm0, -{{[0-9]+}}(%rsp)
-; SSE2-NEXT:    movb -{{[0-9]+}}(%rsp), %al
-; SSE2-NEXT:    andb $1, %al
-; SSE2-NEXT:    movb %al, (%rdi)
+; SSE2-NEXT:    movb %al, 6(%rdi)
 ; SSE2-NEXT:    movb -{{[0-9]+}}(%rsp), %al
 ; SSE2-NEXT:    andb $1, %al
 ; SSE2-NEXT:    movb %al, 6(%rdi)
 ; SSE2-NEXT:    movb -{{[0-9]+}}(%rsp), %al
 ; SSE2-NEXT:    andb $1, %al
-; SSE2-NEXT:    movb %al, 4(%rdi)
+; SSE2-NEXT:    movb %al, 6(%rdi)
 ; SSE2-NEXT:    movb -{{[0-9]+}}(%rsp), %al
 ; SSE2-NEXT:    andb $1, %al
-; SSE2-NEXT:    movb %al, 2(%rdi)
+; SSE2-NEXT:    movb %al, 6(%rdi)
 ; SSE2-NEXT:    movb -{{[0-9]+}}(%rsp), %al
 ; SSE2-NEXT:    andb $1, %al
-; SSE2-NEXT:    movb %al, (%rdi)
+; SSE2-NEXT:    movb %al, 6(%rdi)
 ; SSE2-NEXT:    movb -{{[0-9]+}}(%rsp), %al
 ; SSE2-NEXT:    andb $1, %al
 ; SSE2-NEXT:    movb %al, 6(%rdi)
 ; SSE2-NEXT:    movb -{{[0-9]+}}(%rsp), %al
 ; SSE2-NEXT:    andb $1, %al
-; SSE2-NEXT:    movb %al, 4(%rdi)
+; SSE2-NEXT:    movb %al, 6(%rdi)
 ; SSE2-NEXT:    movb -{{[0-9]+}}(%rsp), %al
 ; SSE2-NEXT:    andb $1, %al
-; SSE2-NEXT:    movb %al, 2(%rdi)
+; SSE2-NEXT:    movb %al, 6(%rdi)
 ; SSE2-NEXT:    movb -{{[0-9]+}}(%rsp), %al
 ; SSE2-NEXT:    andb $1, %al
-; SSE2-NEXT:    movb %al, (%rdi)
+; SSE2-NEXT:    movb %al, 6(%rdi)
 ; SSE2-NEXT:    movb -{{[0-9]+}}(%rsp), %al
 ; SSE2-NEXT:    andb $1, %al
 ; SSE2-NEXT:    movb %al, 6(%rdi)
 ; SSE2-NEXT:    movb -{{[0-9]+}}(%rsp), %al
 ; SSE2-NEXT:    andb $1, %al
-; SSE2-NEXT:    movb %al, 4(%rdi)
+; SSE2-NEXT:    movb %al, 6(%rdi)
 ; SSE2-NEXT:    movb -{{[0-9]+}}(%rsp), %al
 ; SSE2-NEXT:    andb $1, %al
-; SSE2-NEXT:    movb %al, 2(%rdi)
+; SSE2-NEXT:    movb %al, 6(%rdi)
 ; SSE2-NEXT:    movb -{{[0-9]+}}(%rsp), %al
 ; SSE2-NEXT:    andb $1, %al
-; SSE2-NEXT:    movb %al, (%rdi)
+; SSE2-NEXT:    movb %al, 6(%rdi)
 ; SSE2-NEXT:    movb -{{[0-9]+}}(%rsp), %al
+; SSE2-NEXT:    movb -{{[0-9]+}}(%rsp), %cl
+; SSE2-NEXT:    andb $1, %cl
+; SSE2-NEXT:    movb %cl, 6(%rdi)
 ; SSE2-NEXT:    andb $1, %al
 ; SSE2-NEXT:    movb %al, 6(%rdi)
+; SSE2-NEXT:    movdqa %xmm4, -{{[0-9]+}}(%rsp)
 ; SSE2-NEXT:    movb -{{[0-9]+}}(%rsp), %al
 ; SSE2-NEXT:    andb $1, %al
 ; SSE2-NEXT:    movb %al, 4(%rdi)
 ; SSE2-NEXT:    movb -{{[0-9]+}}(%rsp), %al
 ; SSE2-NEXT:    andb $1, %al
-; SSE2-NEXT:    movb %al, 2(%rdi)
-; SSE2-NEXT:    movb -{{[0-9]+}}(%rsp), %al
-; SSE2-NEXT:    andb $1, %al
-; SSE2-NEXT:    movb %al, (%rdi)
-; SSE2-NEXT:    movb -{{[0-9]+}}(%rsp), %al
-; SSE2-NEXT:    andb $1, %al
-; SSE2-NEXT:    movb %al, 6(%rdi)
+; SSE2-NEXT:    movb %al, 4(%rdi)
 ; SSE2-NEXT:    movb -{{[0-9]+}}(%rsp), %al
 ; SSE2-NEXT:    andb $1, %al
 ; SSE2-NEXT:    movb %al, 4(%rdi)
 ; SSE2-NEXT:    movb -{{[0-9]+}}(%rsp), %al
 ; SSE2-NEXT:    andb $1, %al
-; SSE2-NEXT:    movb %al, 2(%rdi)
+; SSE2-NEXT:    movb %al, 4(%rdi)
 ; SSE2-NEXT:    movb -{{[0-9]+}}(%rsp), %al
 ; SSE2-NEXT:    andb $1, %al
-; SSE2-NEXT:    movb %al, (%rdi)
+; SSE2-NEXT:    movb %al, 4(%rdi)
 ; SSE2-NEXT:    movb -{{[0-9]+}}(%rsp), %al
 ; SSE2-NEXT:    andb $1, %al
-; SSE2-NEXT:    movb %al, 6(%rdi)
+; SSE2-NEXT:    movb %al, 4(%rdi)
 ; SSE2-NEXT:    movb -{{[0-9]+}}(%rsp), %al
 ; SSE2-NEXT:    andb $1, %al
 ; SSE2-NEXT:    movb %al, 4(%rdi)
 ; SSE2-NEXT:    movb -{{[0-9]+}}(%rsp), %al
 ; SSE2-NEXT:    andb $1, %al
-; SSE2-NEXT:    movb %al, 2(%rdi)
+; SSE2-NEXT:    movb %al, 4(%rdi)
 ; SSE2-NEXT:    movb -{{[0-9]+}}(%rsp), %al
 ; SSE2-NEXT:    andb $1, %al
-; SSE2-NEXT:    movb %al, (%rdi)
+; SSE2-NEXT:    movb %al, 4(%rdi)
 ; SSE2-NEXT:    movb -{{[0-9]+}}(%rsp), %al
 ; SSE2-NEXT:    andb $1, %al
-; SSE2-NEXT:    movb %al, 6(%rdi)
+; SSE2-NEXT:    movb %al, 4(%rdi)
 ; SSE2-NEXT:    movb -{{[0-9]+}}(%rsp), %al
 ; SSE2-NEXT:    andb $1, %al
 ; SSE2-NEXT:    movb %al, 4(%rdi)
 ; SSE2-NEXT:    movb -{{[0-9]+}}(%rsp), %al
 ; SSE2-NEXT:    andb $1, %al
-; SSE2-NEXT:    movb %al, 2(%rdi)
+; SSE2-NEXT:    movb %al, 4(%rdi)
 ; SSE2-NEXT:    movb -{{[0-9]+}}(%rsp), %al
 ; SSE2-NEXT:    andb $1, %al
-; SSE2-NEXT:    movb %al, (%rdi)
+; SSE2-NEXT:    movb %al, 4(%rdi)
 ; SSE2-NEXT:    movb -{{[0-9]+}}(%rsp), %al
 ; SSE2-NEXT:    andb $1, %al
-; SSE2-NEXT:    movb %al, 6(%rdi)
+; SSE2-NEXT:    movb %al, 4(%rdi)
 ; SSE2-NEXT:    movb -{{[0-9]+}}(%rsp), %al
+; SSE2-NEXT:    movb -{{[0-9]+}}(%rsp), %cl
+; SSE2-NEXT:    andb $1, %cl
+; SSE2-NEXT:    movb %cl, 4(%rdi)
 ; SSE2-NEXT:    andb $1, %al
 ; SSE2-NEXT:    movb %al, 4(%rdi)
+; SSE2-NEXT:    movdqa %xmm2, -{{[0-9]+}}(%rsp)
 ; SSE2-NEXT:    movb -{{[0-9]+}}(%rsp), %al
 ; SSE2-NEXT:    andb $1, %al
 ; SSE2-NEXT:    movb %al, 2(%rdi)
 ; SSE2-NEXT:    movb -{{[0-9]+}}(%rsp), %al
 ; SSE2-NEXT:    andb $1, %al
-; SSE2-NEXT:    movb %al, (%rdi)
+; SSE2-NEXT:    movb %al, 2(%rdi)
 ; SSE2-NEXT:    movb -{{[0-9]+}}(%rsp), %al
 ; SSE2-NEXT:    andb $1, %al
-; SSE2-NEXT:    movb %al, 6(%rdi)
+; SSE2-NEXT:    movb %al, 2(%rdi)
 ; SSE2-NEXT:    movb -{{[0-9]+}}(%rsp), %al
 ; SSE2-NEXT:    andb $1, %al
-; SSE2-NEXT:    movb %al, 4(%rdi)
+; SSE2-NEXT:    movb %al, 2(%rdi)
 ; SSE2-NEXT:    movb -{{[0-9]+}}(%rsp), %al
 ; SSE2-NEXT:    andb $1, %al
 ; SSE2-NEXT:    movb %al, 2(%rdi)
 ; SSE2-NEXT:    movb -{{[0-9]+}}(%rsp), %al
 ; SSE2-NEXT:    andb $1, %al
-; SSE2-NEXT:    movb %al, (%rdi)
+; SSE2-NEXT:    movb %al, 2(%rdi)
 ; SSE2-NEXT:    movb -{{[0-9]+}}(%rsp), %al
 ; SSE2-NEXT:    andb $1, %al
-; SSE2-NEXT:    movb %al, 6(%rdi)
+; SSE2-NEXT:    movb %al, 2(%rdi)
 ; SSE2-NEXT:    movb -{{[0-9]+}}(%rsp), %al
 ; SSE2-NEXT:    andb $1, %al
-; SSE2-NEXT:    movb %al, 4(%rdi)
+; SSE2-NEXT:    movb %al, 2(%rdi)
 ; SSE2-NEXT:    movb -{{[0-9]+}}(%rsp), %al
 ; SSE2-NEXT:    andb $1, %al
 ; SSE2-NEXT:    movb %al, 2(%rdi)
 ; SSE2-NEXT:    movb -{{[0-9]+}}(%rsp), %al
 ; SSE2-NEXT:    andb $1, %al
-; SSE2-NEXT:    movb %al, (%rdi)
+; SSE2-NEXT:    movb %al, 2(%rdi)
 ; SSE2-NEXT:    movb -{{[0-9]+}}(%rsp), %al
 ; SSE2-NEXT:    andb $1, %al
-; SSE2-NEXT:    movb %al, 6(%rdi)
+; SSE2-NEXT:    movb %al, 2(%rdi)
 ; SSE2-NEXT:    movb -{{[0-9]+}}(%rsp), %al
 ; SSE2-NEXT:    andb $1, %al
-; SSE2-NEXT:    movb %al, 4(%rdi)
+; SSE2-NEXT:    movb %al, 2(%rdi)
 ; SSE2-NEXT:    movb -{{[0-9]+}}(%rsp), %al
 ; SSE2-NEXT:    andb $1, %al
 ; SSE2-NEXT:    movb %al, 2(%rdi)
 ; SSE2-NEXT:    movb -{{[0-9]+}}(%rsp), %al
 ; SSE2-NEXT:    andb $1, %al
-; SSE2-NEXT:    movb %al, (%rdi)
+; SSE2-NEXT:    movb %al, 2(%rdi)
 ; SSE2-NEXT:    movb -{{[0-9]+}}(%rsp), %al
+; SSE2-NEXT:    movb -{{[0-9]+}}(%rsp), %cl
+; SSE2-NEXT:    andb $1, %cl
+; SSE2-NEXT:    movb %cl, 2(%rdi)
 ; SSE2-NEXT:    andb $1, %al
-; SSE2-NEXT:    movb %al, 6(%rdi)
+; SSE2-NEXT:    movb %al, 2(%rdi)
+; SSE2-NEXT:    movdqa %xmm0, -{{[0-9]+}}(%rsp)
 ; SSE2-NEXT:    movb -{{[0-9]+}}(%rsp), %al
 ; SSE2-NEXT:    andb $1, %al
-; SSE2-NEXT:    movb %al, 4(%rdi)
+; SSE2-NEXT:    movb %al, (%rdi)
 ; SSE2-NEXT:    movb -{{[0-9]+}}(%rsp), %al
 ; SSE2-NEXT:    andb $1, %al
-; SSE2-NEXT:    movb %al, 2(%rdi)
+; SSE2-NEXT:    movb %al, (%rdi)
 ; SSE2-NEXT:    movb -{{[0-9]+}}(%rsp), %al
 ; SSE2-NEXT:    andb $1, %al
 ; SSE2-NEXT:    movb %al, (%rdi)
 ; SSE2-NEXT:    movb -{{[0-9]+}}(%rsp), %al
 ; SSE2-NEXT:    andb $1, %al
-; SSE2-NEXT:    movb %al, 6(%rdi)
+; SSE2-NEXT:    movb %al, (%rdi)
 ; SSE2-NEXT:    movb -{{[0-9]+}}(%rsp), %al
 ; SSE2-NEXT:    andb $1, %al
-; SSE2-NEXT:    movb %al, 4(%rdi)
+; SSE2-NEXT:    movb %al, (%rdi)
 ; SSE2-NEXT:    movb -{{[0-9]+}}(%rsp), %al
 ; SSE2-NEXT:    andb $1, %al
-; SSE2-NEXT:    movb %al, 2(%rdi)
+; SSE2-NEXT:    movb %al, (%rdi)
 ; SSE2-NEXT:    movb -{{[0-9]+}}(%rsp), %al
 ; SSE2-NEXT:    andb $1, %al
 ; SSE2-NEXT:    movb %al, (%rdi)
 ; SSE2-NEXT:    movb -{{[0-9]+}}(%rsp), %al
 ; SSE2-NEXT:    andb $1, %al
-; SSE2-NEXT:    movb %al, 6(%rdi)
+; SSE2-NEXT:    movb %al, (%rdi)
 ; SSE2-NEXT:    movb -{{[0-9]+}}(%rsp), %al
 ; SSE2-NEXT:    andb $1, %al
-; SSE2-NEXT:    movb %al, 4(%rdi)
+; SSE2-NEXT:    movb %al, (%rdi)
 ; SSE2-NEXT:    movb -{{[0-9]+}}(%rsp), %al
 ; SSE2-NEXT:    andb $1, %al
-; SSE2-NEXT:    movb %al, 2(%rdi)
+; SSE2-NEXT:    movb %al, (%rdi)
 ; SSE2-NEXT:    movb -{{[0-9]+}}(%rsp), %al
 ; SSE2-NEXT:    andb $1, %al
 ; SSE2-NEXT:    movb %al, (%rdi)
 ; SSE2-NEXT:    movb -{{[0-9]+}}(%rsp), %al
 ; SSE2-NEXT:    andb $1, %al
-; SSE2-NEXT:    movb %al, 6(%rdi)
+; SSE2-NEXT:    movb %al, (%rdi)
 ; SSE2-NEXT:    movb -{{[0-9]+}}(%rsp), %al
 ; SSE2-NEXT:    andb $1, %al
-; SSE2-NEXT:    movb %al, 4(%rdi)
+; SSE2-NEXT:    movb %al, (%rdi)
 ; SSE2-NEXT:    movb -{{[0-9]+}}(%rsp), %al
 ; SSE2-NEXT:    andb $1, %al
-; SSE2-NEXT:    movb %al, 2(%rdi)
+; SSE2-NEXT:    movb %al, (%rdi)
 ; SSE2-NEXT:    movb -{{[0-9]+}}(%rsp), %al
+; SSE2-NEXT:    movb -{{[0-9]+}}(%rsp), %cl
+; SSE2-NEXT:    andb $1, %cl
+; SSE2-NEXT:    movb %cl, (%rdi)
 ; SSE2-NEXT:    andb $1, %al
 ; SSE2-NEXT:    movb %al, (%rdi)
 ; SSE2-NEXT:    movq %rdi, %rax
@@ -4671,392 +4671,392 @@ define <128 x i1> @test_cmp_v128i8(<128 x i8> %a0, <128 x i8> %a1) nounwind {
 ; SSE2-NEXT:    movb -{{[0-9]+}}(%rsp), %al
 ; SSE2-NEXT:    andb $1, %al
 ; SSE2-NEXT:    movb %al, 14(%rdi)
-; SSE2-NEXT:    movdqa %xmm6, -{{[0-9]+}}(%rsp)
 ; SSE2-NEXT:    movb -{{[0-9]+}}(%rsp), %al
 ; SSE2-NEXT:    andb $1, %al
-; SSE2-NEXT:    movb %al, 12(%rdi)
-; SSE2-NEXT:    movdqa %xmm5, -{{[0-9]+}}(%rsp)
-; SSE2-NEXT:    movb -{{[0-9]+}}(%rsp), %al
-; SSE2-NEXT:    andb $1, %al
-; SSE2-NEXT:    movb %al, 10(%rdi)
-; SSE2-NEXT:    movdqa %xmm4, -{{[0-9]+}}(%rsp)
-; SSE2-NEXT:    movb -{{[0-9]+}}(%rsp), %al
-; SSE2-NEXT:    andb $1, %al
-; SSE2-NEXT:    movb %al, 8(%rdi)
-; SSE2-NEXT:    movdqa %xmm3, -{{[0-9]+}}(%rsp)
+; SSE2-NEXT:    movb %al, 14(%rdi)
 ; SSE2-NEXT:    movb -{{[0-9]+}}(%rsp), %al
 ; SSE2-NEXT:    andb $1, %al
-; SSE2-NEXT:    movb %al, 6(%rdi)
-; SSE2-NEXT:    movdqa %xmm2, -{{[0-9]+}}(%rsp)
+; SSE2-NEXT:    movb %al, 14(%rdi)
 ; SSE2-NEXT:    movb -{{[0-9]+}}(%rsp), %al
 ; SSE2-NEXT:    andb $1, %al
-; SSE2-NEXT:    movb %al, 4(%rdi)
-; SSE2-NEXT:    movdqa %xmm1, -{{[0-9]+}}(%rsp)
+; SSE2-NEXT:    movb %al, 14(%rdi)
 ; SSE2-NEXT:    movb -{{[0-9]+}}(%rsp), %al
 ; SSE2-NEXT:    andb $1, %al
-; SSE2-NEXT:    movb %al, 2(%rdi)
-; SSE2-NEXT:    movdqa %xmm0, -{{[0-9]+}}(%rsp)
+; SSE2-NEXT:    movb %al, 14(%rdi)
 ; SSE2-NEXT:    movb -{{[0-9]+}}(%rsp), %al
 ; SSE2-NEXT:    andb $1, %al
-; SSE2-NEXT:    movb %al, (%rdi)
+; SSE2-NEXT:    movb %al, 14(%rdi)
 ; SSE2-NEXT:    movb -{{[0-9]+}}(%rsp), %al
 ; SSE2-NEXT:    andb $1, %al
 ; SSE2-NEXT:    movb %al, 14(%rdi)
 ; SSE2-NEXT:    movb -{{[0-9]+}}(%rsp), %al
 ; SSE2-NEXT:    andb $1, %al
-; SSE2-NEXT:    movb %al, 12(%rdi)
+; SSE2-NEXT:    movb %al, 14(%rdi)
 ; SSE2-NEXT:    movb -{{[0-9]+}}(%rsp), %al
 ; SSE2-NEXT:    andb $1, %al
-; SSE2-NEXT:    movb %al, 10(%rdi)
+; SSE2-NEXT:    movb %al, 14(%rdi)
 ; SSE2-NEXT:    movb -{{[0-9]+}}(%rsp), %al
 ; SSE2-NEXT:    andb $1, %al
-; SSE2-NEXT:    movb %al, 8(%rdi)
+; SSE2-NEXT:    movb %al, 14(%rdi)
 ; SSE2-NEXT:    movb -{{[0-9]+}}(%rsp), %al
 ; SSE2-NEXT:    andb $1, %al
-; SSE2-NEXT:    movb %al, 6(%rdi)
+; SSE2-NEXT:    movb %al, 14(%rdi)
 ; SSE2-NEXT:    movb -{{[0-9]+}}(%rsp), %al
 ; SSE2-NEXT:    andb $1, %al
-; SSE2-NEXT:    movb %al, 4(%rdi)
+; SSE2-NEXT:    movb %al, 14(%rdi)
 ; SSE2-NEXT:    movb -{{[0-9]+}}(%rsp), %al
 ; SSE2-NEXT:    andb $1, %al
-; SSE2-NEXT:    movb %al, 2(%rdi)
+; SSE2-NEXT:    movb %al, 14(%rdi)
 ; SSE2-NEXT:    movb -{{[0-9]+}}(%rsp), %al
 ; SSE2-NEXT:    andb $1, %al
-; SSE2-NEXT:    movb %al, (%rdi)
+; SSE2-NEXT:    movb %al, 14(%rdi)
 ; SSE2-NEXT:    movb -{{[0-9]+}}(%rsp), %al
+; SSE2-NEXT:    movb -{{[0-9]+}}(%rsp), %cl
+; SSE2-NEXT:    andb $1, %cl
+; SSE2-NEXT:    movb %cl, 14(%rdi)
 ; SSE2-NEXT:    andb $1, %al
 ; SSE2-NEXT:    movb %al, 14(%rdi)
+; SSE2-NEXT:    movdqa %xmm6, -{{[0-9]+}}(%rsp)
 ; SSE2-NEXT:    movb -{{[0-9]+}}(%rsp), %al
 ; SSE2-NEXT:    andb $1, %al
 ; SSE2-NEXT:    movb %al, 12(%rdi)
 ; SSE2-NEXT:    movb -{{[0-9]+}}(%rsp), %al
 ; SSE2-NEXT:    andb $1, %al
-; SSE2-NEXT:    movb %al, 10(%rdi)
-; SSE2-NEXT:    movb -{{[0-9]+}}(%rsp), %al
-; SSE2-NEXT:    andb $1, %al
-; SSE2-NEXT:    movb %al, 8(%rdi)
-; SSE2-NEXT:    movb -{{[0-9]+}}(%rsp), %al
-; SSE2-NEXT:    andb $1, %al
-; SSE2-NEXT:    movb %al, 6(%rdi)
+; SSE2-NEXT:    movb %al, 12(%rdi)
 ; SSE2-NEXT:    movb -{{[0-9]+}}(%rsp), %al
 ; SSE2-NEXT:    andb $1, %al
-; SSE2-NEXT:    movb %al, 4(%rdi)
+; SSE2-NEXT:    movb %al, 12(%rdi)
 ; SSE2-NEXT:    movb -{{[0-9]+}}(%rsp), %al
 ; SSE2-NEXT:    andb $1, %al
-; SSE2-NEXT:    movb %al, 2(%rdi)
+; SSE2-NEXT:    movb %al, 12(%rdi)
 ; SSE2-NEXT:    movb -{{[0-9]+}}(%rsp), %al
 ; SSE2-NEXT:    andb $1, %al
-; SSE2-NEXT:    movb %al, (%rdi)
+; SSE2-NEXT:    movb %al, 12(%rdi)
 ; SSE2-NEXT:    movb -{{[0-9]+}}(%rsp), %al
 ; SSE2-NEXT:    andb $1, %al
-; SSE2-NEXT:    movb %al, 14(%rdi)
+; SSE2-NEXT:    movb %al, 12(%rdi)
 ; SSE2-NEXT:    movb -{{[0-9]+}}(%rsp), %al
 ; SSE2-NEXT:    andb $1, %al
 ; SSE2-NEXT:    movb %al, 12(%rdi)
 ; SSE2-NEXT:    movb -{{[0-9]+}}(%rsp), %al
 ; SSE2-NEXT:    andb $1, %al
-; SSE2-NEXT:    movb %al, 10(%rdi)
+; SSE2-NEXT:    movb %al, 12(%rdi)
 ; SSE2-NEXT:    movb -{{[0-9]+}}(%rsp), %al
 ; SSE2-NEXT:    andb $1, %al
-; SSE2-NEXT:    movb %al, 8(%rdi)
+; SSE2-NEXT:    movb %al, 12(%rdi)
 ; SSE2-NEXT:    movb -{{[0-9]+}}(%rsp), %al
 ; SSE2-NEXT:    andb $1, %al
-; SSE2-NEXT:    movb %al, 6(%rdi)
+; SSE2-NEXT:    movb %al, 12(%rdi)
 ; SSE2-NEXT:    movb -{{[0-9]+}}(%rsp), %al
 ; SSE2-NEXT:    andb $1, %al
-; SSE2-NEXT:    movb %al, 4(%rdi)
+; SSE2-NEXT:    movb %al, 12(%rdi)
 ; SSE2-NEXT:    movb -{{[0-9]+}}(%rsp), %al
 ; SSE2-NEXT:    andb $1, %al
-; SSE2-NEXT:    movb %al, 2(%rdi)
+; SSE2-NEXT:    movb %al, 12(%rdi)
 ; SSE2-NEXT:    movb -{{[0-9]+}}(%rsp), %al
 ; SSE2-NEXT:    andb $1, %al
-; SSE2-NEXT:    movb %al, (%rdi)
+; SSE2-NEXT:    movb %al, 12(%rdi)
 ; SSE2-NEXT:    movb -{{[0-9]+}}(%rsp), %al
 ; SSE2-NEXT:    andb $1, %al
-; SSE2-NEXT:    movb %al, 14(%rdi)
+; SSE2-NEXT:    movb %al, 12(%rdi)
 ; SSE2-NEXT:    movb -{{[0-9]+}}(%rsp), %al
+; SSE2-NEXT:    movb -{{[0-9]+}}(%rsp), %cl
+; SSE2-NEXT:    andb $1, %cl
+; SSE2-NEXT:    movb %cl, 12(%rdi)
 ; SSE2-NEXT:    andb $1, %al
 ; SSE2-NEXT:    movb %al, 12(%rdi)
+; SSE2-NEXT:    movdqa %xmm5, -{{[0-9]+}}(%rsp)
 ; SSE2-NEXT:    movb -{{[0-9]+}}(%rsp), %al
 ; SSE2-NEXT:    andb $1, %al
 ; SSE2-NEXT:    movb %al, 10(%rdi)
 ; SSE2-NEXT:    movb -{{[0-9]+}}(%rsp), %al
 ; SSE2-NEXT:    andb $1, %al
-; SSE2-NEXT:    movb %al, 8(%rdi)
-; SSE2-NEXT:    movb -{{[0-9]+}}(%rsp), %al
-; SSE2-NEXT:    andb $1, %al
-; SSE2-NEXT:    movb %al, 6(%rdi)
-; SSE2-NEXT:    movb -{{[0-9]+}}(%rsp), %al
-; SSE2-NEXT:    andb $1, %al
-; SSE2-NEXT:    movb %al, 4(%rdi)
+; SSE2-NEXT:    movb %al, 10(%rdi)
 ; SSE2-NEXT:    movb -{{[0-9]+}}(%rsp), %al
 ; SSE2-NEXT:    andb $1, %al
-; SSE2-NEXT:    movb %al, 2(%rdi)
+; SSE2-NEXT:    movb %al, 10(%rdi)
 ; SSE2-NEXT:    movb -{{[0-9]+}}(%rsp), %al
 ; SSE2-NEXT:    andb $1, %al
-; SSE2-NEXT:    movb %al, (%rdi)
+; SSE2-NEXT:    movb %al, 10(%rdi)
 ; SSE2-NEXT:    movb -{{[0-9]+}}(%rsp), %al
 ; SSE2-NEXT:    andb $1, %al
-; SSE2-NEXT:    movb %al, 14(%rdi)
+; SSE2-NEXT:    movb %al, 10(%rdi)
 ; SSE2-NEXT:    movb -{{[0-9]+}}(%rsp), %al
 ; SSE2-NEXT:    andb $1, %al
-; SSE2-NEXT:    movb %al, 12(%rdi)
+; SSE2-NEXT:    movb %al, 10(%rdi)
 ; SSE2-NEXT:    movb -{{[0-9]+}}(%rsp), %al
 ; SSE2-NEXT:    andb $1, %al
 ; SSE2-NEXT:    movb %al, 10(%rdi)
 ; SSE2-NEXT:    movb -{{[0-9]+}}(%rsp), %al
 ; SSE2-NEXT:    andb $1, %al
-; SSE2-NEXT:    movb %al, 8(%rdi)
+; SSE2-NEXT:    movb %al, 10(%rdi)
 ; SSE2-NEXT:    movb -{{[0-9]+}}(%rsp), %al
 ; SSE2-NEXT:    andb $1, %al
-; SSE2-NEXT:    movb %al, 6(%rdi)
+; SSE2-NEXT:    movb %al, 10(%rdi)
 ; SSE2-NEXT:    movb -{{[0-9]+}}(%rsp), %al
 ; SSE2-NEXT:    andb $1, %al
-; SSE2-NEXT:    movb %al, 4(%rdi)
+; SSE2-NEXT:    movb %al, 10(%rdi)
 ; SSE2-NEXT:    movb -{{[0-9]+}}(%rsp), %al
 ; SSE2-NEXT:    andb $1, %al
-; SSE2-NEXT:    movb %al, 2(%rdi)
+; SSE2-NEXT:    movb %al, 10(%rdi)
 ; SSE2-NEXT:    movb -{{[0-9]+}}(%rsp), %al
 ; SSE2-NEXT:    andb $1, %al
-; SSE2-NEXT:    movb %al, (%rdi)
+; SSE2-NEXT:    movb %al, 10(%rdi)
 ; SSE2-NEXT:    movb -{{[0-9]+}}(%rsp), %al
 ; SSE2-NEXT:    andb $1, %al
-; SSE2-NEXT:    movb %al, 14(%rdi)
+; SSE2-NEXT:    movb %al, 10(%rdi)
 ; SSE2-NEXT:    movb -{{[0-9]+}}(%rsp), %al
 ; SSE2-NEXT:    andb $1, %al
-; SSE2-NEXT:    movb %al, 12(%rdi)
+; SSE2-NEXT:    movb %al, 10(%rdi)
 ; SSE2-NEXT:    movb -{{[0-9]+}}(%rsp), %al
+; SSE2-NEXT:    movb -{{[0-9]+}}(%rsp), %cl
+; SSE2-NEXT:    andb $1, %cl
+; SSE2-NEXT:    movb %cl, 10(%rdi)
 ; SSE2-NEXT:    andb $1, %al
 ; SSE2-NEXT:    movb %al, 10(%rdi)
+; SSE2-NEXT:    movdqa %xmm4, -{{[0-9]+}}(%rsp)
 ; SSE2-NEXT:    movb -{{[0-9]+}}(%rsp), %al
 ; SSE2-NEXT:    andb $1, %al
 ; SSE2-NEXT:    movb %al, 8(%rdi)
 ; SSE2-NEXT:    movb -{{[0-9]+}}(%rsp), %al
 ; SSE2-NEXT:    andb $1, %al
-; SSE2-NEXT:    movb %al, 6(%rdi)
-; SSE2-NEXT:    movb -{{[0-9]+}}(%rsp), %al
-; SSE2-NEXT:    andb $1, %al
-; SSE2-NEXT:    movb %al, 4(%rdi)
-; SSE2-NEXT:    movb -{{[0-9]+}}(%rsp), %al
-; SSE2-NEXT:    andb $1, %al
-; SSE2-NEXT:    movb %al, 2(%rdi)
+; SSE2-NEXT:    movb %al, 8(%rdi)
 ; SSE2-NEXT:    movb -{{[0-9]+}}(%rsp), %al
 ; SSE2-NEXT:    andb $1, %al
-; SSE2-NEXT:    movb %al, (%rdi)
+; SSE2-NEXT:    movb %al, 8(%rdi)
 ; SSE2-NEXT:    movb -{{[0-9]+}}(%rsp), %al
 ; SSE2-NEXT:    andb $1, %al
-; SSE2-NEXT:    movb %al, 14(%rdi)
+; SSE2-NEXT:    movb %al, 8(%rdi)
 ; SSE2-NEXT:    movb -{{[0-9]+}}(%rsp), %al
 ; SSE2-NEXT:    andb $1, %al
-; SSE2-NEXT:    movb %al, 12(%rdi)
+; SSE2-NEXT:    movb %al, 8(%rdi)
 ; SSE2-NEXT:    movb -{{[0-9]+}}(%rsp), %al
 ; SSE2-NEXT:    andb $1, %al
-; SSE2-NEXT:    movb %al, 10(%rdi)
+; SSE2-NEXT:    movb %al, 8(%rdi)
 ; SSE2-NEXT:    movb -{{[0-9]+}}(%rsp), %al
 ; SSE2-NEXT:    andb $1, %al
 ; SSE2-NEXT:    movb %al, 8(%rdi)
 ; SSE2-NEXT:    movb -{{[0-9]+}}(%rsp), %al
 ; SSE2-NEXT:    andb $1, %al
-; SSE2-NEXT:    movb %al, 6(%rdi)
+; SSE2-NEXT:    movb %al, 8(%rdi)
 ; SSE2-NEXT:    movb -{{[0-9]+}}(%rsp), %al
 ; SSE2-NEXT:    andb $1, %al
-; SSE2-NEXT:    movb %al, 4(%rdi)
+; SSE2-NEXT:    movb %al, 8(%rdi)
 ; SSE2-NEXT:    movb -{{[0-9]+}}(%rsp), %al
 ; SSE2-NEXT:    andb $1, %al
-; SSE2-NEXT:    movb %al, 2(%rdi)
+; SSE2-NEXT:    movb %al, 8(%rdi)
 ; SSE2-NEXT:    movb -{{[0-9]+}}(%rsp), %al
 ; SSE2-NEXT:    andb $1, %al
-; SSE2-NEXT:    movb %al, (%rdi)
+; SSE2-NEXT:    movb %al, 8(%rdi)
 ; SSE2-NEXT:    movb -{{[0-9]+}}(%rsp), %al
 ; SSE2-NEXT:    andb $1, %al
-; SSE2-NEXT:    movb %al, 14(%rdi)
+; SSE2-NEXT:    movb %al, 8(%rdi)
 ; SSE2-NEXT:    movb -{{[0-9]+}}(%rsp), %al
 ; SSE2-NEXT:    andb $1, %al
-; SSE2-NEXT:    movb %al, 12(%rdi)
+; SSE2-NEXT:    movb %al, 8(%rdi)
 ; SSE2-NEXT:    movb -{{[0-9]+}}(%rsp), %al
 ; SSE2-NEXT:    andb $1, %al
-; SSE2-NEXT:    movb %al, 10(%rdi)
+; SSE2-NEXT:    movb %al, 8(%rdi)
 ; SSE2-NEXT:    movb -{{[0-9]+}}(%rsp), %al
+; SSE2-NEXT:    movb -{{[0-9]+}}(%rsp), %cl
+; SSE2-NEXT:    andb $1, %cl
+; SSE2-NEXT:    movb %cl, 8(%rdi)
 ; SSE2-NEXT:    andb $1, %al
 ; SSE2-NEXT:    movb %al, 8(%rdi)
+; SSE2-NEXT:    movdqa %xmm3, -{{[0-9]+}}(%rsp)
 ; SSE2-NEXT:    movb -{{[0-9]+}}(%rsp), %al
 ; SSE2-NEXT:    andb $1, %al
 ; SSE2-NEXT:    movb %al, 6(%rdi)
 ; SSE2-NEXT:    movb -{{[0-9]+}}(%rsp), %al
 ; SSE2-NEXT:    andb $1, %al
-; SSE2-NEXT:    movb %al, 4(%rdi)
+; SSE2-NEXT:    movb %al, 6(%rdi)
 ; SSE2-NEXT:    movb -{{[0-9]+}}(%rsp), %al
 ; SSE2-NEXT:    andb $1, %al
-; SSE2-NEXT:    movb %al, 2(%rdi)
+; SSE2-NEXT:    movb %al, 6(%rdi)
 ; SSE2-NEXT:    movb -{{[0-9]+}}(%rsp), %al
 ; SSE2-NEXT:    andb $1, %al
-; SSE2-NEXT:    movb %al, (%rdi)
+; SSE2-NEXT:    movb %al, 6(%rdi)
 ; SSE2-NEXT:    movb -{{[0-9]+}}(%rsp), %al
 ; SSE2-NEXT:    andb $1, %al
-; SSE2-NEXT:    movb %al, 14(%rdi)
+; SSE2-NEXT:    movb %al, 6(%rdi)
 ; SSE2-NEXT:    movb -{{[0-9]+}}(%rsp), %al
 ; SSE2-NEXT:    andb $1, %al
-; SSE2-NEXT:    movb %al, 12(%rdi)
+; SSE2-NEXT:    movb %al, 6(%rdi)
 ; SSE2-NEXT:    movb -{{[0-9]+}}(%rsp), %al
 ; SSE2-NEXT:    andb $1, %al
-; SSE2-NEXT:    movb %al, 10(%rdi)
+; SSE2-NEXT:    movb %al, 6(%rdi)
 ; SSE2-NEXT:    movb -{{[0-9]+}}(%rsp), %al
 ; SSE2-NEXT:    andb $1, %al
-; SSE2-NEXT:    movb %al, 8(%rdi)
+; SSE2-NEXT:    movb %al, 6(%rdi)
 ; SSE2-NEXT:    movb -{{[0-9]+}}(%rsp), %al
 ; SSE2-NEXT:    andb $1, %al
 ; SSE2-NEXT:    movb %al, 6(%rdi)
 ; SSE2-NEXT:    movb -{{[0-9]+}}(%rsp), %al
 ; SSE2-NEXT:    andb $1, %al
-; SSE2-NEXT:    movb %al, 4(%rdi)
+; SSE2-NEXT:    movb %al, 6(%rdi)
 ; SSE2-NEXT:    movb -{{[0-9]+}}(%rsp), %al
 ; SSE2-NEXT:    andb $1, %al
-; SSE2-NEXT:    movb %al, 2(%rdi)
+; SSE2-NEXT:    movb %al, 6(%rdi)
 ; SSE2-NEXT:    movb -{{[0-9]+}}(%rsp), %al
 ; SSE2-NEXT:    andb $1, %al
-; SSE2-NEXT:    movb %al, (%rdi)
+; SSE2-NEXT:    movb %al, 6(%rdi)
 ; SSE2-NEXT:    movb -{{[0-9]+}}(%rsp), %al
 ; SSE2-NEXT:    andb $1, %al
-; SSE2-NEXT:    movb %al, 14(%rdi)
+; SSE2-NEXT:    movb %al, 6(%rdi)
 ; SSE2-NEXT:    movb -{{[0-9]+}}(%rsp), %al
 ; SSE2-NEXT:    andb $1, %al
-; SSE2-NEXT:    movb %al, 12(%rdi)
+; SSE2-NEXT:    movb %al, 6(%rdi)
 ; SSE2-NEXT:    movb -{{[0-9]+}}(%rsp), %al
+; SSE2-NEXT:    movb -{{[0-9]+}}(%rsp), %cl
+; SSE2-NEXT:    andb $1, %cl
+; SSE2-NEXT:    movb %cl, 6(%rdi)
 ; SSE2-NEXT:    andb $1, %al
-; SSE2-NEXT:    movb %al, 10(%rdi)
+; SSE2-NEXT:    movb %al, 6(%rdi)
+; SSE2-NEXT:    movdqa %xmm2, -{{[0-9]+}}(%rsp)
 ; SSE2-NEXT:    movb -{{[0-9]+}}(%rsp), %al
 ; SSE2-NEXT:    andb $1, %al
-; SSE2-NEXT:    movb %al, 8(%rdi)
+; SSE2-NEXT:    movb %al, 4(%rdi)
 ; SSE2-NEXT:    movb -{{[0-9]+}}(%rsp), %al
 ; SSE2-NEXT:    andb $1, %al
-; SSE2-NEXT:    movb %al, 6(%rdi)
+; SSE2-NEXT:    movb %al, 4(%rdi)
 ; SSE2-NEXT:    movb -{{[0-9]+}}(%rsp), %al
 ; SSE2-NEXT:    andb $1, %al
 ; SSE2-NEXT:    movb %al, 4(%rdi)
 ; SSE2-NEXT:    movb -{{[0-9]+}}(%rsp), %al
 ; SSE2-NEXT:    andb $1, %al
-; SSE2-NEXT:    movb %al, 2(%rdi)
+; SSE2-NEXT:    movb %al, 4(%rdi)
 ; SSE2-NEXT:    movb -{{[0-9]+}}(%rsp), %al
 ; SSE2-NEXT:    andb $1, %al
-; SSE2-NEXT:    movb %al, (%rdi)
+; SSE2-NEXT:    movb %al, 4(%rdi)
 ; SSE2-NEXT:    movb -{{[0-9]+}}(%rsp), %al
 ; SSE2-NEXT:    andb $1, %al
-; SSE2-NEXT:    movb %al, 14(%rdi)
+; SSE2-NEXT:    movb %al, 4(%rdi)
 ; SSE2-NEXT:    movb -{{[0-9]+}}(%rsp), %al
 ; SSE2-NEXT:    andb $1, %al
-; SSE2-NEXT:    movb %al, 12(%rdi)
+; SSE2-NEXT:    movb %al, 4(%rdi)
 ; SSE2-NEXT:    movb -{{[0-9]+}}(%rsp), %al
 ; SSE2-NEXT:    andb $1, %al
-; SSE2-NEXT:    movb %al, 10(%rdi)
+; SSE2-NEXT:    movb %al, 4(%rdi)
 ; SSE2-NEXT:    movb -{{[0-9]+}}(%rsp), %al
 ; SSE2-NEXT:    andb $1, %al
-; SSE2-NEXT:    movb %al, 8(%rdi)
+; SSE2-NEXT:    movb %al, 4(%rdi)
 ; SSE2-NEXT:    movb -{{[0-9]+}}(%rsp), %al
 ; SSE2-NEXT:    andb $1, %al
-; SSE2-NEXT:    movb %al, 6(%rdi)
+; SSE2-NEXT:    movb %al, 4(%rdi)
 ; SSE2-NEXT:    movb -{{[0-9]+}}(%rsp), %al
 ; SSE2-NEXT:    andb $1, %al
 ; SSE2-NEXT:    movb %al, 4(%rdi)
 ; SSE2-NEXT:    movb -{{[0-9]+}}(%rsp), %al
 ; SSE2-NEXT:    andb $1, %al
-; SSE2-NEXT:    movb %al, 2(%rdi)
+; SSE2-NEXT:    movb %al, 4(%rdi)
 ; SSE2-NEXT:    movb -{{[0-9]+}}(%rsp), %al
 ; SSE2-NEXT:    andb $1, %al
-; SSE2-NEXT:    movb %al, (%rdi)
+; SSE2-NEXT:    movb %al, 4(%rdi)
 ; SSE2-NEXT:    movb -{{[0-9]+}}(%rsp), %al
 ; SSE2-NEXT:    andb $1, %al
-; SSE2-NEXT:    movb %al, 14(%rdi)
+; SSE2-NEXT:    movb %al, 4(%rdi)
 ; SSE2-NEXT:    movb -{{[0-9]+}}(%rsp), %al
+; SSE2-NEXT:    movb -{{[0-9]+}}(%rsp), %cl
+; SSE2-NEXT:    andb $1, %cl
+; SSE2-NEXT:    movb %cl, 4(%rdi)
 ; SSE2-NEXT:    andb $1, %al
-; SSE2-NEXT:    movb %al, 12(%rdi)
+; SSE2-NEXT:    movb %al, 4(%rdi)
+; SSE2-NEXT:    movdqa %xmm1, -{{[0-9]+}}(%rsp)
 ; SSE2-NEXT:    movb -{{[0-9]+}}(%rsp), %al
 ; SSE2-NEXT:    andb $1, %al
-; SSE2-NEXT:    movb %al, 10(%rdi)
+; SSE2-NEXT:    movb %al, 2(%rdi)
 ; SSE2-NEXT:    movb -{{[0-9]+}}(%rsp), %al
 ; SSE2-NEXT:    andb $1, %al
-; SSE2-NEXT:    movb %al, 8(%rdi)
+; SSE2-NEXT:    movb %al, 2(%rdi)
 ; SSE2-NEXT:    movb -{{[0-9]+}}(%rsp), %al
 ; SSE2-NEXT:    andb $1, %al
-; SSE2-NEXT:    movb %al, 6(%rdi)
+; SSE2-NEXT:    movb %al, 2(%rdi)
 ; SSE2-NEXT:    movb -{{[0-9]+}}(%rsp), %al
 ; SSE2-NEXT:    andb $1, %al
-; SSE2-NEXT:    movb %al, 4(%rdi)
+; SSE2-NEXT:    movb %al, 2(%rdi)
 ; SSE2-NEXT:    movb -{{[0-9]+}}(%rsp), %al
 ; SSE2-NEXT:    andb $1, %al
 ; SSE2-NEXT:    movb %al, 2(%rdi)
 ; SSE2-NEXT:    movb -{{[0-9]+}}(%rsp), %al
 ; SSE2-NEXT:    andb $1, %al
-; SSE2-NEXT:    movb %al, (%rdi)
+; SSE2-NEXT:    movb %al, 2(%rdi)
 ; SSE2-NEXT:    movb -{{[0-9]+}}(%rsp), %al
 ; SSE2-NEXT:    andb $1, %al
-; SSE2-NEXT:    movb %al, 14(%rdi)
+; SSE2-NEXT:    movb %al, 2(%rdi)
 ; SSE2-NEXT:    movb -{{[0-9]+}}(%rsp), %al
 ; SSE2-NEXT:    andb $1, %al
-; SSE2-NEXT:    movb %al, 12(%rdi)
+; SSE2-NEXT:    movb %al, 2(%rdi)
 ; SSE2-NEXT:    movb -{{[0-9]+}}(%rsp), %al
 ; SSE2-NEXT:    andb $1, %al
-; SSE2-NEXT:    movb %al, 10(%rdi)
+; SSE2-NEXT:    movb %al, 2(%rdi)
 ; SSE2-NEXT:    movb -{{[0-9]+}}(%rsp), %al
 ; SSE2-NEXT:    andb $1, %al
-; SSE2-NEXT:    movb %al, 8(%rdi)
+; SSE2-NEXT:    movb %al, 2(%rdi)
 ; SSE2-NEXT:    movb -{{[0-9]+}}(%rsp), %al
 ; SSE2-NEXT:    andb $1, %al
-; SSE2-NEXT:    movb %al, 6(%rdi)
+; SSE2-NEXT:    movb %al, 2(%rdi)
 ; SSE2-NEXT:    movb -{{[0-9]+}}(%rsp), %al
 ; SSE2-NEXT:    andb $1, %al
-; SSE2-NEXT:    movb %al, 4(%rdi)
+; SSE2-NEXT:    movb %al, 2(%rdi)
 ; SSE2-NEXT:    movb -{{[0-9]+}}(%rsp), %al
 ; SSE2-NEXT:    andb $1, %al
 ; SSE2-NEXT:    movb %al, 2(%rdi)
 ; SSE2-NEXT:    movb -{{[0-9]+}}(%rsp), %al
 ; SSE2-NEXT:    andb $1, %al
-; SSE2-NEXT:    movb %al, (%rdi)
+; SSE2-NEXT:    movb %al, 2(%rdi)
 ; SSE2-NEXT:    movb -{{[0-9]+}}(%rsp), %al
+; SSE2-NEXT:    movb -{{[0-9]+}}(%rsp), %cl
+; SSE2-NEXT:    andb $1, %cl
+; SSE2-NEXT:    movb %cl, 2(%rdi)
 ; SSE2-NEXT:    andb $1, %al
-; SSE2-NEXT:    movb %al, 14(%rdi)
+; SSE2-NEXT:    movb %al, 2(%rdi)
+; SSE2-NEXT:    movdqa %xmm0, -{{[0-9]+}}(%rsp)
 ; SSE2-NEXT:    movb -{{[0-9]+}}(%rsp), %al
 ; SSE2-NEXT:    andb $1, %al
-; SSE2-NEXT:    movb %al, 12(%rdi)
+; SSE2-NEXT:    movb %al, (%rdi)
 ; SSE2-NEXT:    movb -{{[0-9]+}}(%rsp), %al
 ; SSE2-NEXT:    andb $1, %al
-; SSE2-NEXT:    movb %al, 10(%rdi)
+; SSE2-NEXT:    movb %al, (%rdi)
 ; SSE2-NEXT:    movb -{{[0-9]+}}(%rsp), %al
 ; SSE2-NEXT:    andb $1, %al
-; SSE2-NEXT:    movb %al, 8(%rdi)
+; SSE2-NEXT:    movb %al, (%rdi)
 ; SSE2-NEXT:    movb -{{[0-9]+}}(%rsp), %al
 ; SSE2-NEXT:    andb $1, %al
-; SSE2-NEXT:    movb %al, 6(%rdi)
+; SSE2-NEXT:    movb %al, (%rdi)
 ; SSE2-NEXT:    movb -{{[0-9]+}}(%rsp), %al
 ; SSE2-NEXT:    andb $1, %al
-; SSE2-NEXT:    movb %al, 4(%rdi)
+; SSE2-NEXT:    movb %al, (%rdi)
 ; SSE2-NEXT:    movb -{{[0-9]+}}(%rsp), %al
 ; SSE2-NEXT:    andb $1, %al
-; SSE2-NEXT:    movb %al, 2(%rdi)
+; SSE2-NEXT:    movb %al, (%rdi)
 ; SSE2-NEXT:    movb -{{[0-9]+}}(%rsp), %al
 ; SSE2-NEXT:    andb $1, %al
 ; SSE2-NEXT:    movb %al, (%rdi)
 ; SSE2-NEXT:    movb -{{[0-9]+}}(%rsp), %al
 ; SSE2-NEXT:    andb $1, %al
-; SSE2-NEXT:    movb %al, 14(%rdi)
+; SSE2-NEXT:    movb %al, (%rdi)
 ; SSE2-NEXT:    movb -{{[0-9]+}}(%rsp), %al
 ; SSE2-NEXT:    andb $1, %al
-; SSE2-NEXT:    movb %al, 12(%rdi)
+; SSE2-NEXT:    movb %al, (%rdi)
 ; SSE2-NEXT:    movb -{{[0-9]+}}(%rsp), %al
 ; SSE2-NEXT:    andb $1, %al
-; SSE2-NEXT:    movb %al, 10(%rdi)
+; SSE2-NEXT:    movb %al, (%rdi)
 ; SSE2-NEXT:    movb -{{[0-9]+}}(%rsp), %al
 ; SSE2-NEXT:    andb $1, %al
-; SSE2-NEXT:    movb %al, 8(%rdi)
+; SSE2-NEXT:    movb %al, (%rdi)
 ; SSE2-NEXT:    movb -{{[0-9]+}}(%rsp), %al
 ; SSE2-NEXT:    andb $1, %al
-; SSE2-NEXT:    movb %al, 6(%rdi)
+; SSE2-NEXT:    movb %al, (%rdi)
 ; SSE2-NEXT:    movb -{{[0-9]+}}(%rsp), %al
 ; SSE2-NEXT:    andb $1, %al
-; SSE2-NEXT:    movb %al, 4(%rdi)
+; SSE2-NEXT:    movb %al, (%rdi)
 ; SSE2-NEXT:    movb -{{[0-9]+}}(%rsp), %al
 ; SSE2-NEXT:    andb $1, %al
-; SSE2-NEXT:    movb %al, 2(%rdi)
+; SSE2-NEXT:    movb %al, (%rdi)
 ; SSE2-NEXT:    movb -{{[0-9]+}}(%rsp), %al
+; SSE2-NEXT:    movb -{{[0-9]+}}(%rsp), %cl
+; SSE2-NEXT:    andb $1, %cl
+; SSE2-NEXT:    movb %cl, (%rdi)
 ; SSE2-NEXT:    andb $1, %al
 ; SSE2-NEXT:    movb %al, (%rdi)
 ; SSE2-NEXT:    movq %rdi, %rax
@@ -6438,98 +6438,98 @@ define <32 x i1> @test_cmp_v32f64(<32 x double> %a0, <32 x double> %a1) nounwind
 ; SSE2-NEXT:    movb -{{[0-9]+}}(%rsp), %al
 ; SSE2-NEXT:    andb $1, %al
 ; SSE2-NEXT:    movb %al, 2(%rdi)
-; SSE2-NEXT:    movdqa %xmm0, -{{[0-9]+}}(%rsp)
-; SSE2-NEXT:    movb -{{[0-9]+}}(%rsp), %al
-; SSE2-NEXT:    andb $1, %al
-; SSE2-NEXT:    movb %al, (%rdi)
 ; SSE2-NEXT:    movb -{{[0-9]+}}(%rsp), %al
 ; SSE2-NEXT:    andb $1, %al
 ; SSE2-NEXT:    movb %al, 2(%rdi)
 ; SSE2-NEXT:    movb -{{[0-9]+}}(%rsp), %al
 ; SSE2-NEXT:    andb $1, %al
-; SSE2-NEXT:    movb %al, (%rdi)
-; SSE2-NEXT:    movb -{{[0-9]+}}(%rsp), %al
-; SSE2-NEXT:    andb $1, %al
 ; SSE2-NEXT:    movb %al, 2(%rdi)
 ; SSE2-NEXT:    movb -{{[0-9]+}}(%rsp), %al
 ; SSE2-NEXT:    andb $1, %al
-; SSE2-NEXT:    movb %al, (%rdi)
+; SSE2-NEXT:    movb %al, 2(%rdi)
 ; SSE2-NEXT:    movb -{{[0-9]+}}(%rsp), %al
 ; SSE2-NEXT:    andb $1, %al
 ; SSE2-NEXT:    movb %al, 2(%rdi)
 ; SSE2-NEXT:    movb -{{[0-9]+}}(%rsp), %al
 ; SSE2-NEXT:    andb $1, %al
-; SSE2-NEXT:    movb %al, (%rdi)
+; SSE2-NEXT:    movb %al, 2(%rdi)
 ; SSE2-NEXT:    movb -{{[0-9]+}}(%rsp), %al
 ; SSE2-NEXT:    andb $1, %al
 ; SSE2-NEXT:    movb %al, 2(%rdi)
 ; SSE2-NEXT:    movb -{{[0-9]+}}(%rsp), %al
 ; SSE2-NEXT:    andb $1, %al
-; SSE2-NEXT:    movb %al, (%rdi)
+; SSE2-NEXT:    movb %al, 2(%rdi)
 ; SSE2-NEXT:    movb -{{[0-9]+}}(%rsp), %al
 ; SSE2-NEXT:    andb $1, %al
 ; SSE2-NEXT:    movb %al, 2(%rdi)
 ; SSE2-NEXT:    movb -{{[0-9]+}}(%rsp), %al
 ; SSE2-NEXT:    andb $1, %al
-; SSE2-NEXT:    movb %al, (%rdi)
+; SSE2-NEXT:    movb %al, 2(%rdi)
 ; SSE2-NEXT:    movb -{{[0-9]+}}(%rsp), %al
 ; SSE2-NEXT:    andb $1, %al
 ; SSE2-NEXT:    movb %al, 2(%rdi)
 ; SSE2-NEXT:    movb -{{[0-9]+}}(%rsp), %al
 ; SSE2-NEXT:    andb $1, %al
-; SSE2-NEXT:    movb %al, (%rdi)
+; SSE2-NEXT:    movb %al, 2(%rdi)
 ; SSE2-NEXT:    movb -{{[0-9]+}}(%rsp), %al
 ; SSE2-NEXT:    andb $1, %al
 ; SSE2-NEXT:    movb %al, 2(%rdi)
 ; SSE2-NEXT:    movb -{{[0-9]+}}(%rsp), %al
 ; SSE2-NEXT:    andb $1, %al
-; SSE2-NEXT:    movb %al, (%rdi)
+; SSE2-NEXT:    movb %al, 2(%rdi)
 ; SSE2-NEXT:    movb -{{[0-9]+}}(%rsp), %al
+; SSE2-NEXT:    movb -{{[0-9]+}}(%rsp), %cl
+; SSE2-NEXT:    andb $1, %cl
+; SSE2-NEXT:    movb %cl, 2(%rdi)
 ; SSE2-NEXT:    andb $1, %al
 ; SSE2-NEXT:    movb %al, 2(%rdi)
+; SSE2-NEXT:    movdqa %xmm0, -{{[0-9]+}}(%rsp)
 ; SSE2-NEXT:    movb -{{[0-9]+}}(%rsp), %al
 ; SSE2-NEXT:    andb $1, %al
 ; SSE2-NEXT:    movb %al, (%rdi)
 ; SSE2-NEXT:    movb -{{[0-9]+}}(%rsp), %al
 ; SSE2-NEXT:    andb $1, %al
-; SSE2-NEXT:    movb %al, 2(%rdi)
+; SSE2-NEXT:    movb %al, (%rdi)
 ; SSE2-NEXT:    movb -{{[0-9]+}}(%rsp), %al
 ; SSE2-NEXT:    andb $1, %al
 ; SSE2-NEXT:    movb %al, (%rdi)
 ; SSE2-NEXT:    movb -{{[0-9]+}}(%rsp), %al
 ; SSE2-NEXT:    andb $1, %al
-; SSE2-NEXT:    movb %al, 2(%rdi)
+; SSE2-NEXT:    movb %al, (%rdi)
 ; SSE2-NEXT:    movb -{{[0-9]+}}(%rsp), %al
 ; SSE2-NEXT:    andb $1, %al
 ; SSE2-NEXT:    movb %al, (%rdi)
 ; SSE2-NEXT:    movb -{{[0-9]+}}(%rsp), %al
 ; SSE2-NEXT:    andb $1, %al
-; SSE2-NEXT:    movb %al, 2(%rdi)
+; SSE2-NEXT:    movb %al, (%rdi)
 ; SSE2-NEXT:    movb -{{[0-9]+}}(%rsp), %al
 ; SSE2-NEXT:    andb $1, %al
 ; SSE2-NEXT:    movb %al, (%rdi)
 ; SSE2-NEXT:    movb -{{[0-9]+}}(%rsp), %al
 ; SSE2-NEXT:    andb $1, %al
-; SSE2-NEXT:    movb %al, 2(%rdi)
+; SSE2-NEXT:    movb %al, (%rdi)
 ; SSE2-NEXT:    movb -{{[0-9]+}}(%rsp), %al
 ; SSE2-NEXT:    andb $1, %al
 ; SSE2-NEXT:    movb %al, (%rdi)
 ; SSE2-NEXT:    movb -{{[0-9]+}}(%rsp), %al
 ; SSE2-NEXT:    andb $1, %al
-; SSE2-NEXT:    movb %al, 2(%rdi)
+; SSE2-NEXT:    movb %al, (%rdi)
 ; SSE2-NEXT:    movb -{{[0-9]+}}(%rsp), %al
 ; SSE2-NEXT:    andb $1, %al
 ; SSE2-NEXT:    movb %al, (%rdi)
 ; SSE2-NEXT:    movb -{{[0-9]+}}(%rsp), %al
 ; SSE2-NEXT:    andb $1, %al
-; SSE2-NEXT:    movb %al, 2(%rdi)
+; SSE2-NEXT:    movb %al, (%rdi)
 ; SSE2-NEXT:    movb -{{[0-9]+}}(%rsp), %al
 ; SSE2-NEXT:    andb $1, %al
 ; SSE2-NEXT:    movb %al, (%rdi)
 ; SSE2-NEXT:    movb -{{[0-9]+}}(%rsp), %al
 ; SSE2-NEXT:    andb $1, %al
-; SSE2-NEXT:    movb %al, 2(%rdi)
+; SSE2-NEXT:    movb %al, (%rdi)
 ; SSE2-NEXT:    movb -{{[0-9]+}}(%rsp), %al
+; SSE2-NEXT:    movb -{{[0-9]+}}(%rsp), %cl
+; SSE2-NEXT:    andb $1, %cl
+; SSE2-NEXT:    movb %cl, (%rdi)
 ; SSE2-NEXT:    andb $1, %al
 ; SSE2-NEXT:    movb %al, (%rdi)
 ; SSE2-NEXT:    movq %rdi, %rax
@@ -7340,98 +7340,98 @@ define <32 x i1> @test_cmp_v32i64(<32 x i64> %a0, <32 x i64> %a1) nounwind {
 ; SSE2-NEXT:    movb -{{[0-9]+}}(%rsp), %al
 ; SSE2-NEXT:    andb $1, %al
 ; SSE2-NEXT:    movb %al, 2(%rdi)
-; SSE2-NEXT:    movdqa %xmm0, -{{[0-9]+}}(%rsp)
-; SSE2-NEXT:    movb -{{[0-9]+}}(%rsp), %al
-; SSE2-NEXT:    andb $1, %al
-; SSE2-NEXT:    movb %al, (%rdi)
 ; SSE2-NEXT:    movb -{{[0-9]+}}(%rsp), %al
 ; SSE2-NEXT:    andb $1, %al
 ; SSE2-NEXT:    movb %al, 2(%rdi)
 ; SSE2-NEXT:    movb -{{[0-9]+}}(%rsp), %al
 ; SSE2-NEXT:    andb $1, %al
-; SSE2-NEXT:    movb %al, (%rdi)
-; SSE2-NEXT:    movb -{{[0-9]+}}(%rsp), %al
-; SSE2-NEXT:    andb $1, %al
 ; SSE2-NEXT:    movb %al, 2(%rdi)
 ; SSE2-NEXT:    movb -{{[0-9]+}}(%rsp), %al
 ; SSE2-NEXT:    andb $1, %al
-; SSE2-NEXT:    movb %al, (%rdi)
+; SSE2-NEXT:    movb %al, 2(%rdi)
 ; SSE2-NEXT:    movb -{{[0-9]+}}(%rsp), %al
 ; SSE2-NEXT:    andb $1, %al
 ; SSE2-NEXT:    movb %al, 2(%rdi)
 ; SSE2-NEXT:    movb -{{[0-9]+}}(%rsp), %al
 ; SSE2-NEXT:    andb $1, %al
-; SSE2-NEXT:    movb %al, (%rdi)
+; SSE2-NEXT:    movb %al, 2(%rdi)
 ; SSE2-NEXT:    movb -{{[0-9]+}}(%rsp), %al
 ; SSE2-NEXT:    andb $1, %al
 ; SSE2-NEXT:    movb %al, 2(%rdi)
 ; SSE2-NEXT:    movb -{{[0-9]+}}(%rsp), %al
 ; SSE2-NEXT:    andb $1, %al
-; SSE2-NEXT:    movb %al, (%rdi)
+; SSE2-NEXT:    movb %al, 2(%rdi)
 ; SSE2-NEXT:    movb -{{[0-9]+}}(%rsp), %al
 ; SSE2-NEXT:    andb $1, %al
 ; SSE2-NEXT:    movb %al, 2(%rdi)
 ; SSE2-NEXT:    movb -{{[0-9]+}}(%rsp), %al
 ; SSE2-NEXT:    andb $1, %al
-; SSE2-NEXT:    movb %al, (%rdi)
+; SSE2-NEXT:    movb %al, 2(%rdi)
 ; SSE2-NEXT:    movb -{{[0-9]+}}(%rsp), %al
 ; SSE2-NEXT:    andb $1, %al
 ; SSE2-NEXT:    movb %al, 2(%rdi)
 ; SSE2-NEXT:    movb -{{[0-9]+}}(%rsp), %al
 ; SSE2-NEXT:    andb $1, %al
-; SSE2-NEXT:    movb %al, (%rdi)
+; SSE2-NEXT:    movb %al, 2(%rdi)
 ; SSE2-NEXT:    movb -{{[0-9]+}}(%rsp), %al
 ; SSE2-NEXT:    andb $1, %al
 ; SSE2-NEXT:    movb %al, 2(%rdi)
 ; SSE2-NEXT:    movb -{{[0-9]+}}(%rsp), %al
 ; SSE2-NEXT:    andb $1, %al
-; SSE2-NEXT:    movb %al, (%rdi)
+; SSE2-NEXT:    movb %al, 2(%rdi)
 ; SSE2-NEXT:    movb -{{[0-9]+}}(%rsp), %al
+; SSE2-NEXT:    movb -{{[0-9]+}}(%rsp), %cl
+; SSE2-NEXT:    andb $1, %cl
+; SSE2-NEXT:    movb %cl, 2(%rdi)
 ; SSE2-NEXT:    andb $1, %al
 ; SSE2-NEXT:    movb %al, 2(%rdi)
+; SSE2-NEXT:    movdqa %xmm0, -{{[0-9]+}}(%rsp)
 ; SSE2-NEXT:    movb -{{[0-9]+}}(%rsp), %al
 ; SSE2-NEXT:    andb $1, %al
 ; SSE2-NEXT:    movb %al, (%rdi)
 ; SSE2-NEXT:    movb -{{[0-9]+}}(%rsp), %al
 ; SSE2-NEXT:    andb $1, %al
-; SSE2-NEXT:    movb %al, 2(%rdi)
+; SSE2-NEXT:    movb %al, (%rdi)
 ; SSE2-NEXT:    movb -{{[0-9]+}}(%rsp), %al
 ; SSE2-NEXT:    andb $1, %al
 ; SSE2-NEXT:    movb %al, (%rdi)
 ; SSE2-NEXT:    movb -{{[0-9]+}}(%rsp), %al
 ; SSE2-NEXT:    andb $1, %al
-; SSE2-NEXT:    movb %al, 2(%rdi)
+; SSE2-NEXT:    movb %al, (%rdi)
 ; SSE2-NEXT:    movb -{{[0-9]+}}(%rsp), %al
 ; SSE2-NEXT:    andb $1, %al
 ; SSE2-NEXT:    movb %al, (%rdi)
 ; SSE2-NEXT:    movb -{{[0-9]+}}(%rsp), %al
 ; SSE2-NEXT:    andb $1, %al
-; SSE2-NEXT:    movb %al, 2(%rdi)
+; SSE2-NEXT:    movb %al, (%rdi)
 ; SSE2-NEXT:    movb -{{[0-9]+}}(%rsp), %al
 ; SSE2-NEXT:    andb $1, %al
 ; SSE2-NEXT:    movb %al, (%rdi)
 ; SSE2-NEXT:    movb -{{[0-9]+}}(%rsp), %al
 ; SSE2-NEXT:    andb $1, %al
-; SSE2-NEXT:    movb %al, 2(%rdi)
+; SSE2-NEXT:    movb %al, (%rdi)
 ; SSE2-NEXT:    movb -{{[0-9]+}}(%rsp), %al
 ; SSE2-NEXT:    andb $1, %al
 ; SSE2-NEXT:    movb %al, (%rdi)
 ; SSE2-NEXT:    movb -{{[0-9]+}}(%rsp), %al
 ; SSE2-NEXT:    andb $1, %al
-; SSE2-NEXT:    movb %al, 2(%rdi)
+; SSE2-NEXT:    movb %al, (%rdi)
 ; SSE2-NEXT:    movb -{{[0-9]+}}(%rsp), %al
 ; SSE2-NEXT:    andb $1, %al
 ; SSE2-NEXT:    movb %al, (%rdi)
 ; SSE2-NEXT:    movb -{{[0-9]+}}(%rsp), %al
 ; SSE2-NEXT:    andb $1, %al
-; SSE2-NEXT:    movb %al, 2(%rdi)
+; SSE2-NEXT:    movb %al, (%rdi)
 ; SSE2-NEXT:    movb -{{[0-9]+}}(%rsp), %al
 ; SSE2-NEXT:    andb $1, %al
 ; SSE2-NEXT:    movb %al, (%rdi)
 ; SSE2-NEXT:    movb -{{[0-9]+}}(%rsp), %al
 ; SSE2-NEXT:    andb $1, %al
-; SSE2-NEXT:    movb %al, 2(%rdi)
+; SSE2-NEXT:    movb %al, (%rdi)
 ; SSE2-NEXT:    movb -{{[0-9]+}}(%rsp), %al
+; SSE2-NEXT:    movb -{{[0-9]+}}(%rsp), %cl
+; SSE2-NEXT:    andb $1, %cl
+; SSE2-NEXT:    movb %cl, (%rdi)
 ; SSE2-NEXT:    andb $1, %al
 ; SSE2-NEXT:    movb %al, (%rdi)
 ; SSE2-NEXT:    movq %rdi, %rax
diff --git a/llvm/test/CodeGen/X86/vector-shuffle-variable-128.ll b/llvm/test/CodeGen/X86/vector-shuffle-variable-128.ll
index d130e7ff00b..2b79142ee57 100644
--- a/llvm/test/CodeGen/X86/vector-shuffle-variable-128.ll
+++ b/llvm/test/CodeGen/X86/vector-shuffle-variable-128.ll
@@ -36,8 +36,8 @@ define <2 x i64> @var_shuffle_v2i64_v2i64_xx_i64(<2 x i64> %x, i32 %i0, i32 %i1)
 ; SSE-NEXT:    movslq %edi, %rax
 ; SSE-NEXT:    movaps %xmm0, -{{[0-9]+}}(%rsp)
 ; SSE-NEXT:    movslq %esi, %rcx
-; SSE-NEXT:    movq {{.*#+}} xmm0 = mem[0],zero
 ; SSE-NEXT:    movq {{.*#+}} xmm1 = mem[0],zero
+; SSE-NEXT:    movq {{.*#+}} xmm0 = mem[0],zero
 ; SSE-NEXT:    punpcklqdq {{.*#+}} xmm0 = xmm0[0],xmm1[0]
 ; SSE-NEXT:    retq
 ;
@@ -48,7 +48,7 @@ define <2 x i64> @var_shuffle_v2i64_v2i64_xx_i64(<2 x i64> %x, i32 %i0, i32 %i1)
 ; AVX-NEXT:    movslq %esi, %rcx
 ; AVX-NEXT:    vmovq {{.*#+}} xmm0 = mem[0],zero
 ; AVX-NEXT:    vmovq {{.*#+}} xmm1 = mem[0],zero
-; AVX-NEXT:    vpunpcklqdq {{.*#+}} xmm0 = xmm0[0],xmm1[0]
+; AVX-NEXT:    vpunpcklqdq {{.*#+}} xmm0 = xmm1[0],xmm0[0]
 ; AVX-NEXT:    retq
   %x0 = extractelement <2 x i64> %x, i32 %i0
   %x1 = extractelement <2 x i64> %x, i32 %i1
@@ -67,10 +67,10 @@ define <4 x float> @var_shuffle_v4f32_v4f32_xxxx_i32(<4 x float> %x, i32 %i0, i3
 ; SSE2-NEXT:    movslq %ecx, %rcx
 ; SSE2-NEXT:    movss {{.*#+}} xmm0 = mem[0],zero,zero,zero
 ; SSE2-NEXT:    movss {{.*#+}} xmm1 = mem[0],zero,zero,zero
+; SSE2-NEXT:    unpcklps {{.*#+}} xmm1 = xmm1[0],xmm0[0],xmm1[1],xmm0[1]
 ; SSE2-NEXT:    movss {{.*#+}} xmm2 = mem[0],zero,zero,zero
+; SSE2-NEXT:    movss {{.*#+}} xmm0 = mem[0],zero,zero,zero
 ; SSE2-NEXT:    unpcklps {{.*#+}} xmm0 = xmm0[0],xmm2[0],xmm0[1],xmm2[1]
-; SSE2-NEXT:    movss {{.*#+}} xmm2 = mem[0],zero,zero,zero
-; SSE2-NEXT:    unpcklps {{.*#+}} xmm1 = xmm1[0],xmm2[0],xmm1[1],xmm2[1]
 ; SSE2-NEXT:    unpcklps {{.*#+}} xmm0 = xmm0[0],xmm1[0],xmm0[1],xmm1[1]
 ; SSE2-NEXT:    retq
 ;
@@ -83,10 +83,10 @@ define <4 x float> @var_shuffle_v4f32_v4f32_xxxx_i32(<4 x float> %x, i32 %i0, i3
 ; SSSE3-NEXT:    movslq %ecx, %rcx
 ; SSSE3-NEXT:    movss {{.*#+}} xmm0 = mem[0],zero,zero,zero
 ; SSSE3-NEXT:    movss {{.*#+}} xmm1 = mem[0],zero,zero,zero
+; SSSE3-NEXT:    unpcklps {{.*#+}} xmm1 = xmm1[0],xmm0[0],xmm1[1],xmm0[1]
 ; SSSE3-NEXT:    movss {{.*#+}} xmm2 = mem[0],zero,zero,zero
+; SSSE3-NEXT:    movss {{.*#+}} xmm0 = mem[0],zero,zero,zero
 ; SSSE3-NEXT:    unpcklps {{.*#+}} xmm0 = xmm0[0],xmm2[0],xmm0[1],xmm2[1]
-; SSSE3-NEXT:    movss {{.*#+}} xmm2 = mem[0],zero,zero,zero
-; SSSE3-NEXT:    unpcklps {{.*#+}} xmm1 = xmm1[0],xmm2[0],xmm1[1],xmm2[1]
 ; SSSE3-NEXT:    unpcklps {{.*#+}} xmm0 = xmm0[0],xmm1[0],xmm0[1],xmm1[1]
 ; SSSE3-NEXT:    retq
 ;
@@ -136,10 +136,10 @@ define <4 x i32> @var_shuffle_v4i32_v4i32_xxxx_i32(<4 x i32> %x, i32 %i0, i32 %i
 ; SSE2-NEXT:    movslq %ecx, %rcx
 ; SSE2-NEXT:    movd {{.*#+}} xmm0 = mem[0],zero,zero,zero
 ; SSE2-NEXT:    movd {{.*#+}} xmm1 = mem[0],zero,zero,zero
+; SSE2-NEXT:    punpckldq {{.*#+}} xmm1 = xmm1[0],xmm0[0],xmm1[1],xmm0[1]
 ; SSE2-NEXT:    movd {{.*#+}} xmm2 = mem[0],zero,zero,zero
+; SSE2-NEXT:    movd {{.*#+}} xmm0 = mem[0],zero,zero,zero
 ; SSE2-NEXT:    punpckldq {{.*#+}} xmm0 = xmm0[0],xmm2[0],xmm0[1],xmm2[1]
-; SSE2-NEXT:    movd {{.*#+}} xmm2 = mem[0],zero,zero,zero
-; SSE2-NEXT:    punpckldq {{.*#+}} xmm1 = xmm1[0],xmm2[0],xmm1[1],xmm2[1]
 ; SSE2-NEXT:    punpckldq {{.*#+}} xmm0 = xmm0[0],xmm1[0],xmm0[1],xmm1[1]
 ; SSE2-NEXT:    retq
 ;
@@ -152,10 +152,10 @@ define <4 x i32> @var_shuffle_v4i32_v4i32_xxxx_i32(<4 x i32> %x, i32 %i0, i32 %i
 ; SSSE3-NEXT:    movslq %ecx, %rcx
 ; SSSE3-NEXT:    movd {{.*#+}} xmm0 = mem[0],zero,zero,zero
 ; SSSE3-NEXT:    movd {{.*#+}} xmm1 = mem[0],zero,zero,zero
+; SSSE3-NEXT:    punpckldq {{.*#+}} xmm1 = xmm1[0],xmm0[0],xmm1[1],xmm0[1]
 ; SSSE3-NEXT:    movd {{.*#+}} xmm2 = mem[0],zero,zero,zero
+; SSSE3-NEXT:    movd {{.*#+}} xmm0 = mem[0],zero,zero,zero
 ; SSSE3-NEXT:    punpckldq {{.*#+}} xmm0 = xmm0[0],xmm2[0],xmm0[1],xmm2[1]
-; SSSE3-NEXT:    movd {{.*#+}} xmm2 = mem[0],zero,zero,zero
-; SSSE3-NEXT:    punpckldq {{.*#+}} xmm1 = xmm1[0],xmm2[0],xmm1[1],xmm2[1]
 ; SSSE3-NEXT:    punpckldq {{.*#+}} xmm0 = xmm0[0],xmm1[0],xmm0[1],xmm1[1]
 ; SSSE3-NEXT:    retq
 ;
@@ -204,38 +204,38 @@ define <8 x i16> @var_shuffle_v8i16_v8i16_xxxxxxxx_i16(<8 x i16> %x, i16 %i0, i1
 ; SSE2-NEXT:    # kill: %EDX<def> %EDX<kill> %RDX<def>
 ; SSE2-NEXT:    # kill: %ESI<def> %ESI<kill> %RSI<def>
 ; SSE2-NEXT:    # kill: %EDI<def> %EDI<kill> %RDI<def>
-; SSE2-NEXT:    movswq %di, %rax
+; SSE2-NEXT:    movswq %di, %r10
 ; SSE2-NEXT:    movswq %si, %rsi
-; SSE2-NEXT:    movswq %dx, %rdx
-; SSE2-NEXT:    movswq %cx, %r10
-; SSE2-NEXT:    movswq %r8w, %r11
+; SSE2-NEXT:    movswq %dx, %r11
+; SSE2-NEXT:    movswq %cx, %rcx
+; SSE2-NEXT:    movswq %r8w, %r8
 ; SSE2-NEXT:    movaps %xmm0, -{{[0-9]+}}(%rsp)
-; SSE2-NEXT:    movswq %r9w, %r8
-; SSE2-NEXT:    movswq {{[0-9]+}}(%rsp), %rcx
+; SSE2-NEXT:    movswq %r9w, %rax
+; SSE2-NEXT:    movswq {{[0-9]+}}(%rsp), %rdx
 ; SSE2-NEXT:    movswq {{[0-9]+}}(%rsp), %rdi
-; SSE2-NEXT:    movzwl -24(%rsp,%rcx,2), %ecx
 ; SSE2-NEXT:    movzwl -24(%rsp,%rdi,2), %edi
-; SSE2-NEXT:    movzwl -24(%rsp,%rax,2), %eax
-; SSE2-NEXT:    movzwl -24(%rsp,%rsi,2), %esi
-; SSE2-NEXT:    movd %ecx, %xmm0
-; SSE2-NEXT:    movzwl -24(%rsp,%rdx,2), %ecx
+; SSE2-NEXT:    movd %edi, %xmm0
+; SSE2-NEXT:    movzwl -24(%rsp,%rcx,2), %ecx
 ; SSE2-NEXT:    movd %ecx, %xmm1
 ; SSE2-NEXT:    punpcklwd {{.*#+}} xmm1 = xmm1[0],xmm0[0],xmm1[1],xmm0[1],xmm1[2],xmm0[2],xmm1[3],xmm0[3]
-; SSE2-NEXT:    movzwl -24(%rsp,%r10,2), %ecx
+; SSE2-NEXT:    movzwl -24(%rsp,%rax,2), %eax
 ; SSE2-NEXT:    movd %eax, %xmm0
-; SSE2-NEXT:    movzwl -24(%rsp,%r11,2), %eax
+; SSE2-NEXT:    movzwl -24(%rsp,%rsi,2), %eax
 ; SSE2-NEXT:    movd %eax, %xmm2
-; SSE2-NEXT:    punpcklwd {{.*#+}} xmm0 = xmm0[0],xmm2[0],xmm0[1],xmm2[1],xmm0[2],xmm2[2],xmm0[3],xmm2[3]
-; SSE2-NEXT:    punpcklwd {{.*#+}} xmm0 = xmm0[0],xmm1[0],xmm0[1],xmm1[1],xmm0[2],xmm1[2],xmm0[3],xmm1[3]
-; SSE2-NEXT:    movd %edi, %xmm1
-; SSE2-NEXT:    movd %ecx, %xmm2
+; SSE2-NEXT:    punpcklwd {{.*#+}} xmm2 = xmm2[0],xmm0[0],xmm2[1],xmm0[1],xmm2[2],xmm0[2],xmm2[3],xmm0[3]
 ; SSE2-NEXT:    punpcklwd {{.*#+}} xmm2 = xmm2[0],xmm1[0],xmm2[1],xmm1[1],xmm2[2],xmm1[2],xmm2[3],xmm1[3]
-; SSE2-NEXT:    movd %esi, %xmm1
+; SSE2-NEXT:    movzwl -24(%rsp,%rdx,2), %eax
+; SSE2-NEXT:    movd %eax, %xmm0
+; SSE2-NEXT:    movzwl -24(%rsp,%r11,2), %eax
+; SSE2-NEXT:    movd %eax, %xmm1
+; SSE2-NEXT:    punpcklwd {{.*#+}} xmm1 = xmm1[0],xmm0[0],xmm1[1],xmm0[1],xmm1[2],xmm0[2],xmm1[3],xmm0[3]
 ; SSE2-NEXT:    movzwl -24(%rsp,%r8,2), %eax
 ; SSE2-NEXT:    movd %eax, %xmm3
-; SSE2-NEXT:    punpcklwd {{.*#+}} xmm1 = xmm1[0],xmm3[0],xmm1[1],xmm3[1],xmm1[2],xmm3[2],xmm1[3],xmm3[3]
-; SSE2-NEXT:    punpcklwd {{.*#+}} xmm1 = xmm1[0],xmm2[0],xmm1[1],xmm2[1],xmm1[2],xmm2[2],xmm1[3],xmm2[3]
+; SSE2-NEXT:    movzwl -24(%rsp,%r10,2), %eax
+; SSE2-NEXT:    movd %eax, %xmm0
+; SSE2-NEXT:    punpcklwd {{.*#+}} xmm0 = xmm0[0],xmm3[0],xmm0[1],xmm3[1],xmm0[2],xmm3[2],xmm0[3],xmm3[3]
 ; SSE2-NEXT:    punpcklwd {{.*#+}} xmm0 = xmm0[0],xmm1[0],xmm0[1],xmm1[1],xmm0[2],xmm1[2],xmm0[3],xmm1[3]
+; SSE2-NEXT:    punpcklwd {{.*#+}} xmm0 = xmm0[0],xmm2[0],xmm0[1],xmm2[1],xmm0[2],xmm2[2],xmm0[3],xmm2[3]
 ; SSE2-NEXT:    retq
 ;
 ; SSSE3-LABEL: var_shuffle_v8i16_v8i16_xxxxxxxx_i16:
@@ -246,43 +246,42 @@ define <8 x i16> @var_shuffle_v8i16_v8i16_xxxxxxxx_i16(<8 x i16> %x, i16 %i0, i1
 ; SSSE3-NEXT:    # kill: %EDX<def> %EDX<kill> %RDX<def>
 ; SSSE3-NEXT:    # kill: %ESI<def> %ESI<kill> %RSI<def>
 ; SSSE3-NEXT:    # kill: %EDI<def> %EDI<kill> %RDI<def>
-; SSSE3-NEXT:    movswq %di, %rax
+; SSSE3-NEXT:    movswq %di, %r10
 ; SSSE3-NEXT:    movswq %si, %rsi
-; SSSE3-NEXT:    movswq %dx, %rdx
-; SSSE3-NEXT:    movswq %cx, %r10
-; SSSE3-NEXT:    movswq %r8w, %r11
+; SSSE3-NEXT:    movswq %dx, %r11
+; SSSE3-NEXT:    movswq %cx, %rcx
+; SSSE3-NEXT:    movswq %r8w, %r8
 ; SSSE3-NEXT:    movaps %xmm0, -{{[0-9]+}}(%rsp)
-; SSSE3-NEXT:    movswq %r9w, %r8
-; SSSE3-NEXT:    movswq {{[0-9]+}}(%rsp), %rcx
+; SSSE3-NEXT:    movswq %r9w, %rax
+; SSSE3-NEXT:    movswq {{[0-9]+}}(%rsp), %rdx
 ; SSSE3-NEXT:    movswq {{[0-9]+}}(%rsp), %rdi
-; SSSE3-NEXT:    movzwl -24(%rsp,%rcx,2), %ecx
 ; SSSE3-NEXT:    movzwl -24(%rsp,%rdi,2), %edi
-; SSSE3-NEXT:    movzwl -24(%rsp,%rax,2), %eax
-; SSSE3-NEXT:    movzwl -24(%rsp,%rsi,2), %esi
-; SSSE3-NEXT:    movd %ecx, %xmm0
-; SSSE3-NEXT:    movzwl -24(%rsp,%rdx,2), %ecx
+; SSSE3-NEXT:    movd %edi, %xmm0
+; SSSE3-NEXT:    movzwl -24(%rsp,%rcx,2), %ecx
 ; SSSE3-NEXT:    movd %ecx, %xmm1
 ; SSSE3-NEXT:    punpcklwd {{.*#+}} xmm1 = xmm1[0],xmm0[0],xmm1[1],xmm0[1],xmm1[2],xmm0[2],xmm1[3],xmm0[3]
-; SSSE3-NEXT:    movzwl -24(%rsp,%r10,2), %ecx
+; SSSE3-NEXT:    movzwl -24(%rsp,%rax,2), %eax
 ; SSSE3-NEXT:    movd %eax, %xmm0
-; SSSE3-NEXT:    movzwl -24(%rsp,%r11,2), %eax
+; SSSE3-NEXT:    movzwl -24(%rsp,%rsi,2), %eax
 ; SSSE3-NEXT:    movd %eax, %xmm2
-; SSSE3-NEXT:    punpcklwd {{.*#+}} xmm0 = xmm0[0],xmm2[0],xmm0[1],xmm2[1],xmm0[2],xmm2[2],xmm0[3],xmm2[3]
-; SSSE3-NEXT:    punpcklwd {{.*#+}} xmm0 = xmm0[0],xmm1[0],xmm0[1],xmm1[1],xmm0[2],xmm1[2],xmm0[3],xmm1[3]
-; SSSE3-NEXT:    movd %edi, %xmm1
-; SSSE3-NEXT:    movd %ecx, %xmm2
+; SSSE3-NEXT:    punpcklwd {{.*#+}} xmm2 = xmm2[0],xmm0[0],xmm2[1],xmm0[1],xmm2[2],xmm0[2],xmm2[3],xmm0[3]
 ; SSSE3-NEXT:    punpcklwd {{.*#+}} xmm2 = xmm2[0],xmm1[0],xmm2[1],xmm1[1],xmm2[2],xmm1[2],xmm2[3],xmm1[3]
-; SSSE3-NEXT:    movd %esi, %xmm1
+; SSSE3-NEXT:    movzwl -24(%rsp,%rdx,2), %eax
+; SSSE3-NEXT:    movd %eax, %xmm0
+; SSSE3-NEXT:    movzwl -24(%rsp,%r11,2), %eax
+; SSSE3-NEXT:    movd %eax, %xmm1
+; SSSE3-NEXT:    punpcklwd {{.*#+}} xmm1 = xmm1[0],xmm0[0],xmm1[1],xmm0[1],xmm1[2],xmm0[2],xmm1[3],xmm0[3]
 ; SSSE3-NEXT:    movzwl -24(%rsp,%r8,2), %eax
 ; SSSE3-NEXT:    movd %eax, %xmm3
-; SSSE3-NEXT:    punpcklwd {{.*#+}} xmm1 = xmm1[0],xmm3[0],xmm1[1],xmm3[1],xmm1[2],xmm3[2],xmm1[3],xmm3[3]
-; SSSE3-NEXT:    punpcklwd {{.*#+}} xmm1 = xmm1[0],xmm2[0],xmm1[1],xmm2[1],xmm1[2],xmm2[2],xmm1[3],xmm2[3]
+; SSSE3-NEXT:    movzwl -24(%rsp,%r10,2), %eax
+; SSSE3-NEXT:    movd %eax, %xmm0
+; SSSE3-NEXT:    punpcklwd {{.*#+}} xmm0 = xmm0[0],xmm3[0],xmm0[1],xmm3[1],xmm0[2],xmm3[2],xmm0[3],xmm3[3]
 ; SSSE3-NEXT:    punpcklwd {{.*#+}} xmm0 = xmm0[0],xmm1[0],xmm0[1],xmm1[1],xmm0[2],xmm1[2],xmm0[3],xmm1[3]
+; SSSE3-NEXT:    punpcklwd {{.*#+}} xmm0 = xmm0[0],xmm2[0],xmm0[1],xmm2[1],xmm0[2],xmm2[2],xmm0[3],xmm2[3]
 ; SSSE3-NEXT:    retq
 ;
 ; SSE41-LABEL: var_shuffle_v8i16_v8i16_xxxxxxxx_i16:
 ; SSE41:       # BB#0:
-; SSE41-NEXT:    pushq %rbx
 ; SSE41-NEXT:    # kill: %R9D<def> %R9D<kill> %R9<def>
 ; SSE41-NEXT:    # kill: %R8D<def> %R8D<kill> %R8<def>
 ; SSE41-NEXT:    # kill: %ECX<def> %ECX<kill> %RCX<def>
@@ -290,26 +289,23 @@ define <8 x i16> @var_shuffle_v8i16_v8i16_xxxxxxxx_i16(<8 x i16> %x, i16 %i0, i1
 ; SSE41-NEXT:    # kill: %ESI<def> %ESI<kill> %RSI<def>
 ; SSE41-NEXT:    # kill: %EDI<def> %EDI<kill> %RDI<def>
 ; SSE41-NEXT:    movswq %di, %rax
-; SSE41-NEXT:    movswq %si, %rbx
-; SSE41-NEXT:    movswq %dx, %r11
+; SSE41-NEXT:    movswq %si, %rsi
+; SSE41-NEXT:    movswq %dx, %rdx
 ; SSE41-NEXT:    movswq %cx, %r10
 ; SSE41-NEXT:    movswq %r8w, %rdi
 ; SSE41-NEXT:    movaps %xmm0, -{{[0-9]+}}(%rsp)
 ; SSE41-NEXT:    movswq %r9w, %rcx
-; SSE41-NEXT:    movswq {{[0-9]+}}(%rsp), %rdx
-; SSE41-NEXT:    movswq {{[0-9]+}}(%rsp), %rsi
-; SSE41-NEXT:    movzwl -16(%rsp,%rdx,2), %edx
-; SSE41-NEXT:    movzwl -16(%rsp,%rsi,2), %esi
-; SSE41-NEXT:    movzwl -16(%rsp,%rax,2), %eax
+; SSE41-NEXT:    movswq {{[0-9]+}}(%rsp), %r8
+; SSE41-NEXT:    movswq {{[0-9]+}}(%rsp), %r9
+; SSE41-NEXT:    movzwl -24(%rsp,%rax,2), %eax
 ; SSE41-NEXT:    movd %eax, %xmm0
-; SSE41-NEXT:    pinsrw $1, -16(%rsp,%rbx,2), %xmm0
-; SSE41-NEXT:    pinsrw $2, -16(%rsp,%r11,2), %xmm0
-; SSE41-NEXT:    pinsrw $3, -16(%rsp,%r10,2), %xmm0
-; SSE41-NEXT:    pinsrw $4, -16(%rsp,%rdi,2), %xmm0
-; SSE41-NEXT:    pinsrw $5, -16(%rsp,%rcx,2), %xmm0
-; SSE41-NEXT:    pinsrw $6, %edx, %xmm0
-; SSE41-NEXT:    pinsrw $7, %esi, %xmm0
-; SSE41-NEXT:    popq %rbx
+; SSE41-NEXT:    pinsrw $1, -24(%rsp,%rsi,2), %xmm0
+; SSE41-NEXT:    pinsrw $2, -24(%rsp,%rdx,2), %xmm0
+; SSE41-NEXT:    pinsrw $3, -24(%rsp,%r10,2), %xmm0
+; SSE41-NEXT:    pinsrw $4, -24(%rsp,%rdi,2), %xmm0
+; SSE41-NEXT:    pinsrw $5, -24(%rsp,%rcx,2), %xmm0
+; SSE41-NEXT:    pinsrw $6, -24(%rsp,%r8,2), %xmm0
+; SSE41-NEXT:    pinsrw $7, -24(%rsp,%r9,2), %xmm0
 ; SSE41-NEXT:    retq
 ;
 ; AVX-LABEL: var_shuffle_v8i16_v8i16_xxxxxxxx_i16:
@@ -331,8 +327,6 @@ define <8 x i16> @var_shuffle_v8i16_v8i16_xxxxxxxx_i16(<8 x i16> %x, i16 %i0, i1
 ; AVX-NEXT:    movswq %r9w, %rax
 ; AVX-NEXT:    movswq {{[0-9]+}}(%rsp), %rsi
 ; AVX-NEXT:    movswq {{[0-9]+}}(%rsp), %rdx
-; AVX-NEXT:    movzwl -24(%rsp,%rsi,2), %esi
-; AVX-NEXT:    movzwl -24(%rsp,%rdx,2), %edx
 ; AVX-NEXT:    movzwl -24(%rsp,%r10,2), %ebx
 ; AVX-NEXT:    vmovd %ebx, %xmm0
 ; AVX-NEXT:    vpinsrw $1, -24(%rsp,%r11,2), %xmm0, %xmm0
@@ -340,8 +334,8 @@ define <8 x i16> @var_shuffle_v8i16_v8i16_xxxxxxxx_i16(<8 x i16> %x, i16 %i0, i1
 ; AVX-NEXT:    vpinsrw $3, -24(%rsp,%rcx,2), %xmm0, %xmm0
 ; AVX-NEXT:    vpinsrw $4, -24(%rsp,%rdi,2), %xmm0, %xmm0
 ; AVX-NEXT:    vpinsrw $5, -24(%rsp,%rax,2), %xmm0, %xmm0
-; AVX-NEXT:    vpinsrw $6, %esi, %xmm0, %xmm0
-; AVX-NEXT:    vpinsrw $7, %edx, %xmm0, %xmm0
+; AVX-NEXT:    vpinsrw $6, -24(%rsp,%rsi,2), %xmm0, %xmm0
+; AVX-NEXT:    vpinsrw $7, -24(%rsp,%rdx,2), %xmm0, %xmm0
 ; AVX-NEXT:    popq %rbx
 ; AVX-NEXT:    popq %r14
 ; AVX-NEXT:    retq
@@ -377,67 +371,67 @@ define <16 x i8> @var_shuffle_v16i8_v16i8_xxxxxxxxxxxxxxxx_i8(<16 x i8> %x, i8 %
 ; SSE2-NEXT:    movsbq {{[0-9]+}}(%rsp), %r10
 ; SSE2-NEXT:    leaq -{{[0-9]+}}(%rsp), %r11
 ; SSE2-NEXT:    movzbl (%r10,%r11), %eax
-; SSE2-NEXT:    movd %eax, %xmm15
+; SSE2-NEXT:    movd %eax, %xmm8
 ; SSE2-NEXT:    movsbq {{[0-9]+}}(%rsp), %rax
 ; SSE2-NEXT:    movzbl (%rax,%r11), %eax
-; SSE2-NEXT:    movd %eax, %xmm8
+; SSE2-NEXT:    movd %eax, %xmm15
 ; SSE2-NEXT:    movsbq {{[0-9]+}}(%rsp), %rax
 ; SSE2-NEXT:    movzbl (%rax,%r11), %eax
 ; SSE2-NEXT:    movd %eax, %xmm9
-; SSE2-NEXT:    movsbq %dl, %rax
+; SSE2-NEXT:    movsbq %cl, %rax
 ; SSE2-NEXT:    movzbl (%rax,%r11), %eax
 ; SSE2-NEXT:    movd %eax, %xmm3
 ; SSE2-NEXT:    movsbq {{[0-9]+}}(%rsp), %rax
 ; SSE2-NEXT:    movzbl (%rax,%r11), %eax
 ; SSE2-NEXT:    movd %eax, %xmm10
-; SSE2-NEXT:    movsbq %dil, %rax
+; SSE2-NEXT:    movsbq %r9b, %rax
 ; SSE2-NEXT:    movzbl (%rax,%r11), %eax
-; SSE2-NEXT:    movd %eax, %xmm0
+; SSE2-NEXT:    movd %eax, %xmm7
 ; SSE2-NEXT:    movsbq {{[0-9]+}}(%rsp), %rax
 ; SSE2-NEXT:    movzbl (%rax,%r11), %eax
 ; SSE2-NEXT:    movd %eax, %xmm11
-; SSE2-NEXT:    movsbq %r8b, %rax
+; SSE2-NEXT:    movsbq %sil, %rax
 ; SSE2-NEXT:    movzbl (%rax,%r11), %eax
-; SSE2-NEXT:    movd %eax, %xmm7
+; SSE2-NEXT:    movd %eax, %xmm6
 ; SSE2-NEXT:    movsbq {{[0-9]+}}(%rsp), %rax
 ; SSE2-NEXT:    movzbl (%rax,%r11), %eax
-; SSE2-NEXT:    movd %eax, %xmm2
+; SSE2-NEXT:    movd %eax, %xmm12
 ; SSE2-NEXT:    movsbq {{[0-9]+}}(%rsp), %rax
 ; SSE2-NEXT:    movzbl (%rax,%r11), %eax
-; SSE2-NEXT:    movd %eax, %xmm12
+; SSE2-NEXT:    movd %eax, %xmm5
 ; SSE2-NEXT:    movsbq {{[0-9]+}}(%rsp), %rax
 ; SSE2-NEXT:    movzbl (%rax,%r11), %eax
 ; SSE2-NEXT:    movd %eax, %xmm13
-; SSE2-NEXT:    movsbq %cl, %rax
+; SSE2-NEXT:    movsbq %dl, %rax
 ; SSE2-NEXT:    movzbl (%rax,%r11), %eax
-; SSE2-NEXT:    movd %eax, %xmm6
+; SSE2-NEXT:    movd %eax, %xmm4
 ; SSE2-NEXT:    movsbq {{[0-9]+}}(%rsp), %rax
 ; SSE2-NEXT:    movzbl (%rax,%r11), %eax
 ; SSE2-NEXT:    movd %eax, %xmm14
-; SSE2-NEXT:    movsbq %sil, %rax
+; SSE2-NEXT:    movsbq %r8b, %rax
 ; SSE2-NEXT:    movzbl (%rax,%r11), %eax
-; SSE2-NEXT:    movd %eax, %xmm5
+; SSE2-NEXT:    movd %eax, %xmm1
 ; SSE2-NEXT:    movsbq {{[0-9]+}}(%rsp), %rax
 ; SSE2-NEXT:    movzbl (%rax,%r11), %eax
-; SSE2-NEXT:    movd %eax, %xmm4
-; SSE2-NEXT:    movsbq %r9b, %rax
+; SSE2-NEXT:    movd %eax, %xmm2
+; SSE2-NEXT:    movsbq %dil, %rax
 ; SSE2-NEXT:    movzbl (%rax,%r11), %eax
-; SSE2-NEXT:    movd %eax, %xmm1
+; SSE2-NEXT:    movd %eax, %xmm0
 ; SSE2-NEXT:    punpcklbw {{.*#+}} xmm15 = xmm15[0],xmm8[0],xmm15[1],xmm8[1],xmm15[2],xmm8[2],xmm15[3],xmm8[3],xmm15[4],xmm8[4],xmm15[5],xmm8[5],xmm15[6],xmm8[6],xmm15[7],xmm8[7]
 ; SSE2-NEXT:    punpcklbw {{.*#+}} xmm3 = xmm3[0],xmm9[0],xmm3[1],xmm9[1],xmm3[2],xmm9[2],xmm3[3],xmm9[3],xmm3[4],xmm9[4],xmm3[5],xmm9[5],xmm3[6],xmm9[6],xmm3[7],xmm9[7]
 ; SSE2-NEXT:    punpcklbw {{.*#+}} xmm3 = xmm3[0],xmm15[0],xmm3[1],xmm15[1],xmm3[2],xmm15[2],xmm3[3],xmm15[3],xmm3[4],xmm15[4],xmm3[5],xmm15[5],xmm3[6],xmm15[6],xmm3[7],xmm15[7]
-; SSE2-NEXT:    punpcklbw {{.*#+}} xmm0 = xmm0[0],xmm10[0],xmm0[1],xmm10[1],xmm0[2],xmm10[2],xmm0[3],xmm10[3],xmm0[4],xmm10[4],xmm0[5],xmm10[5],xmm0[6],xmm10[6],xmm0[7],xmm10[7]
-; SSE2-NEXT:    punpcklbw {{.*#+}} xmm7 = xmm7[0],xmm11[0],xmm7[1],xmm11[1],xmm7[2],xmm11[2],xmm7[3],xmm11[3],xmm7[4],xmm11[4],xmm7[5],xmm11[5],xmm7[6],xmm11[6],xmm7[7],xmm11[7]
-; SSE2-NEXT:    punpcklbw {{.*#+}} xmm0 = xmm0[0],xmm7[0],xmm0[1],xmm7[1],xmm0[2],xmm7[2],xmm0[3],xmm7[3],xmm0[4],xmm7[4],xmm0[5],xmm7[5],xmm0[6],xmm7[6],xmm0[7],xmm7[7]
-; SSE2-NEXT:    punpcklbw {{.*#+}} xmm0 = xmm0[0],xmm3[0],xmm0[1],xmm3[1],xmm0[2],xmm3[2],xmm0[3],xmm3[3],xmm0[4],xmm3[4],xmm0[5],xmm3[5],xmm0[6],xmm3[6],xmm0[7],xmm3[7]
-; SSE2-NEXT:    punpcklbw {{.*#+}} xmm2 = xmm2[0],xmm12[0],xmm2[1],xmm12[1],xmm2[2],xmm12[2],xmm2[3],xmm12[3],xmm2[4],xmm12[4],xmm2[5],xmm12[5],xmm2[6],xmm12[6],xmm2[7],xmm12[7]
-; SSE2-NEXT:    punpcklbw {{.*#+}} xmm6 = xmm6[0],xmm13[0],xmm6[1],xmm13[1],xmm6[2],xmm13[2],xmm6[3],xmm13[3],xmm6[4],xmm13[4],xmm6[5],xmm13[5],xmm6[6],xmm13[6],xmm6[7],xmm13[7]
-; SSE2-NEXT:    punpcklbw {{.*#+}} xmm6 = xmm6[0],xmm2[0],xmm6[1],xmm2[1],xmm6[2],xmm2[2],xmm6[3],xmm2[3],xmm6[4],xmm2[4],xmm6[5],xmm2[5],xmm6[6],xmm2[6],xmm6[7],xmm2[7]
-; SSE2-NEXT:    punpcklbw {{.*#+}} xmm5 = xmm5[0],xmm14[0],xmm5[1],xmm14[1],xmm5[2],xmm14[2],xmm5[3],xmm14[3],xmm5[4],xmm14[4],xmm5[5],xmm14[5],xmm5[6],xmm14[6],xmm5[7],xmm14[7]
-; SSE2-NEXT:    punpcklbw {{.*#+}} xmm1 = xmm1[0],xmm4[0],xmm1[1],xmm4[1],xmm1[2],xmm4[2],xmm1[3],xmm4[3],xmm1[4],xmm4[4],xmm1[5],xmm4[5],xmm1[6],xmm4[6],xmm1[7],xmm4[7]
-; SSE2-NEXT:    punpcklbw {{.*#+}} xmm5 = xmm5[0],xmm1[0],xmm5[1],xmm1[1],xmm5[2],xmm1[2],xmm5[3],xmm1[3],xmm5[4],xmm1[4],xmm5[5],xmm1[5],xmm5[6],xmm1[6],xmm5[7],xmm1[7]
-; SSE2-NEXT:    punpcklbw {{.*#+}} xmm5 = xmm5[0],xmm6[0],xmm5[1],xmm6[1],xmm5[2],xmm6[2],xmm5[3],xmm6[3],xmm5[4],xmm6[4],xmm5[5],xmm6[5],xmm5[6],xmm6[6],xmm5[7],xmm6[7]
-; SSE2-NEXT:    punpcklbw {{.*#+}} xmm0 = xmm0[0],xmm5[0],xmm0[1],xmm5[1],xmm0[2],xmm5[2],xmm0[3],xmm5[3],xmm0[4],xmm5[4],xmm0[5],xmm5[5],xmm0[6],xmm5[6],xmm0[7],xmm5[7]
+; SSE2-NEXT:    punpcklbw {{.*#+}} xmm7 = xmm7[0],xmm10[0],xmm7[1],xmm10[1],xmm7[2],xmm10[2],xmm7[3],xmm10[3],xmm7[4],xmm10[4],xmm7[5],xmm10[5],xmm7[6],xmm10[6],xmm7[7],xmm10[7]
+; SSE2-NEXT:    punpcklbw {{.*#+}} xmm6 = xmm6[0],xmm11[0],xmm6[1],xmm11[1],xmm6[2],xmm11[2],xmm6[3],xmm11[3],xmm6[4],xmm11[4],xmm6[5],xmm11[5],xmm6[6],xmm11[6],xmm6[7],xmm11[7]
+; SSE2-NEXT:    punpcklbw {{.*#+}} xmm6 = xmm6[0],xmm7[0],xmm6[1],xmm7[1],xmm6[2],xmm7[2],xmm6[3],xmm7[3],xmm6[4],xmm7[4],xmm6[5],xmm7[5],xmm6[6],xmm7[6],xmm6[7],xmm7[7]
+; SSE2-NEXT:    punpcklbw {{.*#+}} xmm6 = xmm6[0],xmm3[0],xmm6[1],xmm3[1],xmm6[2],xmm3[2],xmm6[3],xmm3[3],xmm6[4],xmm3[4],xmm6[5],xmm3[5],xmm6[6],xmm3[6],xmm6[7],xmm3[7]
+; SSE2-NEXT:    punpcklbw {{.*#+}} xmm5 = xmm5[0],xmm12[0],xmm5[1],xmm12[1],xmm5[2],xmm12[2],xmm5[3],xmm12[3],xmm5[4],xmm12[4],xmm5[5],xmm12[5],xmm5[6],xmm12[6],xmm5[7],xmm12[7]
+; SSE2-NEXT:    punpcklbw {{.*#+}} xmm4 = xmm4[0],xmm13[0],xmm4[1],xmm13[1],xmm4[2],xmm13[2],xmm4[3],xmm13[3],xmm4[4],xmm13[4],xmm4[5],xmm13[5],xmm4[6],xmm13[6],xmm4[7],xmm13[7]
+; SSE2-NEXT:    punpcklbw {{.*#+}} xmm4 = xmm4[0],xmm5[0],xmm4[1],xmm5[1],xmm4[2],xmm5[2],xmm4[3],xmm5[3],xmm4[4],xmm5[4],xmm4[5],xmm5[5],xmm4[6],xmm5[6],xmm4[7],xmm5[7]
+; SSE2-NEXT:    punpcklbw {{.*#+}} xmm1 = xmm1[0],xmm14[0],xmm1[1],xmm14[1],xmm1[2],xmm14[2],xmm1[3],xmm14[3],xmm1[4],xmm14[4],xmm1[5],xmm14[5],xmm1[6],xmm14[6],xmm1[7],xmm14[7]
+; SSE2-NEXT:    punpcklbw {{.*#+}} xmm0 = xmm0[0],xmm2[0],xmm0[1],xmm2[1],xmm0[2],xmm2[2],xmm0[3],xmm2[3],xmm0[4],xmm2[4],xmm0[5],xmm2[5],xmm0[6],xmm2[6],xmm0[7],xmm2[7]
+; SSE2-NEXT:    punpcklbw {{.*#+}} xmm0 = xmm0[0],xmm1[0],xmm0[1],xmm1[1],xmm0[2],xmm1[2],xmm0[3],xmm1[3],xmm0[4],xmm1[4],xmm0[5],xmm1[5],xmm0[6],xmm1[6],xmm0[7],xmm1[7]
+; SSE2-NEXT:    punpcklbw {{.*#+}} xmm0 = xmm0[0],xmm4[0],xmm0[1],xmm4[1],xmm0[2],xmm4[2],xmm0[3],xmm4[3],xmm0[4],xmm4[4],xmm0[5],xmm4[5],xmm0[6],xmm4[6],xmm0[7],xmm4[7]
+; SSE2-NEXT:    punpcklbw {{.*#+}} xmm0 = xmm0[0],xmm6[0],xmm0[1],xmm6[1],xmm0[2],xmm6[2],xmm0[3],xmm6[3],xmm0[4],xmm6[4],xmm0[5],xmm6[5],xmm0[6],xmm6[6],xmm0[7],xmm6[7]
 ; SSE2-NEXT:    retq
 ;
 ; SSSE3-LABEL: var_shuffle_v16i8_v16i8_xxxxxxxxxxxxxxxx_i8:
@@ -452,201 +446,157 @@ define <16 x i8> @var_shuffle_v16i8_v16i8_xxxxxxxxxxxxxxxx_i8(<16 x i8> %x, i8 %
 ; SSSE3-NEXT:    movsbq {{[0-9]+}}(%rsp), %r10
 ; SSSE3-NEXT:    leaq -{{[0-9]+}}(%rsp), %r11
 ; SSSE3-NEXT:    movzbl (%r10,%r11), %eax
-; SSSE3-NEXT:    movd %eax, %xmm15
+; SSSE3-NEXT:    movd %eax, %xmm8
 ; SSSE3-NEXT:    movsbq {{[0-9]+}}(%rsp), %rax
 ; SSSE3-NEXT:    movzbl (%rax,%r11), %eax
-; SSSE3-NEXT:    movd %eax, %xmm8
+; SSSE3-NEXT:    movd %eax, %xmm15
 ; SSSE3-NEXT:    movsbq {{[0-9]+}}(%rsp), %rax
 ; SSSE3-NEXT:    movzbl (%rax,%r11), %eax
 ; SSSE3-NEXT:    movd %eax, %xmm9
-; SSSE3-NEXT:    movsbq %dl, %rax
+; SSSE3-NEXT:    movsbq %cl, %rax
 ; SSSE3-NEXT:    movzbl (%rax,%r11), %eax
 ; SSSE3-NEXT:    movd %eax, %xmm3
 ; SSSE3-NEXT:    movsbq {{[0-9]+}}(%rsp), %rax
 ; SSSE3-NEXT:    movzbl (%rax,%r11), %eax
 ; SSSE3-NEXT:    movd %eax, %xmm10
-; SSSE3-NEXT:    movsbq %dil, %rax
+; SSSE3-NEXT:    movsbq %r9b, %rax
 ; SSSE3-NEXT:    movzbl (%rax,%r11), %eax
-; SSSE3-NEXT:    movd %eax, %xmm0
+; SSSE3-NEXT:    movd %eax, %xmm7
 ; SSSE3-NEXT:    movsbq {{[0-9]+}}(%rsp), %rax
 ; SSSE3-NEXT:    movzbl (%rax,%r11), %eax
 ; SSSE3-NEXT:    movd %eax, %xmm11
-; SSSE3-NEXT:    movsbq %r8b, %rax
+; SSSE3-NEXT:    movsbq %sil, %rax
 ; SSSE3-NEXT:    movzbl (%rax,%r11), %eax
-; SSSE3-NEXT:    movd %eax, %xmm7
+; SSSE3-NEXT:    movd %eax, %xmm6
 ; SSSE3-NEXT:    movsbq {{[0-9]+}}(%rsp), %rax
 ; SSSE3-NEXT:    movzbl (%rax,%r11), %eax
-; SSSE3-NEXT:    movd %eax, %xmm2
+; SSSE3-NEXT:    movd %eax, %xmm12
 ; SSSE3-NEXT:    movsbq {{[0-9]+}}(%rsp), %rax
 ; SSSE3-NEXT:    movzbl (%rax,%r11), %eax
-; SSSE3-NEXT:    movd %eax, %xmm12
+; SSSE3-NEXT:    movd %eax, %xmm5
 ; SSSE3-NEXT:    movsbq {{[0-9]+}}(%rsp), %rax
 ; SSSE3-NEXT:    movzbl (%rax,%r11), %eax
 ; SSSE3-NEXT:    movd %eax, %xmm13
-; SSSE3-NEXT:    movsbq %cl, %rax
+; SSSE3-NEXT:    movsbq %dl, %rax
 ; SSSE3-NEXT:    movzbl (%rax,%r11), %eax
-; SSSE3-NEXT:    movd %eax, %xmm6
+; SSSE3-NEXT:    movd %eax, %xmm4
 ; SSSE3-NEXT:    movsbq {{[0-9]+}}(%rsp), %rax
 ; SSSE3-NEXT:    movzbl (%rax,%r11), %eax
 ; SSSE3-NEXT:    movd %eax, %xmm14
-; SSSE3-NEXT:    movsbq %sil, %rax
+; SSSE3-NEXT:    movsbq %r8b, %rax
 ; SSSE3-NEXT:    movzbl (%rax,%r11), %eax
-; SSSE3-NEXT:    movd %eax, %xmm5
+; SSSE3-NEXT:    movd %eax, %xmm1
 ; SSSE3-NEXT:    movsbq {{[0-9]+}}(%rsp), %rax
 ; SSSE3-NEXT:    movzbl (%rax,%r11), %eax
-; SSSE3-NEXT:    movd %eax, %xmm4
-; SSSE3-NEXT:    movsbq %r9b, %rax
+; SSSE3-NEXT:    movd %eax, %xmm2
+; SSSE3-NEXT:    movsbq %dil, %rax
 ; SSSE3-NEXT:    movzbl (%rax,%r11), %eax
-; SSSE3-NEXT:    movd %eax, %xmm1
+; SSSE3-NEXT:    movd %eax, %xmm0
 ; SSSE3-NEXT:    punpcklbw {{.*#+}} xmm15 = xmm15[0],xmm8[0],xmm15[1],xmm8[1],xmm15[2],xmm8[2],xmm15[3],xmm8[3],xmm15[4],xmm8[4],xmm15[5],xmm8[5],xmm15[6],xmm8[6],xmm15[7],xmm8[7]
 ; SSSE3-NEXT:    punpcklbw {{.*#+}} xmm3 = xmm3[0],xmm9[0],xmm3[1],xmm9[1],xmm3[2],xmm9[2],xmm3[3],xmm9[3],xmm3[4],xmm9[4],xmm3[5],xmm9[5],xmm3[6],xmm9[6],xmm3[7],xmm9[7]
 ; SSSE3-NEXT:    punpcklbw {{.*#+}} xmm3 = xmm3[0],xmm15[0],xmm3[1],xmm15[1],xmm3[2],xmm15[2],xmm3[3],xmm15[3],xmm3[4],xmm15[4],xmm3[5],xmm15[5],xmm3[6],xmm15[6],xmm3[7],xmm15[7]
-; SSSE3-NEXT:    punpcklbw {{.*#+}} xmm0 = xmm0[0],xmm10[0],xmm0[1],xmm10[1],xmm0[2],xmm10[2],xmm0[3],xmm10[3],xmm0[4],xmm10[4],xmm0[5],xmm10[5],xmm0[6],xmm10[6],xmm0[7],xmm10[7]
-; SSSE3-NEXT:    punpcklbw {{.*#+}} xmm7 = xmm7[0],xmm11[0],xmm7[1],xmm11[1],xmm7[2],xmm11[2],xmm7[3],xmm11[3],xmm7[4],xmm11[4],xmm7[5],xmm11[5],xmm7[6],xmm11[6],xmm7[7],xmm11[7]
-; SSSE3-NEXT:    punpcklbw {{.*#+}} xmm0 = xmm0[0],xmm7[0],xmm0[1],xmm7[1],xmm0[2],xmm7[2],xmm0[3],xmm7[3],xmm0[4],xmm7[4],xmm0[5],xmm7[5],xmm0[6],xmm7[6],xmm0[7],xmm7[7]
-; SSSE3-NEXT:    punpcklbw {{.*#+}} xmm0 = xmm0[0],xmm3[0],xmm0[1],xmm3[1],xmm0[2],xmm3[2],xmm0[3],xmm3[3],xmm0[4],xmm3[4],xmm0[5],xmm3[5],xmm0[6],xmm3[6],xmm0[7],xmm3[7]
-; SSSE3-NEXT:    punpcklbw {{.*#+}} xmm2 = xmm2[0],xmm12[0],xmm2[1],xmm12[1],xmm2[2],xmm12[2],xmm2[3],xmm12[3],xmm2[4],xmm12[4],xmm2[5],xmm12[5],xmm2[6],xmm12[6],xmm2[7],xmm12[7]
-; SSSE3-NEXT:    punpcklbw {{.*#+}} xmm6 = xmm6[0],xmm13[0],xmm6[1],xmm13[1],xmm6[2],xmm13[2],xmm6[3],xmm13[3],xmm6[4],xmm13[4],xmm6[5],xmm13[5],xmm6[6],xmm13[6],xmm6[7],xmm13[7]
-; SSSE3-NEXT:    punpcklbw {{.*#+}} xmm6 = xmm6[0],xmm2[0],xmm6[1],xmm2[1],xmm6[2],xmm2[2],xmm6[3],xmm2[3],xmm6[4],xmm2[4],xmm6[5],xmm2[5],xmm6[6],xmm2[6],xmm6[7],xmm2[7]
-; SSSE3-NEXT:    punpcklbw {{.*#+}} xmm5 = xmm5[0],xmm14[0],xmm5[1],xmm14[1],xmm5[2],xmm14[2],xmm5[3],xmm14[3],xmm5[4],xmm14[4],xmm5[5],xmm14[5],xmm5[6],xmm14[6],xmm5[7],xmm14[7]
-; SSSE3-NEXT:    punpcklbw {{.*#+}} xmm1 = xmm1[0],xmm4[0],xmm1[1],xmm4[1],xmm1[2],xmm4[2],xmm1[3],xmm4[3],xmm1[4],xmm4[4],xmm1[5],xmm4[5],xmm1[6],xmm4[6],xmm1[7],xmm4[7]
-; SSSE3-NEXT:    punpcklbw {{.*#+}} xmm5 = xmm5[0],xmm1[0],xmm5[1],xmm1[1],xmm5[2],xmm1[2],xmm5[3],xmm1[3],xmm5[4],xmm1[4],xmm5[5],xmm1[5],xmm5[6],xmm1[6],xmm5[7],xmm1[7]
-; SSSE3-NEXT:    punpcklbw {{.*#+}} xmm5 = xmm5[0],xmm6[0],xmm5[1],xmm6[1],xmm5[2],xmm6[2],xmm5[3],xmm6[3],xmm5[4],xmm6[4],xmm5[5],xmm6[5],xmm5[6],xmm6[6],xmm5[7],xmm6[7]
-; SSSE3-NEXT:    punpcklbw {{.*#+}} xmm0 = xmm0[0],xmm5[0],xmm0[1],xmm5[1],xmm0[2],xmm5[2],xmm0[3],xmm5[3],xmm0[4],xmm5[4],xmm0[5],xmm5[5],xmm0[6],xmm5[6],xmm0[7],xmm5[7]
+; SSSE3-NEXT:    punpcklbw {{.*#+}} xmm7 = xmm7[0],xmm10[0],xmm7[1],xmm10[1],xmm7[2],xmm10[2],xmm7[3],xmm10[3],xmm7[4],xmm10[4],xmm7[5],xmm10[5],xmm7[6],xmm10[6],xmm7[7],xmm10[7]
+; SSSE3-NEXT:    punpcklbw {{.*#+}} xmm6 = xmm6[0],xmm11[0],xmm6[1],xmm11[1],xmm6[2],xmm11[2],xmm6[3],xmm11[3],xmm6[4],xmm11[4],xmm6[5],xmm11[5],xmm6[6],xmm11[6],xmm6[7],xmm11[7]
+; SSSE3-NEXT:    punpcklbw {{.*#+}} xmm6 = xmm6[0],xmm7[0],xmm6[1],xmm7[1],xmm6[2],xmm7[2],xmm6[3],xmm7[3],xmm6[4],xmm7[4],xmm6[5],xmm7[5],xmm6[6],xmm7[6],xmm6[7],xmm7[7]
+; SSSE3-NEXT:    punpcklbw {{.*#+}} xmm6 = xmm6[0],xmm3[0],xmm6[1],xmm3[1],xmm6[2],xmm3[2],xmm6[3],xmm3[3],xmm6[4],xmm3[4],xmm6[5],xmm3[5],xmm6[6],xmm3[6],xmm6[7],xmm3[7]
+; SSSE3-NEXT:    punpcklbw {{.*#+}} xmm5 = xmm5[0],xmm12[0],xmm5[1],xmm12[1],xmm5[2],xmm12[2],xmm5[3],xmm12[3],xmm5[4],xmm12[4],xmm5[5],xmm12[5],xmm5[6],xmm12[6],xmm5[7],xmm12[7]
+; SSSE3-NEXT:    punpcklbw {{.*#+}} xmm4 = xmm4[0],xmm13[0],xmm4[1],xmm13[1],xmm4[2],xmm13[2],xmm4[3],xmm13[3],xmm4[4],xmm13[4],xmm4[5],xmm13[5],xmm4[6],xmm13[6],xmm4[7],xmm13[7]
+; SSSE3-NEXT:    punpcklbw {{.*#+}} xmm4 = xmm4[0],xmm5[0],xmm4[1],xmm5[1],xmm4[2],xmm5[2],xmm4[3],xmm5[3],xmm4[4],xmm5[4],xmm4[5],xmm5[5],xmm4[6],xmm5[6],xmm4[7],xmm5[7]
+; SSSE3-NEXT:    punpcklbw {{.*#+}} xmm1 = xmm1[0],xmm14[0],xmm1[1],xmm14[1],xmm1[2],xmm14[2],xmm1[3],xmm14[3],xmm1[4],xmm14[4],xmm1[5],xmm14[5],xmm1[6],xmm14[6],xmm1[7],xmm14[7]
+; SSSE3-NEXT:    punpcklbw {{.*#+}} xmm0 = xmm0[0],xmm2[0],xmm0[1],xmm2[1],xmm0[2],xmm2[2],xmm0[3],xmm2[3],xmm0[4],xmm2[4],xmm0[5],xmm2[5],xmm0[6],xmm2[6],xmm0[7],xmm2[7]
+; SSSE3-NEXT:    punpcklbw {{.*#+}} xmm0 = xmm0[0],xmm1[0],xmm0[1],xmm1[1],xmm0[2],xmm1[2],xmm0[3],xmm1[3],xmm0[4],xmm1[4],xmm0[5],xmm1[5],xmm0[6],xmm1[6],xmm0[7],xmm1[7]
+; SSSE3-NEXT:    punpcklbw {{.*#+}} xmm0 = xmm0[0],xmm4[0],xmm0[1],xmm4[1],xmm0[2],xmm4[2],xmm0[3],xmm4[3],xmm0[4],xmm4[4],xmm0[5],xmm4[5],xmm0[6],xmm4[6],xmm0[7],xmm4[7]
+; SSSE3-NEXT:    punpcklbw {{.*#+}} xmm0 = xmm0[0],xmm6[0],xmm0[1],xmm6[1],xmm0[2],xmm6[2],xmm0[3],xmm6[3],xmm0[4],xmm6[4],xmm0[5],xmm6[5],xmm0[6],xmm6[6],xmm0[7],xmm6[7]
 ; SSSE3-NEXT:    retq
 ;
 ; SSE41-LABEL: var_shuffle_v16i8_v16i8_xxxxxxxxxxxxxxxx_i8:
 ; SSE41:       # BB#0:
-; SSE41-NEXT:    pushq %rbp
-; SSE41-NEXT:    pushq %r15
-; SSE41-NEXT:    pushq %r14
-; SSE41-NEXT:    pushq %r13
-; SSE41-NEXT:    pushq %r12
-; SSE41-NEXT:    pushq %rbx
 ; SSE41-NEXT:    # kill: %R9D<def> %R9D<kill> %R9<def>
 ; SSE41-NEXT:    # kill: %R8D<def> %R8D<kill> %R8<def>
 ; SSE41-NEXT:    # kill: %ECX<def> %ECX<kill> %RCX<def>
 ; SSE41-NEXT:    # kill: %EDX<def> %EDX<kill> %RDX<def>
 ; SSE41-NEXT:    # kill: %ESI<def> %ESI<kill> %RSI<def>
 ; SSE41-NEXT:    # kill: %EDI<def> %EDI<kill> %RDI<def>
-; SSE41-NEXT:    movsbq %dil, %r15
-; SSE41-NEXT:    movsbq %sil, %r14
-; SSE41-NEXT:    movsbq %dl, %r11
-; SSE41-NEXT:    movsbq %cl, %r10
-; SSE41-NEXT:    movsbq %r8b, %r8
+; SSE41-NEXT:    movsbq %dil, %rdi
 ; SSE41-NEXT:    movaps %xmm0, -{{[0-9]+}}(%rsp)
-; SSE41-NEXT:    movsbq %r9b, %r9
-; SSE41-NEXT:    movsbq {{[0-9]+}}(%rsp), %r12
-; SSE41-NEXT:    movsbq {{[0-9]+}}(%rsp), %r13
-; SSE41-NEXT:    movsbq {{[0-9]+}}(%rsp), %rbp
-; SSE41-NEXT:    movsbq {{[0-9]+}}(%rsp), %rbx
 ; SSE41-NEXT:    leaq -{{[0-9]+}}(%rsp), %rax
-; SSE41-NEXT:    movzbl (%r15,%rax), %ecx
-; SSE41-NEXT:    movd %ecx, %xmm0
-; SSE41-NEXT:    movsbq {{[0-9]+}}(%rsp), %r15
-; SSE41-NEXT:    pinsrb $1, (%r14,%rax), %xmm0
-; SSE41-NEXT:    movsbq {{[0-9]+}}(%rsp), %r14
-; SSE41-NEXT:    pinsrb $2, (%r11,%rax), %xmm0
-; SSE41-NEXT:    movsbq {{[0-9]+}}(%rsp), %r11
-; SSE41-NEXT:    pinsrb $3, (%r10,%rax), %xmm0
-; SSE41-NEXT:    movsbq {{[0-9]+}}(%rsp), %r10
-; SSE41-NEXT:    pinsrb $4, (%r8,%rax), %xmm0
+; SSE41-NEXT:    movzbl (%rdi,%rax), %edi
+; SSE41-NEXT:    movd %edi, %xmm0
+; SSE41-NEXT:    movsbq %sil, %rsi
+; SSE41-NEXT:    pinsrb $1, (%rsi,%rax), %xmm0
+; SSE41-NEXT:    movsbq %dl, %rdx
+; SSE41-NEXT:    pinsrb $2, (%rdx,%rax), %xmm0
+; SSE41-NEXT:    movsbq %cl, %rcx
+; SSE41-NEXT:    pinsrb $3, (%rcx,%rax), %xmm0
+; SSE41-NEXT:    movsbq %r8b, %rcx
+; SSE41-NEXT:    pinsrb $4, (%rcx,%rax), %xmm0
+; SSE41-NEXT:    movsbq %r9b, %rcx
+; SSE41-NEXT:    pinsrb $5, (%rcx,%rax), %xmm0
 ; SSE41-NEXT:    movsbq {{[0-9]+}}(%rsp), %rcx
-; SSE41-NEXT:    pinsrb $5, (%r9,%rax), %xmm0
-; SSE41-NEXT:    movsbq {{[0-9]+}}(%rsp), %rdx
-; SSE41-NEXT:    movzbl (%r12,%rax), %esi
-; SSE41-NEXT:    movzbl (%r13,%rax), %edi
-; SSE41-NEXT:    movzbl (%rbp,%rax), %ebp
-; SSE41-NEXT:    movzbl (%rbx,%rax), %ebx
-; SSE41-NEXT:    movzbl (%r15,%rax), %r8d
-; SSE41-NEXT:    movzbl (%r14,%rax), %r9d
-; SSE41-NEXT:    movzbl (%r11,%rax), %r11d
-; SSE41-NEXT:    movzbl (%r10,%rax), %r10d
-; SSE41-NEXT:    movzbl (%rcx,%rax), %ecx
-; SSE41-NEXT:    movzbl (%rdx,%rax), %eax
-; SSE41-NEXT:    pinsrb $6, %esi, %xmm0
-; SSE41-NEXT:    pinsrb $7, %edi, %xmm0
-; SSE41-NEXT:    pinsrb $8, %ebp, %xmm0
-; SSE41-NEXT:    pinsrb $9, %ebx, %xmm0
-; SSE41-NEXT:    pinsrb $10, %r8d, %xmm0
-; SSE41-NEXT:    pinsrb $11, %r9d, %xmm0
-; SSE41-NEXT:    pinsrb $12, %r11d, %xmm0
-; SSE41-NEXT:    pinsrb $13, %r10d, %xmm0
-; SSE41-NEXT:    pinsrb $14, %ecx, %xmm0
-; SSE41-NEXT:    pinsrb $15, %eax, %xmm0
-; SSE41-NEXT:    popq %rbx
-; SSE41-NEXT:    popq %r12
-; SSE41-NEXT:    popq %r13
-; SSE41-NEXT:    popq %r14
-; SSE41-NEXT:    popq %r15
-; SSE41-NEXT:    popq %rbp
+; SSE41-NEXT:    pinsrb $6, (%rcx,%rax), %xmm0
+; SSE41-NEXT:    movsbq {{[0-9]+}}(%rsp), %rcx
+; SSE41-NEXT:    pinsrb $7, (%rcx,%rax), %xmm0
+; SSE41-NEXT:    movsbq {{[0-9]+}}(%rsp), %rcx
+; SSE41-NEXT:    pinsrb $8, (%rcx,%rax), %xmm0
+; SSE41-NEXT:    movsbq {{[0-9]+}}(%rsp), %rcx
+; SSE41-NEXT:    pinsrb $9, (%rcx,%rax), %xmm0
+; SSE41-NEXT:    movsbq {{[0-9]+}}(%rsp), %rcx
+; SSE41-NEXT:    pinsrb $10, (%rcx,%rax), %xmm0
+; SSE41-NEXT:    movsbq {{[0-9]+}}(%rsp), %rcx
+; SSE41-NEXT:    pinsrb $11, (%rcx,%rax), %xmm0
+; SSE41-NEXT:    movsbq {{[0-9]+}}(%rsp), %rcx
+; SSE41-NEXT:    pinsrb $12, (%rcx,%rax), %xmm0
+; SSE41-NEXT:    movsbq {{[0-9]+}}(%rsp), %rcx
+; SSE41-NEXT:    pinsrb $13, (%rcx,%rax), %xmm0
+; SSE41-NEXT:    movsbq {{[0-9]+}}(%rsp), %rcx
+; SSE41-NEXT:    pinsrb $14, (%rcx,%rax), %xmm0
+; SSE41-NEXT:    movsbq {{[0-9]+}}(%rsp), %rcx
+; SSE41-NEXT:    pinsrb $15, (%rcx,%rax), %xmm0
 ; SSE41-NEXT:    retq
 ;
 ; AVX-LABEL: var_shuffle_v16i8_v16i8_xxxxxxxxxxxxxxxx_i8:
 ; AVX:       # BB#0:
-; AVX-NEXT:    pushq %rbp
-; AVX-NEXT:    pushq %r15
-; AVX-NEXT:    pushq %r14
-; AVX-NEXT:    pushq %r13
-; AVX-NEXT:    pushq %r12
-; AVX-NEXT:    pushq %rbx
 ; AVX-NEXT:    # kill: %R9D<def> %R9D<kill> %R9<def>
 ; AVX-NEXT:    # kill: %R8D<def> %R8D<kill> %R8<def>
 ; AVX-NEXT:    # kill: %ECX<def> %ECX<kill> %RCX<def>
 ; AVX-NEXT:    # kill: %EDX<def> %EDX<kill> %RDX<def>
 ; AVX-NEXT:    # kill: %ESI<def> %ESI<kill> %RSI<def>
 ; AVX-NEXT:    # kill: %EDI<def> %EDI<kill> %RDI<def>
-; AVX-NEXT:    movsbq %dil, %r10
-; AVX-NEXT:    movsbq %sil, %r11
-; AVX-NEXT:    movsbq %dl, %r14
-; AVX-NEXT:    movsbq %cl, %r15
-; AVX-NEXT:    movsbq %r8b, %r8
+; AVX-NEXT:    movsbq %dil, %rax
 ; AVX-NEXT:    vmovaps %xmm0, -{{[0-9]+}}(%rsp)
-; AVX-NEXT:    movsbq %r9b, %r9
-; AVX-NEXT:    movsbq {{[0-9]+}}(%rsp), %r12
-; AVX-NEXT:    movsbq {{[0-9]+}}(%rsp), %r13
-; AVX-NEXT:    movsbq {{[0-9]+}}(%rsp), %rbp
-; AVX-NEXT:    movsbq {{[0-9]+}}(%rsp), %rcx
 ; AVX-NEXT:    leaq -{{[0-9]+}}(%rsp), %rdi
-; AVX-NEXT:    movzbl (%r10,%rdi), %eax
+; AVX-NEXT:    movzbl (%rax,%rdi), %eax
 ; AVX-NEXT:    vmovd %eax, %xmm0
-; AVX-NEXT:    movsbq {{[0-9]+}}(%rsp), %r10
-; AVX-NEXT:    vpinsrb $1, (%r11,%rdi), %xmm0, %xmm0
-; AVX-NEXT:    movsbq {{[0-9]+}}(%rsp), %r11
-; AVX-NEXT:    vpinsrb $2, (%r14,%rdi), %xmm0, %xmm0
-; AVX-NEXT:    movsbq {{[0-9]+}}(%rsp), %r14
-; AVX-NEXT:    vpinsrb $3, (%r15,%rdi), %xmm0, %xmm0
-; AVX-NEXT:    movsbq {{[0-9]+}}(%rsp), %r15
-; AVX-NEXT:    vpinsrb $4, (%r8,%rdi), %xmm0, %xmm0
-; AVX-NEXT:    movsbq {{[0-9]+}}(%rsp), %r8
-; AVX-NEXT:    vpinsrb $5, (%r9,%rdi), %xmm0, %xmm0
-; AVX-NEXT:    movsbq {{[0-9]+}}(%rsp), %rsi
-; AVX-NEXT:    movzbl (%r12,%rdi), %edx
-; AVX-NEXT:    movzbl (%r13,%rdi), %ebx
-; AVX-NEXT:    movzbl (%rbp,%rdi), %ebp
-; AVX-NEXT:    movzbl (%rcx,%rdi), %ecx
-; AVX-NEXT:    movzbl (%r10,%rdi), %eax
-; AVX-NEXT:    movzbl (%r11,%rdi), %r9d
-; AVX-NEXT:    movzbl (%r14,%rdi), %r10d
-; AVX-NEXT:    movzbl (%r15,%rdi), %r11d
-; AVX-NEXT:    movzbl (%r8,%rdi), %r8d
-; AVX-NEXT:    movzbl (%rsi,%rdi), %esi
-; AVX-NEXT:    vpinsrb $6, %edx, %xmm0, %xmm0
-; AVX-NEXT:    vpinsrb $7, %ebx, %xmm0, %xmm0
-; AVX-NEXT:    vpinsrb $8, %ebp, %xmm0, %xmm0
-; AVX-NEXT:    vpinsrb $9, %ecx, %xmm0, %xmm0
-; AVX-NEXT:    vpinsrb $10, %eax, %xmm0, %xmm0
-; AVX-NEXT:    vpinsrb $11, %r9d, %xmm0, %xmm0
-; AVX-NEXT:    vpinsrb $12, %r10d, %xmm0, %xmm0
-; AVX-NEXT:    vpinsrb $13, %r11d, %xmm0, %xmm0
-; AVX-NEXT:    vpinsrb $14, %r8d, %xmm0, %xmm0
-; AVX-NEXT:    vpinsrb $15, %esi, %xmm0, %xmm0
-; AVX-NEXT:    popq %rbx
-; AVX-NEXT:    popq %r12
-; AVX-NEXT:    popq %r13
-; AVX-NEXT:    popq %r14
-; AVX-NEXT:    popq %r15
-; AVX-NEXT:    popq %rbp
+; AVX-NEXT:    movsbq %sil, %rax
+; AVX-NEXT:    vpinsrb $1, (%rax,%rdi), %xmm0, %xmm0
+; AVX-NEXT:    movsbq %dl, %rax
+; AVX-NEXT:    vpinsrb $2, (%rax,%rdi), %xmm0, %xmm0
+; AVX-NEXT:    movsbq %cl, %rax
+; AVX-NEXT:    vpinsrb $3, (%rax,%rdi), %xmm0, %xmm0
+; AVX-NEXT:    movsbq %r8b, %rax
+; AVX-NEXT:    vpinsrb $4, (%rax,%rdi), %xmm0, %xmm0
+; AVX-NEXT:    movsbq %r9b, %rax
+; AVX-NEXT:    vpinsrb $5, (%rax,%rdi), %xmm0, %xmm0
+; AVX-NEXT:    movsbq {{[0-9]+}}(%rsp), %rax
+; AVX-NEXT:    vpinsrb $6, (%rax,%rdi), %xmm0, %xmm0
+; AVX-NEXT:    movsbq {{[0-9]+}}(%rsp), %rax
+; AVX-NEXT:    vpinsrb $7, (%rax,%rdi), %xmm0, %xmm0
+; AVX-NEXT:    movsbq {{[0-9]+}}(%rsp), %rax
+; AVX-NEXT:    vpinsrb $8, (%rax,%rdi), %xmm0, %xmm0
+; AVX-NEXT:    movsbq {{[0-9]+}}(%rsp), %rax
+; AVX-NEXT:    vpinsrb $9, (%rax,%rdi), %xmm0, %xmm0
+; AVX-NEXT:    movsbq {{[0-9]+}}(%rsp), %rax
+; AVX-NEXT:    vpinsrb $10, (%rax,%rdi), %xmm0, %xmm0
+; AVX-NEXT:    movsbq {{[0-9]+}}(%rsp), %rax
+; AVX-NEXT:    vpinsrb $11, (%rax,%rdi), %xmm0, %xmm0
+; AVX-NEXT:    movsbq {{[0-9]+}}(%rsp), %rax
+; AVX-NEXT:    vpinsrb $12, (%rax,%rdi), %xmm0, %xmm0
+; AVX-NEXT:    movsbq {{[0-9]+}}(%rsp), %rax
+; AVX-NEXT:    vpinsrb $13, (%rax,%rdi), %xmm0, %xmm0
+; AVX-NEXT:    movsbq {{[0-9]+}}(%rsp), %rax
+; AVX-NEXT:    vpinsrb $14, (%rax,%rdi), %xmm0, %xmm0
+; AVX-NEXT:    movsbq {{[0-9]+}}(%rsp), %rax
+; AVX-NEXT:    vpinsrb $15, (%rax,%rdi), %xmm0, %xmm0
 ; AVX-NEXT:    retq
   %x0  = extractelement <16 x i8> %x, i8 %i0
   %x1  = extractelement <16 x i8> %x, i8 %i1
@@ -697,11 +647,11 @@ define <4 x i32> @mem_shuffle_v4i32_v4i32_xxxx_i32(<4 x i32> %x, i32* %i) nounwi
 ; SSE2-NEXT:    movslq 12(%rdi), %rsi
 ; SSE2-NEXT:    movd {{.*#+}} xmm0 = mem[0],zero,zero,zero
 ; SSE2-NEXT:    movd {{.*#+}} xmm1 = mem[0],zero,zero,zero
+; SSE2-NEXT:    punpckldq {{.*#+}} xmm1 = xmm1[0],xmm0[0],xmm1[1],xmm0[1]
 ; SSE2-NEXT:    movd {{.*#+}} xmm2 = mem[0],zero,zero,zero
+; SSE2-NEXT:    movd {{.*#+}} xmm0 = mem[0],zero,zero,zero
 ; SSE2-NEXT:    punpckldq {{.*#+}} xmm0 = xmm0[0],xmm2[0],xmm0[1],xmm2[1]
-; SSE2-NEXT:    movd {{.*#+}} xmm2 = mem[0],zero,zero,zero
-; SSE2-NEXT:    punpckldq {{.*#+}} xmm2 = xmm2[0],xmm1[0],xmm2[1],xmm1[1]
-; SSE2-NEXT:    punpckldq {{.*#+}} xmm0 = xmm0[0],xmm2[0],xmm0[1],xmm2[1]
+; SSE2-NEXT:    punpckldq {{.*#+}} xmm0 = xmm0[0],xmm1[0],xmm0[1],xmm1[1]
 ; SSE2-NEXT:    retq
 ;
 ; SSSE3-LABEL: mem_shuffle_v4i32_v4i32_xxxx_i32:
@@ -713,11 +663,11 @@ define <4 x i32> @mem_shuffle_v4i32_v4i32_xxxx_i32(<4 x i32> %x, i32* %i) nounwi
 ; SSSE3-NEXT:    movslq 12(%rdi), %rsi
 ; SSSE3-NEXT:    movd {{.*#+}} xmm0 = mem[0],zero,zero,zero
 ; SSSE3-NEXT:    movd {{.*#+}} xmm1 = mem[0],zero,zero,zero
+; SSSE3-NEXT:    punpckldq {{.*#+}} xmm1 = xmm1[0],xmm0[0],xmm1[1],xmm0[1]
 ; SSSE3-NEXT:    movd {{.*#+}} xmm2 = mem[0],zero,zero,zero
+; SSSE3-NEXT:    movd {{.*#+}} xmm0 = mem[0],zero,zero,zero
 ; SSSE3-NEXT:    punpckldq {{.*#+}} xmm0 = xmm0[0],xmm2[0],xmm0[1],xmm2[1]
-; SSSE3-NEXT:    movd {{.*#+}} xmm2 = mem[0],zero,zero,zero
-; SSSE3-NEXT:    punpckldq {{.*#+}} xmm2 = xmm2[0],xmm1[0],xmm2[1],xmm1[1]
-; SSSE3-NEXT:    punpckldq {{.*#+}} xmm0 = xmm0[0],xmm2[0],xmm0[1],xmm2[1]
+; SSSE3-NEXT:    punpckldq {{.*#+}} xmm0 = xmm0[0],xmm1[0],xmm0[1],xmm1[1]
 ; SSSE3-NEXT:    retq
 ;
 ; SSE41-LABEL: mem_shuffle_v4i32_v4i32_xxxx_i32:
@@ -767,270 +717,218 @@ define <4 x i32> @mem_shuffle_v4i32_v4i32_xxxx_i32(<4 x i32> %x, i32* %i) nounwi
 define <16 x i8> @mem_shuffle_v16i8_v16i8_xxxxxxxxxxxxxxxx_i8(<16 x i8> %x, i8* %i) nounwind {
 ; SSE2-LABEL: mem_shuffle_v16i8_v16i8_xxxxxxxxxxxxxxxx_i8:
 ; SSE2:       # BB#0:
-; SSE2-NEXT:    movsbq (%rdi), %rcx
+; SSE2-NEXT:    movsbq (%rdi), %rax
 ; SSE2-NEXT:    movaps %xmm0, -{{[0-9]+}}(%rsp)
-; SSE2-NEXT:    leaq -{{[0-9]+}}(%rsp), %rax
-; SSE2-NEXT:    movzbl (%rcx,%rax), %ecx
-; SSE2-NEXT:    movd %ecx, %xmm0
-; SSE2-NEXT:    movsbq 8(%rdi), %rcx
-; SSE2-NEXT:    movzbl (%rcx,%rax), %ecx
-; SSE2-NEXT:    movd %ecx, %xmm8
-; SSE2-NEXT:    movsbq 12(%rdi), %rcx
-; SSE2-NEXT:    movzbl (%rcx,%rax), %ecx
-; SSE2-NEXT:    movd %ecx, %xmm9
-; SSE2-NEXT:    movsbq 4(%rdi), %rcx
-; SSE2-NEXT:    movzbl (%rcx,%rax), %ecx
-; SSE2-NEXT:    movd %ecx, %xmm3
-; SSE2-NEXT:    movsbq 14(%rdi), %rcx
-; SSE2-NEXT:    movzbl (%rcx,%rax), %ecx
-; SSE2-NEXT:    movd %ecx, %xmm10
-; SSE2-NEXT:    movsbq 6(%rdi), %rcx
-; SSE2-NEXT:    movzbl (%rcx,%rax), %ecx
-; SSE2-NEXT:    movd %ecx, %xmm5
-; SSE2-NEXT:    movsbq 10(%rdi), %rcx
-; SSE2-NEXT:    movzbl (%rcx,%rax), %ecx
-; SSE2-NEXT:    movd %ecx, %xmm11
-; SSE2-NEXT:    movsbq 2(%rdi), %rcx
-; SSE2-NEXT:    movzbl (%rcx,%rax), %ecx
-; SSE2-NEXT:    movd %ecx, %xmm7
-; SSE2-NEXT:    movsbq 15(%rdi), %rcx
-; SSE2-NEXT:    movzbl (%rcx,%rax), %ecx
-; SSE2-NEXT:    movd %ecx, %xmm12
-; SSE2-NEXT:    movsbq 7(%rdi), %rcx
-; SSE2-NEXT:    movzbl (%rcx,%rax), %ecx
-; SSE2-NEXT:    movd %ecx, %xmm2
-; SSE2-NEXT:    movsbq 11(%rdi), %rcx
-; SSE2-NEXT:    movzbl (%rcx,%rax), %ecx
-; SSE2-NEXT:    movd %ecx, %xmm13
-; SSE2-NEXT:    movsbq 3(%rdi), %rcx
-; SSE2-NEXT:    movzbl (%rcx,%rax), %ecx
-; SSE2-NEXT:    movd %ecx, %xmm6
-; SSE2-NEXT:    movsbq 13(%rdi), %rcx
-; SSE2-NEXT:    movzbl (%rcx,%rax), %ecx
-; SSE2-NEXT:    movd %ecx, %xmm14
-; SSE2-NEXT:    movsbq 5(%rdi), %rcx
-; SSE2-NEXT:    movzbl (%rcx,%rax), %ecx
-; SSE2-NEXT:    movd %ecx, %xmm4
-; SSE2-NEXT:    movsbq 9(%rdi), %rcx
-; SSE2-NEXT:    movzbl (%rcx,%rax), %ecx
-; SSE2-NEXT:    movd %ecx, %xmm15
-; SSE2-NEXT:    movsbq 1(%rdi), %rcx
-; SSE2-NEXT:    movzbl (%rcx,%rax), %eax
-; SSE2-NEXT:    movd %eax, %xmm1
-; SSE2-NEXT:    punpcklbw {{.*#+}} xmm0 = xmm0[0],xmm8[0],xmm0[1],xmm8[1],xmm0[2],xmm8[2],xmm0[3],xmm8[3],xmm0[4],xmm8[4],xmm0[5],xmm8[5],xmm0[6],xmm8[6],xmm0[7],xmm8[7]
+; SSE2-NEXT:    movsbq 15(%rdi), %rdx
+; SSE2-NEXT:    leaq -{{[0-9]+}}(%rsp), %rcx
+; SSE2-NEXT:    movzbl (%rdx,%rcx), %edx
+; SSE2-NEXT:    movd %edx, %xmm8
+; SSE2-NEXT:    movsbq 7(%rdi), %rdx
+; SSE2-NEXT:    movzbl (%rdx,%rcx), %edx
+; SSE2-NEXT:    movd %edx, %xmm15
+; SSE2-NEXT:    movsbq 11(%rdi), %rdx
+; SSE2-NEXT:    movzbl (%rdx,%rcx), %edx
+; SSE2-NEXT:    movd %edx, %xmm9
+; SSE2-NEXT:    movsbq 3(%rdi), %rdx
+; SSE2-NEXT:    movzbl (%rdx,%rcx), %edx
+; SSE2-NEXT:    movd %edx, %xmm3
+; SSE2-NEXT:    movsbq 13(%rdi), %rdx
+; SSE2-NEXT:    movzbl (%rdx,%rcx), %edx
+; SSE2-NEXT:    movd %edx, %xmm10
+; SSE2-NEXT:    movsbq 5(%rdi), %rdx
+; SSE2-NEXT:    movzbl (%rdx,%rcx), %edx
+; SSE2-NEXT:    movd %edx, %xmm7
+; SSE2-NEXT:    movsbq 9(%rdi), %rdx
+; SSE2-NEXT:    movzbl (%rdx,%rcx), %edx
+; SSE2-NEXT:    movd %edx, %xmm11
+; SSE2-NEXT:    movsbq 1(%rdi), %rdx
+; SSE2-NEXT:    movzbl (%rdx,%rcx), %edx
+; SSE2-NEXT:    movd %edx, %xmm6
+; SSE2-NEXT:    movsbq 14(%rdi), %rdx
+; SSE2-NEXT:    movzbl (%rdx,%rcx), %edx
+; SSE2-NEXT:    movd %edx, %xmm12
+; SSE2-NEXT:    movsbq 6(%rdi), %rdx
+; SSE2-NEXT:    movzbl (%rdx,%rcx), %edx
+; SSE2-NEXT:    movd %edx, %xmm5
+; SSE2-NEXT:    movsbq 10(%rdi), %rdx
+; SSE2-NEXT:    movzbl (%rdx,%rcx), %edx
+; SSE2-NEXT:    movd %edx, %xmm13
+; SSE2-NEXT:    movsbq 2(%rdi), %rdx
+; SSE2-NEXT:    movzbl (%rdx,%rcx), %edx
+; SSE2-NEXT:    movd %edx, %xmm4
+; SSE2-NEXT:    movsbq 12(%rdi), %rdx
+; SSE2-NEXT:    movzbl (%rdx,%rcx), %edx
+; SSE2-NEXT:    movd %edx, %xmm14
+; SSE2-NEXT:    movsbq 4(%rdi), %rdx
+; SSE2-NEXT:    movzbl (%rdx,%rcx), %edx
+; SSE2-NEXT:    movd %edx, %xmm1
+; SSE2-NEXT:    movsbq 8(%rdi), %rdx
+; SSE2-NEXT:    movzbl (%rdx,%rcx), %edx
+; SSE2-NEXT:    movd %edx, %xmm2
+; SSE2-NEXT:    movzbl (%rax,%rcx), %eax
+; SSE2-NEXT:    movd %eax, %xmm0
+; SSE2-NEXT:    punpcklbw {{.*#+}} xmm15 = xmm15[0],xmm8[0],xmm15[1],xmm8[1],xmm15[2],xmm8[2],xmm15[3],xmm8[3],xmm15[4],xmm8[4],xmm15[5],xmm8[5],xmm15[6],xmm8[6],xmm15[7],xmm8[7]
 ; SSE2-NEXT:    punpcklbw {{.*#+}} xmm3 = xmm3[0],xmm9[0],xmm3[1],xmm9[1],xmm3[2],xmm9[2],xmm3[3],xmm9[3],xmm3[4],xmm9[4],xmm3[5],xmm9[5],xmm3[6],xmm9[6],xmm3[7],xmm9[7]
-; SSE2-NEXT:    punpcklbw {{.*#+}} xmm0 = xmm0[0],xmm3[0],xmm0[1],xmm3[1],xmm0[2],xmm3[2],xmm0[3],xmm3[3],xmm0[4],xmm3[4],xmm0[5],xmm3[5],xmm0[6],xmm3[6],xmm0[7],xmm3[7]
-; SSE2-NEXT:    punpcklbw {{.*#+}} xmm5 = xmm5[0],xmm10[0],xmm5[1],xmm10[1],xmm5[2],xmm10[2],xmm5[3],xmm10[3],xmm5[4],xmm10[4],xmm5[5],xmm10[5],xmm5[6],xmm10[6],xmm5[7],xmm10[7]
-; SSE2-NEXT:    punpcklbw {{.*#+}} xmm7 = xmm7[0],xmm11[0],xmm7[1],xmm11[1],xmm7[2],xmm11[2],xmm7[3],xmm11[3],xmm7[4],xmm11[4],xmm7[5],xmm11[5],xmm7[6],xmm11[6],xmm7[7],xmm11[7]
-; SSE2-NEXT:    punpcklbw {{.*#+}} xmm7 = xmm7[0],xmm5[0],xmm7[1],xmm5[1],xmm7[2],xmm5[2],xmm7[3],xmm5[3],xmm7[4],xmm5[4],xmm7[5],xmm5[5],xmm7[6],xmm5[6],xmm7[7],xmm5[7]
-; SSE2-NEXT:    punpcklbw {{.*#+}} xmm0 = xmm0[0],xmm7[0],xmm0[1],xmm7[1],xmm0[2],xmm7[2],xmm0[3],xmm7[3],xmm0[4],xmm7[4],xmm0[5],xmm7[5],xmm0[6],xmm7[6],xmm0[7],xmm7[7]
-; SSE2-NEXT:    punpcklbw {{.*#+}} xmm2 = xmm2[0],xmm12[0],xmm2[1],xmm12[1],xmm2[2],xmm12[2],xmm2[3],xmm12[3],xmm2[4],xmm12[4],xmm2[5],xmm12[5],xmm2[6],xmm12[6],xmm2[7],xmm12[7]
-; SSE2-NEXT:    punpcklbw {{.*#+}} xmm6 = xmm6[0],xmm13[0],xmm6[1],xmm13[1],xmm6[2],xmm13[2],xmm6[3],xmm13[3],xmm6[4],xmm13[4],xmm6[5],xmm13[5],xmm6[6],xmm13[6],xmm6[7],xmm13[7]
-; SSE2-NEXT:    punpcklbw {{.*#+}} xmm6 = xmm6[0],xmm2[0],xmm6[1],xmm2[1],xmm6[2],xmm2[2],xmm6[3],xmm2[3],xmm6[4],xmm2[4],xmm6[5],xmm2[5],xmm6[6],xmm2[6],xmm6[7],xmm2[7]
-; SSE2-NEXT:    punpcklbw {{.*#+}} xmm4 = xmm4[0],xmm14[0],xmm4[1],xmm14[1],xmm4[2],xmm14[2],xmm4[3],xmm14[3],xmm4[4],xmm14[4],xmm4[5],xmm14[5],xmm4[6],xmm14[6],xmm4[7],xmm14[7]
-; SSE2-NEXT:    punpcklbw {{.*#+}} xmm1 = xmm1[0],xmm15[0],xmm1[1],xmm15[1],xmm1[2],xmm15[2],xmm1[3],xmm15[3],xmm1[4],xmm15[4],xmm1[5],xmm15[5],xmm1[6],xmm15[6],xmm1[7],xmm15[7]
-; SSE2-NEXT:    punpcklbw {{.*#+}} xmm1 = xmm1[0],xmm4[0],xmm1[1],xmm4[1],xmm1[2],xmm4[2],xmm1[3],xmm4[3],xmm1[4],xmm4[4],xmm1[5],xmm4[5],xmm1[6],xmm4[6],xmm1[7],xmm4[7]
-; SSE2-NEXT:    punpcklbw {{.*#+}} xmm1 = xmm1[0],xmm6[0],xmm1[1],xmm6[1],xmm1[2],xmm6[2],xmm1[3],xmm6[3],xmm1[4],xmm6[4],xmm1[5],xmm6[5],xmm1[6],xmm6[6],xmm1[7],xmm6[7]
+; SSE2-NEXT:    punpcklbw {{.*#+}} xmm3 = xmm3[0],xmm15[0],xmm3[1],xmm15[1],xmm3[2],xmm15[2],xmm3[3],xmm15[3],xmm3[4],xmm15[4],xmm3[5],xmm15[5],xmm3[6],xmm15[6],xmm3[7],xmm15[7]
+; SSE2-NEXT:    punpcklbw {{.*#+}} xmm7 = xmm7[0],xmm10[0],xmm7[1],xmm10[1],xmm7[2],xmm10[2],xmm7[3],xmm10[3],xmm7[4],xmm10[4],xmm7[5],xmm10[5],xmm7[6],xmm10[6],xmm7[7],xmm10[7]
+; SSE2-NEXT:    punpcklbw {{.*#+}} xmm6 = xmm6[0],xmm11[0],xmm6[1],xmm11[1],xmm6[2],xmm11[2],xmm6[3],xmm11[3],xmm6[4],xmm11[4],xmm6[5],xmm11[5],xmm6[6],xmm11[6],xmm6[7],xmm11[7]
+; SSE2-NEXT:    punpcklbw {{.*#+}} xmm6 = xmm6[0],xmm7[0],xmm6[1],xmm7[1],xmm6[2],xmm7[2],xmm6[3],xmm7[3],xmm6[4],xmm7[4],xmm6[5],xmm7[5],xmm6[6],xmm7[6],xmm6[7],xmm7[7]
+; SSE2-NEXT:    punpcklbw {{.*#+}} xmm6 = xmm6[0],xmm3[0],xmm6[1],xmm3[1],xmm6[2],xmm3[2],xmm6[3],xmm3[3],xmm6[4],xmm3[4],xmm6[5],xmm3[5],xmm6[6],xmm3[6],xmm6[7],xmm3[7]
+; SSE2-NEXT:    punpcklbw {{.*#+}} xmm5 = xmm5[0],xmm12[0],xmm5[1],xmm12[1],xmm5[2],xmm12[2],xmm5[3],xmm12[3],xmm5[4],xmm12[4],xmm5[5],xmm12[5],xmm5[6],xmm12[6],xmm5[7],xmm12[7]
+; SSE2-NEXT:    punpcklbw {{.*#+}} xmm4 = xmm4[0],xmm13[0],xmm4[1],xmm13[1],xmm4[2],xmm13[2],xmm4[3],xmm13[3],xmm4[4],xmm13[4],xmm4[5],xmm13[5],xmm4[6],xmm13[6],xmm4[7],xmm13[7]
+; SSE2-NEXT:    punpcklbw {{.*#+}} xmm4 = xmm4[0],xmm5[0],xmm4[1],xmm5[1],xmm4[2],xmm5[2],xmm4[3],xmm5[3],xmm4[4],xmm5[4],xmm4[5],xmm5[5],xmm4[6],xmm5[6],xmm4[7],xmm5[7]
+; SSE2-NEXT:    punpcklbw {{.*#+}} xmm1 = xmm1[0],xmm14[0],xmm1[1],xmm14[1],xmm1[2],xmm14[2],xmm1[3],xmm14[3],xmm1[4],xmm14[4],xmm1[5],xmm14[5],xmm1[6],xmm14[6],xmm1[7],xmm14[7]
+; SSE2-NEXT:    punpcklbw {{.*#+}} xmm0 = xmm0[0],xmm2[0],xmm0[1],xmm2[1],xmm0[2],xmm2[2],xmm0[3],xmm2[3],xmm0[4],xmm2[4],xmm0[5],xmm2[5],xmm0[6],xmm2[6],xmm0[7],xmm2[7]
 ; SSE2-NEXT:    punpcklbw {{.*#+}} xmm0 = xmm0[0],xmm1[0],xmm0[1],xmm1[1],xmm0[2],xmm1[2],xmm0[3],xmm1[3],xmm0[4],xmm1[4],xmm0[5],xmm1[5],xmm0[6],xmm1[6],xmm0[7],xmm1[7]
+; SSE2-NEXT:    punpcklbw {{.*#+}} xmm0 = xmm0[0],xmm4[0],xmm0[1],xmm4[1],xmm0[2],xmm4[2],xmm0[3],xmm4[3],xmm0[4],xmm4[4],xmm0[5],xmm4[5],xmm0[6],xmm4[6],xmm0[7],xmm4[7]
+; SSE2-NEXT:    punpcklbw {{.*#+}} xmm0 = xmm0[0],xmm6[0],xmm0[1],xmm6[1],xmm0[2],xmm6[2],xmm0[3],xmm6[3],xmm0[4],xmm6[4],xmm0[5],xmm6[5],xmm0[6],xmm6[6],xmm0[7],xmm6[7]
 ; SSE2-NEXT:    retq
 ;
 ; SSSE3-LABEL: mem_shuffle_v16i8_v16i8_xxxxxxxxxxxxxxxx_i8:
 ; SSSE3:       # BB#0:
-; SSSE3-NEXT:    movsbq (%rdi), %rcx
+; SSSE3-NEXT:    movsbq (%rdi), %rax
 ; SSSE3-NEXT:    movaps %xmm0, -{{[0-9]+}}(%rsp)
-; SSSE3-NEXT:    leaq -{{[0-9]+}}(%rsp), %rax
-; SSSE3-NEXT:    movzbl (%rcx,%rax), %ecx
-; SSSE3-NEXT:    movd %ecx, %xmm0
-; SSSE3-NEXT:    movsbq 8(%rdi), %rcx
-; SSSE3-NEXT:    movzbl (%rcx,%rax), %ecx
-; SSSE3-NEXT:    movd %ecx, %xmm8
-; SSSE3-NEXT:    movsbq 12(%rdi), %rcx
-; SSSE3-NEXT:    movzbl (%rcx,%rax), %ecx
-; SSSE3-NEXT:    movd %ecx, %xmm9
-; SSSE3-NEXT:    movsbq 4(%rdi), %rcx
-; SSSE3-NEXT:    movzbl (%rcx,%rax), %ecx
-; SSSE3-NEXT:    movd %ecx, %xmm3
-; SSSE3-NEXT:    movsbq 14(%rdi), %rcx
-; SSSE3-NEXT:    movzbl (%rcx,%rax), %ecx
-; SSSE3-NEXT:    movd %ecx, %xmm10
-; SSSE3-NEXT:    movsbq 6(%rdi), %rcx
-; SSSE3-NEXT:    movzbl (%rcx,%rax), %ecx
-; SSSE3-NEXT:    movd %ecx, %xmm5
-; SSSE3-NEXT:    movsbq 10(%rdi), %rcx
-; SSSE3-NEXT:    movzbl (%rcx,%rax), %ecx
-; SSSE3-NEXT:    movd %ecx, %xmm11
-; SSSE3-NEXT:    movsbq 2(%rdi), %rcx
-; SSSE3-NEXT:    movzbl (%rcx,%rax), %ecx
-; SSSE3-NEXT:    movd %ecx, %xmm7
-; SSSE3-NEXT:    movsbq 15(%rdi), %rcx
-; SSSE3-NEXT:    movzbl (%rcx,%rax), %ecx
-; SSSE3-NEXT:    movd %ecx, %xmm12
-; SSSE3-NEXT:    movsbq 7(%rdi), %rcx
-; SSSE3-NEXT:    movzbl (%rcx,%rax), %ecx
-; SSSE3-NEXT:    movd %ecx, %xmm2
-; SSSE3-NEXT:    movsbq 11(%rdi), %rcx
-; SSSE3-NEXT:    movzbl (%rcx,%rax), %ecx
-; SSSE3-NEXT:    movd %ecx, %xmm13
-; SSSE3-NEXT:    movsbq 3(%rdi), %rcx
-; SSSE3-NEXT:    movzbl (%rcx,%rax), %ecx
-; SSSE3-NEXT:    movd %ecx, %xmm6
-; SSSE3-NEXT:    movsbq 13(%rdi), %rcx
-; SSSE3-NEXT:    movzbl (%rcx,%rax), %ecx
-; SSSE3-NEXT:    movd %ecx, %xmm14
-; SSSE3-NEXT:    movsbq 5(%rdi), %rcx
-; SSSE3-NEXT:    movzbl (%rcx,%rax), %ecx
-; SSSE3-NEXT:    movd %ecx, %xmm4
-; SSSE3-NEXT:    movsbq 9(%rdi), %rcx
-; SSSE3-NEXT:    movzbl (%rcx,%rax), %ecx
-; SSSE3-NEXT:    movd %ecx, %xmm15
-; SSSE3-NEXT:    movsbq 1(%rdi), %rcx
-; SSSE3-NEXT:    movzbl (%rcx,%rax), %eax
-; SSSE3-NEXT:    movd %eax, %xmm1
-; SSSE3-NEXT:    punpcklbw {{.*#+}} xmm0 = xmm0[0],xmm8[0],xmm0[1],xmm8[1],xmm0[2],xmm8[2],xmm0[3],xmm8[3],xmm0[4],xmm8[4],xmm0[5],xmm8[5],xmm0[6],xmm8[6],xmm0[7],xmm8[7]
+; SSSE3-NEXT:    movsbq 15(%rdi), %rdx
+; SSSE3-NEXT:    leaq -{{[0-9]+}}(%rsp), %rcx
+; SSSE3-NEXT:    movzbl (%rdx,%rcx), %edx
+; SSSE3-NEXT:    movd %edx, %xmm8
+; SSSE3-NEXT:    movsbq 7(%rdi), %rdx
+; SSSE3-NEXT:    movzbl (%rdx,%rcx), %edx
+; SSSE3-NEXT:    movd %edx, %xmm15
+; SSSE3-NEXT:    movsbq 11(%rdi), %rdx
+; SSSE3-NEXT:    movzbl (%rdx,%rcx), %edx
+; SSSE3-NEXT:    movd %edx, %xmm9
+; SSSE3-NEXT:    movsbq 3(%rdi), %rdx
+; SSSE3-NEXT:    movzbl (%rdx,%rcx), %edx
+; SSSE3-NEXT:    movd %edx, %xmm3
+; SSSE3-NEXT:    movsbq 13(%rdi), %rdx
+; SSSE3-NEXT:    movzbl (%rdx,%rcx), %edx
+; SSSE3-NEXT:    movd %edx, %xmm10
+; SSSE3-NEXT:    movsbq 5(%rdi), %rdx
+; SSSE3-NEXT:    movzbl (%rdx,%rcx), %edx
+; SSSE3-NEXT:    movd %edx, %xmm7
+; SSSE3-NEXT:    movsbq 9(%rdi), %rdx
+; SSSE3-NEXT:    movzbl (%rdx,%rcx), %edx
+; SSSE3-NEXT:    movd %edx, %xmm11
+; SSSE3-NEXT:    movsbq 1(%rdi), %rdx
+; SSSE3-NEXT:    movzbl (%rdx,%rcx), %edx
+; SSSE3-NEXT:    movd %edx, %xmm6
+; SSSE3-NEXT:    movsbq 14(%rdi), %rdx
+; SSSE3-NEXT:    movzbl (%rdx,%rcx), %edx
+; SSSE3-NEXT:    movd %edx, %xmm12
+; SSSE3-NEXT:    movsbq 6(%rdi), %rdx
+; SSSE3-NEXT:    movzbl (%rdx,%rcx), %edx
+; SSSE3-NEXT:    movd %edx, %xmm5
+; SSSE3-NEXT:    movsbq 10(%rdi), %rdx
+; SSSE3-NEXT:    movzbl (%rdx,%rcx), %edx
+; SSSE3-NEXT:    movd %edx, %xmm13
+; SSSE3-NEXT:    movsbq 2(%rdi), %rdx
+; SSSE3-NEXT:    movzbl (%rdx,%rcx), %edx
+; SSSE3-NEXT:    movd %edx, %xmm4
+; SSSE3-NEXT:    movsbq 12(%rdi), %rdx
+; SSSE3-NEXT:    movzbl (%rdx,%rcx), %edx
+; SSSE3-NEXT:    movd %edx, %xmm14
+; SSSE3-NEXT:    movsbq 4(%rdi), %rdx
+; SSSE3-NEXT:    movzbl (%rdx,%rcx), %edx
+; SSSE3-NEXT:    movd %edx, %xmm1
+; SSSE3-NEXT:    movsbq 8(%rdi), %rdx
+; SSSE3-NEXT:    movzbl (%rdx,%rcx), %edx
+; SSSE3-NEXT:    movd %edx, %xmm2
+; SSSE3-NEXT:    movzbl (%rax,%rcx), %eax
+; SSSE3-NEXT:    movd %eax, %xmm0
+; SSSE3-NEXT:    punpcklbw {{.*#+}} xmm15 = xmm15[0],xmm8[0],xmm15[1],xmm8[1],xmm15[2],xmm8[2],xmm15[3],xmm8[3],xmm15[4],xmm8[4],xmm15[5],xmm8[5],xmm15[6],xmm8[6],xmm15[7],xmm8[7]
 ; SSSE3-NEXT:    punpcklbw {{.*#+}} xmm3 = xmm3[0],xmm9[0],xmm3[1],xmm9[1],xmm3[2],xmm9[2],xmm3[3],xmm9[3],xmm3[4],xmm9[4],xmm3[5],xmm9[5],xmm3[6],xmm9[6],xmm3[7],xmm9[7]
-; SSSE3-NEXT:    punpcklbw {{.*#+}} xmm0 = xmm0[0],xmm3[0],xmm0[1],xmm3[1],xmm0[2],xmm3[2],xmm0[3],xmm3[3],xmm0[4],xmm3[4],xmm0[5],xmm3[5],xmm0[6],xmm3[6],xmm0[7],xmm3[7]
-; SSSE3-NEXT:    punpcklbw {{.*#+}} xmm5 = xmm5[0],xmm10[0],xmm5[1],xmm10[1],xmm5[2],xmm10[2],xmm5[3],xmm10[3],xmm5[4],xmm10[4],xmm5[5],xmm10[5],xmm5[6],xmm10[6],xmm5[7],xmm10[7]
-; SSSE3-NEXT:    punpcklbw {{.*#+}} xmm7 = xmm7[0],xmm11[0],xmm7[1],xmm11[1],xmm7[2],xmm11[2],xmm7[3],xmm11[3],xmm7[4],xmm11[4],xmm7[5],xmm11[5],xmm7[6],xmm11[6],xmm7[7],xmm11[7]
-; SSSE3-NEXT:    punpcklbw {{.*#+}} xmm7 = xmm7[0],xmm5[0],xmm7[1],xmm5[1],xmm7[2],xmm5[2],xmm7[3],xmm5[3],xmm7[4],xmm5[4],xmm7[5],xmm5[5],xmm7[6],xmm5[6],xmm7[7],xmm5[7]
-; SSSE3-NEXT:    punpcklbw {{.*#+}} xmm0 = xmm0[0],xmm7[0],xmm0[1],xmm7[1],xmm0[2],xmm7[2],xmm0[3],xmm7[3],xmm0[4],xmm7[4],xmm0[5],xmm7[5],xmm0[6],xmm7[6],xmm0[7],xmm7[7]
-; SSSE3-NEXT:    punpcklbw {{.*#+}} xmm2 = xmm2[0],xmm12[0],xmm2[1],xmm12[1],xmm2[2],xmm12[2],xmm2[3],xmm12[3],xmm2[4],xmm12[4],xmm2[5],xmm12[5],xmm2[6],xmm12[6],xmm2[7],xmm12[7]
-; SSSE3-NEXT:    punpcklbw {{.*#+}} xmm6 = xmm6[0],xmm13[0],xmm6[1],xmm13[1],xmm6[2],xmm13[2],xmm6[3],xmm13[3],xmm6[4],xmm13[4],xmm6[5],xmm13[5],xmm6[6],xmm13[6],xmm6[7],xmm13[7]
-; SSSE3-NEXT:    punpcklbw {{.*#+}} xmm6 = xmm6[0],xmm2[0],xmm6[1],xmm2[1],xmm6[2],xmm2[2],xmm6[3],xmm2[3],xmm6[4],xmm2[4],xmm6[5],xmm2[5],xmm6[6],xmm2[6],xmm6[7],xmm2[7]
-; SSSE3-NEXT:    punpcklbw {{.*#+}} xmm4 = xmm4[0],xmm14[0],xmm4[1],xmm14[1],xmm4[2],xmm14[2],xmm4[3],xmm14[3],xmm4[4],xmm14[4],xmm4[5],xmm14[5],xmm4[6],xmm14[6],xmm4[7],xmm14[7]
-; SSSE3-NEXT:    punpcklbw {{.*#+}} xmm1 = xmm1[0],xmm15[0],xmm1[1],xmm15[1],xmm1[2],xmm15[2],xmm1[3],xmm15[3],xmm1[4],xmm15[4],xmm1[5],xmm15[5],xmm1[6],xmm15[6],xmm1[7],xmm15[7]
-; SSSE3-NEXT:    punpcklbw {{.*#+}} xmm1 = xmm1[0],xmm4[0],xmm1[1],xmm4[1],xmm1[2],xmm4[2],xmm1[3],xmm4[3],xmm1[4],xmm4[4],xmm1[5],xmm4[5],xmm1[6],xmm4[6],xmm1[7],xmm4[7]
-; SSSE3-NEXT:    punpcklbw {{.*#+}} xmm1 = xmm1[0],xmm6[0],xmm1[1],xmm6[1],xmm1[2],xmm6[2],xmm1[3],xmm6[3],xmm1[4],xmm6[4],xmm1[5],xmm6[5],xmm1[6],xmm6[6],xmm1[7],xmm6[7]
+; SSSE3-NEXT:    punpcklbw {{.*#+}} xmm3 = xmm3[0],xmm15[0],xmm3[1],xmm15[1],xmm3[2],xmm15[2],xmm3[3],xmm15[3],xmm3[4],xmm15[4],xmm3[5],xmm15[5],xmm3[6],xmm15[6],xmm3[7],xmm15[7]
+; SSSE3-NEXT:    punpcklbw {{.*#+}} xmm7 = xmm7[0],xmm10[0],xmm7[1],xmm10[1],xmm7[2],xmm10[2],xmm7[3],xmm10[3],xmm7[4],xmm10[4],xmm7[5],xmm10[5],xmm7[6],xmm10[6],xmm7[7],xmm10[7]
+; SSSE3-NEXT:    punpcklbw {{.*#+}} xmm6 = xmm6[0],xmm11[0],xmm6[1],xmm11[1],xmm6[2],xmm11[2],xmm6[3],xmm11[3],xmm6[4],xmm11[4],xmm6[5],xmm11[5],xmm6[6],xmm11[6],xmm6[7],xmm11[7]
+; SSSE3-NEXT:    punpcklbw {{.*#+}} xmm6 = xmm6[0],xmm7[0],xmm6[1],xmm7[1],xmm6[2],xmm7[2],xmm6[3],xmm7[3],xmm6[4],xmm7[4],xmm6[5],xmm7[5],xmm6[6],xmm7[6],xmm6[7],xmm7[7]
+; SSSE3-NEXT:    punpcklbw {{.*#+}} xmm6 = xmm6[0],xmm3[0],xmm6[1],xmm3[1],xmm6[2],xmm3[2],xmm6[3],xmm3[3],xmm6[4],xmm3[4],xmm6[5],xmm3[5],xmm6[6],xmm3[6],xmm6[7],xmm3[7]
+; SSSE3-NEXT:    punpcklbw {{.*#+}} xmm5 = xmm5[0],xmm12[0],xmm5[1],xmm12[1],xmm5[2],xmm12[2],xmm5[3],xmm12[3],xmm5[4],xmm12[4],xmm5[5],xmm12[5],xmm5[6],xmm12[6],xmm5[7],xmm12[7]
+; SSSE3-NEXT:    punpcklbw {{.*#+}} xmm4 = xmm4[0],xmm13[0],xmm4[1],xmm13[1],xmm4[2],xmm13[2],xmm4[3],xmm13[3],xmm4[4],xmm13[4],xmm4[5],xmm13[5],xmm4[6],xmm13[6],xmm4[7],xmm13[7]
+; SSSE3-NEXT:    punpcklbw {{.*#+}} xmm4 = xmm4[0],xmm5[0],xmm4[1],xmm5[1],xmm4[2],xmm5[2],xmm4[3],xmm5[3],xmm4[4],xmm5[4],xmm4[5],xmm5[5],xmm4[6],xmm5[6],xmm4[7],xmm5[7]
+; SSSE3-NEXT:    punpcklbw {{.*#+}} xmm1 = xmm1[0],xmm14[0],xmm1[1],xmm14[1],xmm1[2],xmm14[2],xmm1[3],xmm14[3],xmm1[4],xmm14[4],xmm1[5],xmm14[5],xmm1[6],xmm14[6],xmm1[7],xmm14[7]
+; SSSE3-NEXT:    punpcklbw {{.*#+}} xmm0 = xmm0[0],xmm2[0],xmm0[1],xmm2[1],xmm0[2],xmm2[2],xmm0[3],xmm2[3],xmm0[4],xmm2[4],xmm0[5],xmm2[5],xmm0[6],xmm2[6],xmm0[7],xmm2[7]
 ; SSSE3-NEXT:    punpcklbw {{.*#+}} xmm0 = xmm0[0],xmm1[0],xmm0[1],xmm1[1],xmm0[2],xmm1[2],xmm0[3],xmm1[3],xmm0[4],xmm1[4],xmm0[5],xmm1[5],xmm0[6],xmm1[6],xmm0[7],xmm1[7]
+; SSSE3-NEXT:    punpcklbw {{.*#+}} xmm0 = xmm0[0],xmm4[0],xmm0[1],xmm4[1],xmm0[2],xmm4[2],xmm0[3],xmm4[3],xmm0[4],xmm4[4],xmm0[5],xmm4[5],xmm0[6],xmm4[6],xmm0[7],xmm4[7]
+; SSSE3-NEXT:    punpcklbw {{.*#+}} xmm0 = xmm0[0],xmm6[0],xmm0[1],xmm6[1],xmm0[2],xmm6[2],xmm0[3],xmm6[3],xmm0[4],xmm6[4],xmm0[5],xmm6[5],xmm0[6],xmm6[6],xmm0[7],xmm6[7]
 ; SSSE3-NEXT:    retq
 ;
 ; SSE41-LABEL: mem_shuffle_v16i8_v16i8_xxxxxxxxxxxxxxxx_i8:
 ; SSE41:       # BB#0:
-; SSE41-NEXT:    pushq %rbp
-; SSE41-NEXT:    pushq %r15
-; SSE41-NEXT:    pushq %r14
-; SSE41-NEXT:    pushq %r13
-; SSE41-NEXT:    pushq %r12
-; SSE41-NEXT:    pushq %rbx
-; SSE41-NEXT:    movsbq (%rdi), %rax
+; SSE41-NEXT:    movsbq (%rdi), %rcx
 ; SSE41-NEXT:    movaps %xmm0, -{{[0-9]+}}(%rsp)
-; SSE41-NEXT:    movsbq 1(%rdi), %r15
-; SSE41-NEXT:    movsbq 2(%rdi), %r8
-; SSE41-NEXT:    movsbq 3(%rdi), %r9
-; SSE41-NEXT:    movsbq 4(%rdi), %r10
-; SSE41-NEXT:    movsbq 5(%rdi), %r11
-; SSE41-NEXT:    movsbq 6(%rdi), %r14
-; SSE41-NEXT:    movsbq 7(%rdi), %r12
-; SSE41-NEXT:    movsbq 8(%rdi), %r13
-; SSE41-NEXT:    movsbq 9(%rdi), %rdx
+; SSE41-NEXT:    leaq -{{[0-9]+}}(%rsp), %rax
+; SSE41-NEXT:    movzbl (%rcx,%rax), %ecx
+; SSE41-NEXT:    movd %ecx, %xmm0
+; SSE41-NEXT:    movsbq 1(%rdi), %rcx
+; SSE41-NEXT:    pinsrb $1, (%rcx,%rax), %xmm0
+; SSE41-NEXT:    movsbq 2(%rdi), %rcx
+; SSE41-NEXT:    pinsrb $2, (%rcx,%rax), %xmm0
+; SSE41-NEXT:    movsbq 3(%rdi), %rcx
+; SSE41-NEXT:    pinsrb $3, (%rcx,%rax), %xmm0
+; SSE41-NEXT:    movsbq 4(%rdi), %rcx
+; SSE41-NEXT:    pinsrb $4, (%rcx,%rax), %xmm0
+; SSE41-NEXT:    movsbq 5(%rdi), %rcx
+; SSE41-NEXT:    pinsrb $5, (%rcx,%rax), %xmm0
+; SSE41-NEXT:    movsbq 6(%rdi), %rcx
+; SSE41-NEXT:    pinsrb $6, (%rcx,%rax), %xmm0
+; SSE41-NEXT:    movsbq 7(%rdi), %rcx
+; SSE41-NEXT:    pinsrb $7, (%rcx,%rax), %xmm0
+; SSE41-NEXT:    movsbq 8(%rdi), %rcx
+; SSE41-NEXT:    pinsrb $8, (%rcx,%rax), %xmm0
+; SSE41-NEXT:    movsbq 9(%rdi), %rcx
+; SSE41-NEXT:    pinsrb $9, (%rcx,%rax), %xmm0
 ; SSE41-NEXT:    movsbq 10(%rdi), %rcx
-; SSE41-NEXT:    movsbq 11(%rdi), %rsi
-; SSE41-NEXT:    movsbq 12(%rdi), %rbx
-; SSE41-NEXT:    leaq -{{[0-9]+}}(%rsp), %rbp
-; SSE41-NEXT:    movzbl (%rax,%rbp), %eax
-; SSE41-NEXT:    movd %eax, %xmm0
-; SSE41-NEXT:    movsbq 13(%rdi), %rax
-; SSE41-NEXT:    pinsrb $1, (%r15,%rbp), %xmm0
-; SSE41-NEXT:    movsbq 14(%rdi), %r15
-; SSE41-NEXT:    movsbq 15(%rdi), %rdi
-; SSE41-NEXT:    movzbl (%rdi,%rbp), %edi
-; SSE41-NEXT:    movzbl (%r15,%rbp), %r15d
-; SSE41-NEXT:    movzbl (%rax,%rbp), %eax
-; SSE41-NEXT:    movzbl (%rbx,%rbp), %ebx
-; SSE41-NEXT:    movzbl (%rsi,%rbp), %esi
-; SSE41-NEXT:    movzbl (%rcx,%rbp), %ecx
-; SSE41-NEXT:    movzbl (%rdx,%rbp), %edx
-; SSE41-NEXT:    movzbl (%r13,%rbp), %r13d
-; SSE41-NEXT:    movzbl (%r12,%rbp), %r12d
-; SSE41-NEXT:    movzbl (%r14,%rbp), %r14d
-; SSE41-NEXT:    movzbl (%r11,%rbp), %r11d
-; SSE41-NEXT:    movzbl (%r10,%rbp), %r10d
-; SSE41-NEXT:    movzbl (%r9,%rbp), %r9d
-; SSE41-NEXT:    movzbl (%r8,%rbp), %ebp
-; SSE41-NEXT:    pinsrb $2, %ebp, %xmm0
-; SSE41-NEXT:    pinsrb $3, %r9d, %xmm0
-; SSE41-NEXT:    pinsrb $4, %r10d, %xmm0
-; SSE41-NEXT:    pinsrb $5, %r11d, %xmm0
-; SSE41-NEXT:    pinsrb $6, %r14d, %xmm0
-; SSE41-NEXT:    pinsrb $7, %r12d, %xmm0
-; SSE41-NEXT:    pinsrb $8, %r13d, %xmm0
-; SSE41-NEXT:    pinsrb $9, %edx, %xmm0
-; SSE41-NEXT:    pinsrb $10, %ecx, %xmm0
-; SSE41-NEXT:    pinsrb $11, %esi, %xmm0
-; SSE41-NEXT:    pinsrb $12, %ebx, %xmm0
-; SSE41-NEXT:    pinsrb $13, %eax, %xmm0
-; SSE41-NEXT:    pinsrb $14, %r15d, %xmm0
-; SSE41-NEXT:    pinsrb $15, %edi, %xmm0
-; SSE41-NEXT:    popq %rbx
-; SSE41-NEXT:    popq %r12
-; SSE41-NEXT:    popq %r13
-; SSE41-NEXT:    popq %r14
-; SSE41-NEXT:    popq %r15
-; SSE41-NEXT:    popq %rbp
+; SSE41-NEXT:    pinsrb $10, (%rcx,%rax), %xmm0
+; SSE41-NEXT:    movsbq 11(%rdi), %rcx
+; SSE41-NEXT:    pinsrb $11, (%rcx,%rax), %xmm0
+; SSE41-NEXT:    movsbq 12(%rdi), %rcx
+; SSE41-NEXT:    pinsrb $12, (%rcx,%rax), %xmm0
+; SSE41-NEXT:    movsbq 13(%rdi), %rcx
+; SSE41-NEXT:    pinsrb $13, (%rcx,%rax), %xmm0
+; SSE41-NEXT:    movsbq 14(%rdi), %rcx
+; SSE41-NEXT:    pinsrb $14, (%rcx,%rax), %xmm0
+; SSE41-NEXT:    movsbq 15(%rdi), %rcx
+; SSE41-NEXT:    pinsrb $15, (%rcx,%rax), %xmm0
 ; SSE41-NEXT:    retq
 ;
 ; AVX-LABEL: mem_shuffle_v16i8_v16i8_xxxxxxxxxxxxxxxx_i8:
 ; AVX:       # BB#0:
-; AVX-NEXT:    pushq %rbp
-; AVX-NEXT:    pushq %r15
-; AVX-NEXT:    pushq %r14
-; AVX-NEXT:    pushq %r13
-; AVX-NEXT:    pushq %r12
-; AVX-NEXT:    pushq %rbx
-; AVX-NEXT:    movsbq (%rdi), %rsi
+; AVX-NEXT:    movsbq (%rdi), %rax
 ; AVX-NEXT:    vmovaps %xmm0, -{{[0-9]+}}(%rsp)
-; AVX-NEXT:    movsbq 1(%rdi), %r15
-; AVX-NEXT:    movsbq 2(%rdi), %r8
-; AVX-NEXT:    movsbq 3(%rdi), %r9
-; AVX-NEXT:    movsbq 4(%rdi), %r10
-; AVX-NEXT:    movsbq 5(%rdi), %r11
-; AVX-NEXT:    movsbq 6(%rdi), %r14
-; AVX-NEXT:    movsbq 7(%rdi), %r12
-; AVX-NEXT:    movsbq 8(%rdi), %r13
-; AVX-NEXT:    movsbq 9(%rdi), %rdx
+; AVX-NEXT:    leaq -{{[0-9]+}}(%rsp), %rcx
+; AVX-NEXT:    movzbl (%rax,%rcx), %eax
+; AVX-NEXT:    vmovd %eax, %xmm0
+; AVX-NEXT:    movsbq 1(%rdi), %rax
+; AVX-NEXT:    vpinsrb $1, (%rax,%rcx), %xmm0, %xmm0
+; AVX-NEXT:    movsbq 2(%rdi), %rax
+; AVX-NEXT:    vpinsrb $2, (%rax,%rcx), %xmm0, %xmm0
+; AVX-NEXT:    movsbq 3(%rdi), %rax
+; AVX-NEXT:    vpinsrb $3, (%rax,%rcx), %xmm0, %xmm0
+; AVX-NEXT:    movsbq 4(%rdi), %rax
+; AVX-NEXT:    vpinsrb $4, (%rax,%rcx), %xmm0, %xmm0
+; AVX-NEXT:    movsbq 5(%rdi), %rax
+; AVX-NEXT:    vpinsrb $5, (%rax,%rcx), %xmm0, %xmm0
+; AVX-NEXT:    movsbq 6(%rdi), %rax
+; AVX-NEXT:    vpinsrb $6, (%rax,%rcx), %xmm0, %xmm0
+; AVX-NEXT:    movsbq 7(%rdi), %rax
+; AVX-NEXT:    vpinsrb $7, (%rax,%rcx), %xmm0, %xmm0
+; AVX-NEXT:    movsbq 8(%rdi), %rax
+; AVX-NEXT:    vpinsrb $8, (%rax,%rcx), %xmm0, %xmm0
+; AVX-NEXT:    movsbq 9(%rdi), %rax
+; AVX-NEXT:    vpinsrb $9, (%rax,%rcx), %xmm0, %xmm0
 ; AVX-NEXT:    movsbq 10(%rdi), %rax
-; AVX-NEXT:    movsbq 11(%rdi), %rcx
-; AVX-NEXT:    movsbq 12(%rdi), %rbx
-; AVX-NEXT:    leaq -{{[0-9]+}}(%rsp), %rbp
-; AVX-NEXT:    movzbl (%rsi,%rbp), %esi
-; AVX-NEXT:    vmovd %esi, %xmm0
-; AVX-NEXT:    movsbq 13(%rdi), %rsi
-; AVX-NEXT:    vpinsrb $1, (%r15,%rbp), %xmm0, %xmm0
-; AVX-NEXT:    movsbq 14(%rdi), %r15
-; AVX-NEXT:    movsbq 15(%rdi), %rdi
-; AVX-NEXT:    movzbl (%rdi,%rbp), %edi
-; AVX-NEXT:    movzbl (%r15,%rbp), %r15d
-; AVX-NEXT:    movzbl (%rsi,%rbp), %esi
-; AVX-NEXT:    movzbl (%rbx,%rbp), %ebx
-; AVX-NEXT:    movzbl (%rcx,%rbp), %ecx
-; AVX-NEXT:    movzbl (%rax,%rbp), %eax
-; AVX-NEXT:    movzbl (%rdx,%rbp), %edx
-; AVX-NEXT:    movzbl (%r13,%rbp), %r13d
-; AVX-NEXT:    movzbl (%r12,%rbp), %r12d
-; AVX-NEXT:    movzbl (%r14,%rbp), %r14d
-; AVX-NEXT:    movzbl (%r11,%rbp), %r11d
-; AVX-NEXT:    movzbl (%r10,%rbp), %r10d
-; AVX-NEXT:    movzbl (%r9,%rbp), %r9d
-; AVX-NEXT:    movzbl (%r8,%rbp), %ebp
-; AVX-NEXT:    vpinsrb $2, %ebp, %xmm0, %xmm0
-; AVX-NEXT:    vpinsrb $3, %r9d, %xmm0, %xmm0
-; AVX-NEXT:    vpinsrb $4, %r10d, %xmm0, %xmm0
-; AVX-NEXT:    vpinsrb $5, %r11d, %xmm0, %xmm0
-; AVX-NEXT:    vpinsrb $6, %r14d, %xmm0, %xmm0
-; AVX-NEXT:    vpinsrb $7, %r12d, %xmm0, %xmm0
-; AVX-NEXT:    vpinsrb $8, %r13d, %xmm0, %xmm0
-; AVX-NEXT:    vpinsrb $9, %edx, %xmm0, %xmm0
-; AVX-NEXT:    vpinsrb $10, %eax, %xmm0, %xmm0
-; AVX-NEXT:    vpinsrb $11, %ecx, %xmm0, %xmm0
-; AVX-NEXT:    vpinsrb $12, %ebx, %xmm0, %xmm0
-; AVX-NEXT:    vpinsrb $13, %esi, %xmm0, %xmm0
-; AVX-NEXT:    vpinsrb $14, %r15d, %xmm0, %xmm0
-; AVX-NEXT:    vpinsrb $15, %edi, %xmm0, %xmm0
-; AVX-NEXT:    popq %rbx
-; AVX-NEXT:    popq %r12
-; AVX-NEXT:    popq %r13
-; AVX-NEXT:    popq %r14
-; AVX-NEXT:    popq %r15
-; AVX-NEXT:    popq %rbp
+; AVX-NEXT:    vpinsrb $10, (%rax,%rcx), %xmm0, %xmm0
+; AVX-NEXT:    movsbq 11(%rdi), %rax
+; AVX-NEXT:    vpinsrb $11, (%rax,%rcx), %xmm0, %xmm0
+; AVX-NEXT:    movsbq 12(%rdi), %rax
+; AVX-NEXT:    vpinsrb $12, (%rax,%rcx), %xmm0, %xmm0
+; AVX-NEXT:    movsbq 13(%rdi), %rax
+; AVX-NEXT:    vpinsrb $13, (%rax,%rcx), %xmm0, %xmm0
+; AVX-NEXT:    movsbq 14(%rdi), %rax
+; AVX-NEXT:    vpinsrb $14, (%rax,%rcx), %xmm0, %xmm0
+; AVX-NEXT:    movsbq 15(%rdi), %rax
+; AVX-NEXT:    vpinsrb $15, (%rax,%rcx), %xmm0, %xmm0
 ; AVX-NEXT:    retq
   %p0  = getelementptr inbounds i8, i8* %i, i64 0
   %p1  = getelementptr inbounds i8, i8* %i, i64 1
@@ -1159,27 +1057,27 @@ define <8 x i16> @var_shuffle_v8i16_v8i16_xyxyxy00_i16(<8 x i16> %x, <8 x i16> %
 ; SSE2-NEXT:    movswq %r8w, %rdi
 ; SSE2-NEXT:    movaps %xmm1, -{{[0-9]+}}(%rsp)
 ; SSE2-NEXT:    movswq %r9w, %rax
-; SSE2-NEXT:    movzwl -24(%rsp,%rsi,2), %esi
 ; SSE2-NEXT:    xorl %edx, %edx
 ; SSE2-NEXT:    movd %edx, %xmm0
 ; SSE2-NEXT:    movzwl -24(%rsp,%rcx,2), %ecx
 ; SSE2-NEXT:    movd %ecx, %xmm1
 ; SSE2-NEXT:    punpcklwd {{.*#+}} xmm1 = xmm1[0],xmm0[0],xmm1[1],xmm0[1],xmm1[2],xmm0[2],xmm1[3],xmm0[3]
-; SSE2-NEXT:    movd %esi, %xmm2
 ; SSE2-NEXT:    movzwl -24(%rsp,%rax,2), %eax
+; SSE2-NEXT:    movd %eax, %xmm2
+; SSE2-NEXT:    movzwl -24(%rsp,%rsi,2), %eax
 ; SSE2-NEXT:    movd %eax, %xmm3
-; SSE2-NEXT:    punpcklwd {{.*#+}} xmm2 = xmm2[0],xmm3[0],xmm2[1],xmm3[1],xmm2[2],xmm3[2],xmm2[3],xmm3[3]
-; SSE2-NEXT:    punpcklwd {{.*#+}} xmm2 = xmm2[0],xmm1[0],xmm2[1],xmm1[1],xmm2[2],xmm1[2],xmm2[3],xmm1[3]
-; SSE2-NEXT:    movzwl -40(%rsp,%r10,2), %eax
-; SSE2-NEXT:    movzwl -40(%rsp,%r11,2), %ecx
-; SSE2-NEXT:    movd %ecx, %xmm1
+; SSE2-NEXT:    punpcklwd {{.*#+}} xmm3 = xmm3[0],xmm2[0],xmm3[1],xmm2[1],xmm3[2],xmm2[2],xmm3[3],xmm2[3]
+; SSE2-NEXT:    punpcklwd {{.*#+}} xmm3 = xmm3[0],xmm1[0],xmm3[1],xmm1[1],xmm3[2],xmm1[2],xmm3[3],xmm1[3]
+; SSE2-NEXT:    movzwl -40(%rsp,%r11,2), %eax
+; SSE2-NEXT:    movd %eax, %xmm1
 ; SSE2-NEXT:    punpcklwd {{.*#+}} xmm1 = xmm1[0],xmm0[0],xmm1[1],xmm0[1],xmm1[2],xmm0[2],xmm1[3],xmm0[3]
-; SSE2-NEXT:    movd %eax, %xmm0
 ; SSE2-NEXT:    movzwl -40(%rsp,%rdi,2), %eax
-; SSE2-NEXT:    movd %eax, %xmm3
-; SSE2-NEXT:    punpcklwd {{.*#+}} xmm0 = xmm0[0],xmm3[0],xmm0[1],xmm3[1],xmm0[2],xmm3[2],xmm0[3],xmm3[3]
-; SSE2-NEXT:    punpcklwd {{.*#+}} xmm0 = xmm0[0],xmm1[0],xmm0[1],xmm1[1],xmm0[2],xmm1[2],xmm0[3],xmm1[3]
+; SSE2-NEXT:    movd %eax, %xmm2
+; SSE2-NEXT:    movzwl -40(%rsp,%r10,2), %eax
+; SSE2-NEXT:    movd %eax, %xmm0
 ; SSE2-NEXT:    punpcklwd {{.*#+}} xmm0 = xmm0[0],xmm2[0],xmm0[1],xmm2[1],xmm0[2],xmm2[2],xmm0[3],xmm2[3]
+; SSE2-NEXT:    punpcklwd {{.*#+}} xmm0 = xmm0[0],xmm1[0],xmm0[1],xmm1[1],xmm0[2],xmm1[2],xmm0[3],xmm1[3]
+; SSE2-NEXT:    punpcklwd {{.*#+}} xmm0 = xmm0[0],xmm3[0],xmm0[1],xmm3[1],xmm0[2],xmm3[2],xmm0[3],xmm3[3]
 ; SSE2-NEXT:    retq
 ;
 ; SSSE3-LABEL: var_shuffle_v8i16_v8i16_xyxyxy00_i16:
@@ -1198,27 +1096,27 @@ define <8 x i16> @var_shuffle_v8i16_v8i16_xyxyxy00_i16(<8 x i16> %x, <8 x i16> %
 ; SSSE3-NEXT:    movswq %r8w, %rdi
 ; SSSE3-NEXT:    movaps %xmm1, -{{[0-9]+}}(%rsp)
 ; SSSE3-NEXT:    movswq %r9w, %rax
-; SSSE3-NEXT:    movzwl -24(%rsp,%rsi,2), %esi
 ; SSSE3-NEXT:    xorl %edx, %edx
 ; SSSE3-NEXT:    movd %edx, %xmm0
 ; SSSE3-NEXT:    movzwl -24(%rsp,%rcx,2), %ecx
 ; SSSE3-NEXT:    movd %ecx, %xmm1
 ; SSSE3-NEXT:    punpcklwd {{.*#+}} xmm1 = xmm1[0],xmm0[0],xmm1[1],xmm0[1],xmm1[2],xmm0[2],xmm1[3],xmm0[3]
-; SSSE3-NEXT:    movd %esi, %xmm2
 ; SSSE3-NEXT:    movzwl -24(%rsp,%rax,2), %eax
+; SSSE3-NEXT:    movd %eax, %xmm2
+; SSSE3-NEXT:    movzwl -24(%rsp,%rsi,2), %eax
 ; SSSE3-NEXT:    movd %eax, %xmm3
-; SSSE3-NEXT:    punpcklwd {{.*#+}} xmm2 = xmm2[0],xmm3[0],xmm2[1],xmm3[1],xmm2[2],xmm3[2],xmm2[3],xmm3[3]
-; SSSE3-NEXT:    punpcklwd {{.*#+}} xmm2 = xmm2[0],xmm1[0],xmm2[1],xmm1[1],xmm2[2],xmm1[2],xmm2[3],xmm1[3]
-; SSSE3-NEXT:    movzwl -40(%rsp,%r10,2), %eax
-; SSSE3-NEXT:    movzwl -40(%rsp,%r11,2), %ecx
-; SSSE3-NEXT:    movd %ecx, %xmm1
+; SSSE3-NEXT:    punpcklwd {{.*#+}} xmm3 = xmm3[0],xmm2[0],xmm3[1],xmm2[1],xmm3[2],xmm2[2],xmm3[3],xmm2[3]
+; SSSE3-NEXT:    punpcklwd {{.*#+}} xmm3 = xmm3[0],xmm1[0],xmm3[1],xmm1[1],xmm3[2],xmm1[2],xmm3[3],xmm1[3]
+; SSSE3-NEXT:    movzwl -40(%rsp,%r11,2), %eax
+; SSSE3-NEXT:    movd %eax, %xmm1
 ; SSSE3-NEXT:    punpcklwd {{.*#+}} xmm1 = xmm1[0],xmm0[0],xmm1[1],xmm0[1],xmm1[2],xmm0[2],xmm1[3],xmm0[3]
-; SSSE3-NEXT:    movd %eax, %xmm0
 ; SSSE3-NEXT:    movzwl -40(%rsp,%rdi,2), %eax
-; SSSE3-NEXT:    movd %eax, %xmm3
-; SSSE3-NEXT:    punpcklwd {{.*#+}} xmm0 = xmm0[0],xmm3[0],xmm0[1],xmm3[1],xmm0[2],xmm3[2],xmm0[3],xmm3[3]
-; SSSE3-NEXT:    punpcklwd {{.*#+}} xmm0 = xmm0[0],xmm1[0],xmm0[1],xmm1[1],xmm0[2],xmm1[2],xmm0[3],xmm1[3]
+; SSSE3-NEXT:    movd %eax, %xmm2
+; SSSE3-NEXT:    movzwl -40(%rsp,%r10,2), %eax
+; SSSE3-NEXT:    movd %eax, %xmm0
 ; SSSE3-NEXT:    punpcklwd {{.*#+}} xmm0 = xmm0[0],xmm2[0],xmm0[1],xmm2[1],xmm0[2],xmm2[2],xmm0[3],xmm2[3]
+; SSSE3-NEXT:    punpcklwd {{.*#+}} xmm0 = xmm0[0],xmm1[0],xmm0[1],xmm1[1],xmm0[2],xmm1[2],xmm0[3],xmm1[3]
+; SSSE3-NEXT:    punpcklwd {{.*#+}} xmm0 = xmm0[0],xmm3[0],xmm0[1],xmm3[1],xmm0[2],xmm3[2],xmm0[3],xmm3[3]
 ; SSSE3-NEXT:    retq
 ;
 ; SSE41-LABEL: var_shuffle_v8i16_v8i16_xyxyxy00_i16:
diff --git a/llvm/test/CodeGen/X86/vector-shuffle-variable-256.ll b/llvm/test/CodeGen/X86/vector-shuffle-variable-256.ll
index bd59328aaf8..a34c596ced4 100644
--- a/llvm/test/CodeGen/X86/vector-shuffle-variable-256.ll
+++ b/llvm/test/CodeGen/X86/vector-shuffle-variable-256.ll
@@ -1,4 +1,4 @@
-; NOTE: Assertions have been autogenerated by utils/update_llc_test_checks.py
+; NOTE: Assertions have been autogenerated by update_llc_test_checks.py
 ; RUN: llc < %s -mtriple=x86_64-unknown-unknown -mcpu=x86-64 -mattr=+avx | FileCheck %s --check-prefix=ALL --check-prefix=AVX --check-prefix=AVX1
 ; RUN: llc < %s -mtriple=x86_64-unknown-unknown -mcpu=x86-64 -mattr=+avx2 | FileCheck %s --check-prefix=ALL --check-prefix=AVX --check-prefix=AVX2
 
@@ -18,7 +18,7 @@ define <4 x double> @var_shuffle_v4f64_v4f64_xxxx_i64(<4 x double> %x, i64 %i0,
 ; ALL-NEXT:    vmovhpd {{.*#+}} xmm0 = xmm0[0],mem[0]
 ; ALL-NEXT:    vmovsd {{.*#+}} xmm1 = mem[0],zero
 ; ALL-NEXT:    vmovhpd {{.*#+}} xmm1 = xmm1[0],mem[0]
-; ALL-NEXT:    vinsertf128 $1, %xmm1, %ymm0, %ymm0
+; ALL-NEXT:    vinsertf128 $1, %xmm0, %ymm1, %ymm0
 ; ALL-NEXT:    movq %rbp, %rsp
 ; ALL-NEXT:    popq %rbp
 ; ALL-NEXT:    retq
@@ -67,7 +67,7 @@ define <4 x double> @var_shuffle_v4f64_v2f64_xxxx_i64(<2 x double> %x, i64 %i0,
 ; ALL-NEXT:    vmovhpd {{.*#+}} xmm0 = xmm0[0],mem[0]
 ; ALL-NEXT:    vmovsd {{.*#+}} xmm1 = mem[0],zero
 ; ALL-NEXT:    vmovhpd {{.*#+}} xmm1 = xmm1[0],mem[0]
-; ALL-NEXT:    vinsertf128 $1, %xmm1, %ymm0, %ymm0
+; ALL-NEXT:    vinsertf128 $1, %xmm0, %ymm1, %ymm0
 ; ALL-NEXT:    retq
   %x0 = extractelement <2 x double> %x, i64 %i0
   %x1 = extractelement <2 x double> %x, i64 %i1
@@ -90,11 +90,11 @@ define <4 x i64> @var_shuffle_v4i64_v4i64_xxxx_i64(<4 x i64> %x, i64 %i0, i64 %i
 ; AVX1-NEXT:    vmovaps %ymm0, (%rsp)
 ; AVX1-NEXT:    vmovq {{.*#+}} xmm0 = mem[0],zero
 ; AVX1-NEXT:    vmovq {{.*#+}} xmm1 = mem[0],zero
-; AVX1-NEXT:    vpunpcklqdq {{.*#+}} xmm0 = xmm0[0],xmm1[0]
+; AVX1-NEXT:    vpunpcklqdq {{.*#+}} xmm0 = xmm1[0],xmm0[0]
 ; AVX1-NEXT:    vmovq {{.*#+}} xmm1 = mem[0],zero
 ; AVX1-NEXT:    vmovq {{.*#+}} xmm2 = mem[0],zero
-; AVX1-NEXT:    vpunpcklqdq {{.*#+}} xmm1 = xmm1[0],xmm2[0]
-; AVX1-NEXT:    vinsertf128 $1, %xmm1, %ymm0, %ymm0
+; AVX1-NEXT:    vpunpcklqdq {{.*#+}} xmm1 = xmm2[0],xmm1[0]
+; AVX1-NEXT:    vinsertf128 $1, %xmm0, %ymm1, %ymm0
 ; AVX1-NEXT:    movq %rbp, %rsp
 ; AVX1-NEXT:    popq %rbp
 ; AVX1-NEXT:    retq
@@ -108,11 +108,11 @@ define <4 x i64> @var_shuffle_v4i64_v4i64_xxxx_i64(<4 x i64> %x, i64 %i0, i64 %i
 ; AVX2-NEXT:    vmovaps %ymm0, (%rsp)
 ; AVX2-NEXT:    vmovq {{.*#+}} xmm0 = mem[0],zero
 ; AVX2-NEXT:    vmovq {{.*#+}} xmm1 = mem[0],zero
-; AVX2-NEXT:    vpunpcklqdq {{.*#+}} xmm0 = xmm0[0],xmm1[0]
+; AVX2-NEXT:    vpunpcklqdq {{.*#+}} xmm0 = xmm1[0],xmm0[0]
 ; AVX2-NEXT:    vmovq {{.*#+}} xmm1 = mem[0],zero
 ; AVX2-NEXT:    vmovq {{.*#+}} xmm2 = mem[0],zero
-; AVX2-NEXT:    vpunpcklqdq {{.*#+}} xmm1 = xmm1[0],xmm2[0]
-; AVX2-NEXT:    vinserti128 $1, %xmm1, %ymm0, %ymm0
+; AVX2-NEXT:    vpunpcklqdq {{.*#+}} xmm1 = xmm2[0],xmm1[0]
+; AVX2-NEXT:    vinserti128 $1, %xmm0, %ymm1, %ymm0
 ; AVX2-NEXT:    movq %rbp, %rsp
 ; AVX2-NEXT:    popq %rbp
 ; AVX2-NEXT:    retq
@@ -137,7 +137,7 @@ define <4 x i64> @var_shuffle_v4i64_v4i64_xx00_i64(<4 x i64> %x, i64 %i0, i64 %i
 ; AVX1-NEXT:    vmovaps %ymm0, (%rsp)
 ; AVX1-NEXT:    vmovq {{.*#+}} xmm0 = mem[0],zero
 ; AVX1-NEXT:    vmovq {{.*#+}} xmm1 = mem[0],zero
-; AVX1-NEXT:    vpunpcklqdq {{.*#+}} xmm0 = xmm0[0],xmm1[0]
+; AVX1-NEXT:    vpunpcklqdq {{.*#+}} xmm0 = xmm1[0],xmm0[0]
 ; AVX1-NEXT:    vpxor %xmm1, %xmm1, %xmm1
 ; AVX1-NEXT:    vinsertf128 $1, %xmm1, %ymm0, %ymm0
 ; AVX1-NEXT:    movq %rbp, %rsp
@@ -153,7 +153,7 @@ define <4 x i64> @var_shuffle_v4i64_v4i64_xx00_i64(<4 x i64> %x, i64 %i0, i64 %i
 ; AVX2-NEXT:    vmovaps %ymm0, (%rsp)
 ; AVX2-NEXT:    vmovq {{.*#+}} xmm0 = mem[0],zero
 ; AVX2-NEXT:    vmovq {{.*#+}} xmm1 = mem[0],zero
-; AVX2-NEXT:    vpunpcklqdq {{.*#+}} xmm0 = xmm0[0],xmm1[0]
+; AVX2-NEXT:    vpunpcklqdq {{.*#+}} xmm0 = xmm1[0],xmm0[0]
 ; AVX2-NEXT:    vpxor %xmm1, %xmm1, %xmm1
 ; AVX2-NEXT:    vinserti128 $1, %xmm1, %ymm0, %ymm0
 ; AVX2-NEXT:    movq %rbp, %rsp
@@ -176,11 +176,11 @@ define <4 x i64> @var_shuffle_v4i64_v2i64_xxxx_i64(<2 x i64> %x, i64 %i0, i64 %i
 ; AVX1-NEXT:    vmovaps %xmm0, -{{[0-9]+}}(%rsp)
 ; AVX1-NEXT:    vmovq {{.*#+}} xmm0 = mem[0],zero
 ; AVX1-NEXT:    vmovq {{.*#+}} xmm1 = mem[0],zero
-; AVX1-NEXT:    vpunpcklqdq {{.*#+}} xmm0 = xmm0[0],xmm1[0]
+; AVX1-NEXT:    vpunpcklqdq {{.*#+}} xmm0 = xmm1[0],xmm0[0]
 ; AVX1-NEXT:    vmovq {{.*#+}} xmm1 = mem[0],zero
 ; AVX1-NEXT:    vmovq {{.*#+}} xmm2 = mem[0],zero
-; AVX1-NEXT:    vpunpcklqdq {{.*#+}} xmm1 = xmm1[0],xmm2[0]
-; AVX1-NEXT:    vinsertf128 $1, %xmm1, %ymm0, %ymm0
+; AVX1-NEXT:    vpunpcklqdq {{.*#+}} xmm1 = xmm2[0],xmm1[0]
+; AVX1-NEXT:    vinsertf128 $1, %xmm0, %ymm1, %ymm0
 ; AVX1-NEXT:    retq
 ;
 ; AVX2-LABEL: var_shuffle_v4i64_v2i64_xxxx_i64:
@@ -188,11 +188,11 @@ define <4 x i64> @var_shuffle_v4i64_v2i64_xxxx_i64(<2 x i64> %x, i64 %i0, i64 %i
 ; AVX2-NEXT:    vmovaps %xmm0, -{{[0-9]+}}(%rsp)
 ; AVX2-NEXT:    vmovq {{.*#+}} xmm0 = mem[0],zero
 ; AVX2-NEXT:    vmovq {{.*#+}} xmm1 = mem[0],zero
-; AVX2-NEXT:    vpunpcklqdq {{.*#+}} xmm0 = xmm0[0],xmm1[0]
+; AVX2-NEXT:    vpunpcklqdq {{.*#+}} xmm0 = xmm1[0],xmm0[0]
 ; AVX2-NEXT:    vmovq {{.*#+}} xmm1 = mem[0],zero
 ; AVX2-NEXT:    vmovq {{.*#+}} xmm2 = mem[0],zero
-; AVX2-NEXT:    vpunpcklqdq {{.*#+}} xmm1 = xmm1[0],xmm2[0]
-; AVX2-NEXT:    vinserti128 $1, %xmm1, %ymm0, %ymm0
+; AVX2-NEXT:    vpunpcklqdq {{.*#+}} xmm1 = xmm2[0],xmm1[0]
+; AVX2-NEXT:    vinserti128 $1, %xmm0, %ymm1, %ymm0
 ; AVX2-NEXT:    retq
   %x0 = extractelement <2 x i64> %x, i64 %i0
   %x1 = extractelement <2 x i64> %x, i64 %i1
@@ -210,29 +210,29 @@ define <8 x float> @var_shuffle_v8f32_v8f32_xxxxxxxx_i32(<8 x float> %x, i32 %i0
 ; AVX1:       # BB#0:
 ; AVX1-NEXT:    pushq %rbp
 ; AVX1-NEXT:    movq %rsp, %rbp
+; AVX1-NEXT:    pushq %rbx
 ; AVX1-NEXT:    andq $-32, %rsp
 ; AVX1-NEXT:    subq $64, %rsp
 ; AVX1-NEXT:    movslq %edi, %rax
-; AVX1-NEXT:    movslq %esi, %rsi
-; AVX1-NEXT:    movslq %edx, %rdx
-; AVX1-NEXT:    movslq %ecx, %r11
-; AVX1-NEXT:    movslq %r8d, %r10
+; AVX1-NEXT:    movslq %esi, %rbx
+; AVX1-NEXT:    movslq %edx, %r11
+; AVX1-NEXT:    movslq %ecx, %r10
+; AVX1-NEXT:    movslq %r8d, %rdi
 ; AVX1-NEXT:    vmovaps %ymm0, (%rsp)
-; AVX1-NEXT:    movslq %r9d, %r8
-; AVX1-NEXT:    movslq 16(%rbp), %rdi
-; AVX1-NEXT:    movslq 24(%rbp), %rcx
+; AVX1-NEXT:    movslq %r9d, %rcx
+; AVX1-NEXT:    movslq 16(%rbp), %rdx
+; AVX1-NEXT:    movslq 24(%rbp), %rsi
 ; AVX1-NEXT:    vmovss {{.*#+}} xmm0 = mem[0],zero,zero,zero
+; AVX1-NEXT:    vinsertps {{.*#+}} xmm0 = xmm0[0],mem[0],xmm0[2,3]
+; AVX1-NEXT:    vinsertps {{.*#+}} xmm0 = xmm0[0,1],mem[0],xmm0[3]
+; AVX1-NEXT:    vinsertps {{.*#+}} xmm0 = xmm0[0,1,2],mem[0]
 ; AVX1-NEXT:    vmovss {{.*#+}} xmm1 = mem[0],zero,zero,zero
-; AVX1-NEXT:    vmovss {{.*#+}} xmm2 = mem[0],zero,zero,zero
-; AVX1-NEXT:    vinsertps {{.*#+}} xmm2 = xmm2[0],mem[0],xmm2[2,3]
-; AVX1-NEXT:    vinsertps {{.*#+}} xmm2 = xmm2[0,1],mem[0],xmm2[3]
-; AVX1-NEXT:    vinsertps {{.*#+}} xmm2 = xmm2[0,1,2],mem[0]
-; AVX1-NEXT:    vmovss {{.*#+}} xmm3 = mem[0],zero,zero,zero
-; AVX1-NEXT:    vinsertps {{.*#+}} xmm3 = xmm3[0],mem[0],xmm3[2,3]
-; AVX1-NEXT:    vinsertps {{.*#+}} xmm0 = xmm3[0,1],xmm0[0],xmm3[3]
-; AVX1-NEXT:    vinsertps {{.*#+}} xmm0 = xmm0[0,1,2],xmm1[0]
-; AVX1-NEXT:    vinsertf128 $1, %xmm0, %ymm2, %ymm0
-; AVX1-NEXT:    movq %rbp, %rsp
+; AVX1-NEXT:    vinsertps {{.*#+}} xmm1 = xmm1[0],mem[0],xmm1[2,3]
+; AVX1-NEXT:    vinsertps {{.*#+}} xmm1 = xmm1[0,1],mem[0],xmm1[3]
+; AVX1-NEXT:    vinsertps {{.*#+}} xmm1 = xmm1[0,1,2],mem[0]
+; AVX1-NEXT:    vinsertf128 $1, %xmm0, %ymm1, %ymm0
+; AVX1-NEXT:    leaq -8(%rbp), %rsp
+; AVX1-NEXT:    popq %rbx
 ; AVX1-NEXT:    popq %rbp
 ; AVX1-NEXT:    retq
 ;
@@ -284,26 +284,26 @@ define <8 x float> @var_shuffle_v8f32_v8f32_xxxxxxxx_i32(<8 x float> %x, i32 %i0
 define <8 x float> @var_shuffle_v8f32_v4f32_xxxxxxxx_i32(<4 x float> %x, i32 %i0, i32 %i1, i32 %i2, i32 %i3, i32 %i4, i32 %i5, i32 %i6, i32 %i7) nounwind {
 ; ALL-LABEL: var_shuffle_v8f32_v4f32_xxxxxxxx_i32:
 ; ALL:       # BB#0:
+; ALL-NEXT:    pushq %rbx
 ; ALL-NEXT:    movslq %edi, %rax
-; ALL-NEXT:    movslq %esi, %rsi
-; ALL-NEXT:    movslq %edx, %rdx
-; ALL-NEXT:    movslq %ecx, %r11
-; ALL-NEXT:    movslq %r8d, %r10
+; ALL-NEXT:    movslq %esi, %rbx
+; ALL-NEXT:    movslq %edx, %r11
+; ALL-NEXT:    movslq %ecx, %r10
+; ALL-NEXT:    movslq %r8d, %rdi
 ; ALL-NEXT:    vmovaps %xmm0, -{{[0-9]+}}(%rsp)
-; ALL-NEXT:    movslq %r9d, %r8
-; ALL-NEXT:    movslq {{[0-9]+}}(%rsp), %rdi
-; ALL-NEXT:    movslq {{[0-9]+}}(%rsp), %rcx
+; ALL-NEXT:    movslq %r9d, %rcx
+; ALL-NEXT:    movslq {{[0-9]+}}(%rsp), %rdx
+; ALL-NEXT:    movslq {{[0-9]+}}(%rsp), %rsi
 ; ALL-NEXT:    vmovss {{.*#+}} xmm0 = mem[0],zero,zero,zero
+; ALL-NEXT:    vinsertps {{.*#+}} xmm0 = xmm0[0],mem[0],xmm0[2,3]
+; ALL-NEXT:    vinsertps {{.*#+}} xmm0 = xmm0[0,1],mem[0],xmm0[3]
+; ALL-NEXT:    vinsertps {{.*#+}} xmm0 = xmm0[0,1,2],mem[0]
 ; ALL-NEXT:    vmovss {{.*#+}} xmm1 = mem[0],zero,zero,zero
-; ALL-NEXT:    vmovss {{.*#+}} xmm2 = mem[0],zero,zero,zero
-; ALL-NEXT:    vinsertps {{.*#+}} xmm2 = xmm2[0],mem[0],xmm2[2,3]
-; ALL-NEXT:    vinsertps {{.*#+}} xmm2 = xmm2[0,1],mem[0],xmm2[3]
-; ALL-NEXT:    vinsertps {{.*#+}} xmm2 = xmm2[0,1,2],mem[0]
-; ALL-NEXT:    vmovss {{.*#+}} xmm3 = mem[0],zero,zero,zero
-; ALL-NEXT:    vinsertps {{.*#+}} xmm3 = xmm3[0],mem[0],xmm3[2,3]
-; ALL-NEXT:    vinsertps {{.*#+}} xmm0 = xmm3[0,1],xmm0[0],xmm3[3]
-; ALL-NEXT:    vinsertps {{.*#+}} xmm0 = xmm0[0,1,2],xmm1[0]
-; ALL-NEXT:    vinsertf128 $1, %xmm0, %ymm2, %ymm0
+; ALL-NEXT:    vinsertps {{.*#+}} xmm1 = xmm1[0],mem[0],xmm1[2,3]
+; ALL-NEXT:    vinsertps {{.*#+}} xmm1 = xmm1[0,1],mem[0],xmm1[3]
+; ALL-NEXT:    vinsertps {{.*#+}} xmm1 = xmm1[0,1,2],mem[0]
+; ALL-NEXT:    vinsertf128 $1, %xmm0, %ymm1, %ymm0
+; ALL-NEXT:    popq %rbx
 ; ALL-NEXT:    retq
   %x0 = extractelement <4 x float> %x, i32 %i0
   %x1 = extractelement <4 x float> %x, i32 %i1
@@ -336,26 +336,19 @@ define <16 x i16> @var_shuffle_v16i16_v16i16_xxxxxxxxxxxxxxxx_i16(<16 x i16> %x,
 ; AVX1-NEXT:    movzwl (%rsp,%rax,2), %eax
 ; AVX1-NEXT:    vmovd %eax, %xmm0
 ; AVX1-NEXT:    movslq 40(%rbp), %rax
-; AVX1-NEXT:    movzwl (%rsp,%rax,2), %eax
-; AVX1-NEXT:    vpinsrw $1, %eax, %xmm0, %xmm0
+; AVX1-NEXT:    vpinsrw $1, (%rsp,%rax,2), %xmm0, %xmm0
 ; AVX1-NEXT:    movslq 48(%rbp), %rax
-; AVX1-NEXT:    movzwl (%rsp,%rax,2), %eax
-; AVX1-NEXT:    vpinsrw $2, %eax, %xmm0, %xmm0
+; AVX1-NEXT:    vpinsrw $2, (%rsp,%rax,2), %xmm0, %xmm0
 ; AVX1-NEXT:    movslq 56(%rbp), %rax
-; AVX1-NEXT:    movzwl (%rsp,%rax,2), %eax
-; AVX1-NEXT:    vpinsrw $3, %eax, %xmm0, %xmm0
+; AVX1-NEXT:    vpinsrw $3, (%rsp,%rax,2), %xmm0, %xmm0
 ; AVX1-NEXT:    movslq 64(%rbp), %rax
-; AVX1-NEXT:    movzwl (%rsp,%rax,2), %eax
-; AVX1-NEXT:    vpinsrw $4, %eax, %xmm0, %xmm0
+; AVX1-NEXT:    vpinsrw $4, (%rsp,%rax,2), %xmm0, %xmm0
 ; AVX1-NEXT:    movslq 72(%rbp), %rax
-; AVX1-NEXT:    movzwl (%rsp,%rax,2), %eax
-; AVX1-NEXT:    vpinsrw $5, %eax, %xmm0, %xmm0
+; AVX1-NEXT:    vpinsrw $5, (%rsp,%rax,2), %xmm0, %xmm0
 ; AVX1-NEXT:    movslq 80(%rbp), %rax
-; AVX1-NEXT:    movzwl (%rsp,%rax,2), %eax
-; AVX1-NEXT:    vpinsrw $6, %eax, %xmm0, %xmm0
+; AVX1-NEXT:    vpinsrw $6, (%rsp,%rax,2), %xmm0, %xmm0
 ; AVX1-NEXT:    movslq 88(%rbp), %rax
-; AVX1-NEXT:    movzwl (%rsp,%rax,2), %eax
-; AVX1-NEXT:    vpinsrw $7, %eax, %xmm0, %xmm0
+; AVX1-NEXT:    vpinsrw $7, (%rsp,%rax,2), %xmm0, %xmm0
 ; AVX1-NEXT:    movslq %edi, %rax
 ; AVX1-NEXT:    movzwl (%rsp,%rax,2), %eax
 ; AVX1-NEXT:    vmovd %eax, %xmm1
@@ -370,11 +363,9 @@ define <16 x i16> @var_shuffle_v16i16_v16i16_xxxxxxxxxxxxxxxx_i16(<16 x i16> %x,
 ; AVX1-NEXT:    movslq %r9d, %rax
 ; AVX1-NEXT:    vpinsrw $5, (%rsp,%rax,2), %xmm1, %xmm1
 ; AVX1-NEXT:    movslq 16(%rbp), %rax
-; AVX1-NEXT:    movzwl (%rsp,%rax,2), %eax
-; AVX1-NEXT:    vpinsrw $6, %eax, %xmm1, %xmm1
+; AVX1-NEXT:    vpinsrw $6, (%rsp,%rax,2), %xmm1, %xmm1
 ; AVX1-NEXT:    movslq 24(%rbp), %rax
-; AVX1-NEXT:    movzwl (%rsp,%rax,2), %eax
-; AVX1-NEXT:    vpinsrw $7, %eax, %xmm1, %xmm1
+; AVX1-NEXT:    vpinsrw $7, (%rsp,%rax,2), %xmm1, %xmm1
 ; AVX1-NEXT:    vinsertf128 $1, %xmm0, %ymm1, %ymm0
 ; AVX1-NEXT:    movq %rbp, %rsp
 ; AVX1-NEXT:    popq %rbp
@@ -391,26 +382,19 @@ define <16 x i16> @var_shuffle_v16i16_v16i16_xxxxxxxxxxxxxxxx_i16(<16 x i16> %x,
 ; AVX2-NEXT:    movzwl (%rsp,%rax,2), %eax
 ; AVX2-NEXT:    vmovd %eax, %xmm0
 ; AVX2-NEXT:    movslq 40(%rbp), %rax
-; AVX2-NEXT:    movzwl (%rsp,%rax,2), %eax
-; AVX2-NEXT:    vpinsrw $1, %eax, %xmm0, %xmm0
+; AVX2-NEXT:    vpinsrw $1, (%rsp,%rax,2), %xmm0, %xmm0
 ; AVX2-NEXT:    movslq 48(%rbp), %rax
-; AVX2-NEXT:    movzwl (%rsp,%rax,2), %eax
-; AVX2-NEXT:    vpinsrw $2, %eax, %xmm0, %xmm0
+; AVX2-NEXT:    vpinsrw $2, (%rsp,%rax,2), %xmm0, %xmm0
 ; AVX2-NEXT:    movslq 56(%rbp), %rax
-; AVX2-NEXT:    movzwl (%rsp,%rax,2), %eax
-; AVX2-NEXT:    vpinsrw $3, %eax, %xmm0, %xmm0
+; AVX2-NEXT:    vpinsrw $3, (%rsp,%rax,2), %xmm0, %xmm0
 ; AVX2-NEXT:    movslq 64(%rbp), %rax
-; AVX2-NEXT:    movzwl (%rsp,%rax,2), %eax
-; AVX2-NEXT:    vpinsrw $4, %eax, %xmm0, %xmm0
+; AVX2-NEXT:    vpinsrw $4, (%rsp,%rax,2), %xmm0, %xmm0
 ; AVX2-NEXT:    movslq 72(%rbp), %rax
-; AVX2-NEXT:    movzwl (%rsp,%rax,2), %eax
-; AVX2-NEXT:    vpinsrw $5, %eax, %xmm0, %xmm0
+; AVX2-NEXT:    vpinsrw $5, (%rsp,%rax,2), %xmm0, %xmm0
 ; AVX2-NEXT:    movslq 80(%rbp), %rax
-; AVX2-NEXT:    movzwl (%rsp,%rax,2), %eax
-; AVX2-NEXT:    vpinsrw $6, %eax, %xmm0, %xmm0
+; AVX2-NEXT:    vpinsrw $6, (%rsp,%rax,2), %xmm0, %xmm0
 ; AVX2-NEXT:    movslq 88(%rbp), %rax
-; AVX2-NEXT:    movzwl (%rsp,%rax,2), %eax
-; AVX2-NEXT:    vpinsrw $7, %eax, %xmm0, %xmm0
+; AVX2-NEXT:    vpinsrw $7, (%rsp,%rax,2), %xmm0, %xmm0
 ; AVX2-NEXT:    movslq %edi, %rax
 ; AVX2-NEXT:    movzwl (%rsp,%rax,2), %eax
 ; AVX2-NEXT:    vmovd %eax, %xmm1
@@ -425,11 +409,9 @@ define <16 x i16> @var_shuffle_v16i16_v16i16_xxxxxxxxxxxxxxxx_i16(<16 x i16> %x,
 ; AVX2-NEXT:    movslq %r9d, %rax
 ; AVX2-NEXT:    vpinsrw $5, (%rsp,%rax,2), %xmm1, %xmm1
 ; AVX2-NEXT:    movslq 16(%rbp), %rax
-; AVX2-NEXT:    movzwl (%rsp,%rax,2), %eax
-; AVX2-NEXT:    vpinsrw $6, %eax, %xmm1, %xmm1
+; AVX2-NEXT:    vpinsrw $6, (%rsp,%rax,2), %xmm1, %xmm1
 ; AVX2-NEXT:    movslq 24(%rbp), %rax
-; AVX2-NEXT:    movzwl (%rsp,%rax,2), %eax
-; AVX2-NEXT:    vpinsrw $7, %eax, %xmm1, %xmm1
+; AVX2-NEXT:    vpinsrw $7, (%rsp,%rax,2), %xmm1, %xmm1
 ; AVX2-NEXT:    vinserti128 $1, %xmm0, %ymm1, %ymm0
 ; AVX2-NEXT:    movq %rbp, %rsp
 ; AVX2-NEXT:    popq %rbp
@@ -477,26 +459,19 @@ define <16 x i16> @var_shuffle_v16i16_v8i16_xxxxxxxxxxxxxxxx_i16(<8 x i16> %x, i
 ; AVX1-NEXT:    movzwl -24(%rsp,%rax,2), %eax
 ; AVX1-NEXT:    vmovd %eax, %xmm0
 ; AVX1-NEXT:    movslq {{[0-9]+}}(%rsp), %rax
-; AVX1-NEXT:    movzwl -24(%rsp,%rax,2), %eax
-; AVX1-NEXT:    vpinsrw $1, %eax, %xmm0, %xmm0
+; AVX1-NEXT:    vpinsrw $1, -24(%rsp,%rax,2), %xmm0, %xmm0
 ; AVX1-NEXT:    movslq {{[0-9]+}}(%rsp), %rax
-; AVX1-NEXT:    movzwl -24(%rsp,%rax,2), %eax
-; AVX1-NEXT:    vpinsrw $2, %eax, %xmm0, %xmm0
+; AVX1-NEXT:    vpinsrw $2, -24(%rsp,%rax,2), %xmm0, %xmm0
 ; AVX1-NEXT:    movslq {{[0-9]+}}(%rsp), %rax
-; AVX1-NEXT:    movzwl -24(%rsp,%rax,2), %eax
-; AVX1-NEXT:    vpinsrw $3, %eax, %xmm0, %xmm0
+; AVX1-NEXT:    vpinsrw $3, -24(%rsp,%rax,2), %xmm0, %xmm0
 ; AVX1-NEXT:    movslq {{[0-9]+}}(%rsp), %rax
-; AVX1-NEXT:    movzwl -24(%rsp,%rax,2), %eax
-; AVX1-NEXT:    vpinsrw $4, %eax, %xmm0, %xmm0
+; AVX1-NEXT:    vpinsrw $4, -24(%rsp,%rax,2), %xmm0, %xmm0
 ; AVX1-NEXT:    movslq {{[0-9]+}}(%rsp), %rax
-; AVX1-NEXT:    movzwl -24(%rsp,%rax,2), %eax
-; AVX1-NEXT:    vpinsrw $5, %eax, %xmm0, %xmm0
+; AVX1-NEXT:    vpinsrw $5, -24(%rsp,%rax,2), %xmm0, %xmm0
 ; AVX1-NEXT:    movslq {{[0-9]+}}(%rsp), %rax
-; AVX1-NEXT:    movzwl -24(%rsp,%rax,2), %eax
-; AVX1-NEXT:    vpinsrw $6, %eax, %xmm0, %xmm0
+; AVX1-NEXT:    vpinsrw $6, -24(%rsp,%rax,2), %xmm0, %xmm0
 ; AVX1-NEXT:    movslq {{[0-9]+}}(%rsp), %rax
-; AVX1-NEXT:    movzwl -24(%rsp,%rax,2), %eax
-; AVX1-NEXT:    vpinsrw $7, %eax, %xmm0, %xmm0
+; AVX1-NEXT:    vpinsrw $7, -24(%rsp,%rax,2), %xmm0, %xmm0
 ; AVX1-NEXT:    movslq %edi, %rax
 ; AVX1-NEXT:    movzwl -24(%rsp,%rax,2), %eax
 ; AVX1-NEXT:    vmovd %eax, %xmm1
@@ -511,11 +486,9 @@ define <16 x i16> @var_shuffle_v16i16_v8i16_xxxxxxxxxxxxxxxx_i16(<8 x i16> %x, i
 ; AVX1-NEXT:    movslq %r9d, %rax
 ; AVX1-NEXT:    vpinsrw $5, -24(%rsp,%rax,2), %xmm1, %xmm1
 ; AVX1-NEXT:    movslq {{[0-9]+}}(%rsp), %rax
-; AVX1-NEXT:    movzwl -24(%rsp,%rax,2), %eax
-; AVX1-NEXT:    vpinsrw $6, %eax, %xmm1, %xmm1
+; AVX1-NEXT:    vpinsrw $6, -24(%rsp,%rax,2), %xmm1, %xmm1
 ; AVX1-NEXT:    movslq {{[0-9]+}}(%rsp), %rax
-; AVX1-NEXT:    movzwl -24(%rsp,%rax,2), %eax
-; AVX1-NEXT:    vpinsrw $7, %eax, %xmm1, %xmm1
+; AVX1-NEXT:    vpinsrw $7, -24(%rsp,%rax,2), %xmm1, %xmm1
 ; AVX1-NEXT:    vinsertf128 $1, %xmm0, %ymm1, %ymm0
 ; AVX1-NEXT:    retq
 ;
@@ -526,26 +499,19 @@ define <16 x i16> @var_shuffle_v16i16_v8i16_xxxxxxxxxxxxxxxx_i16(<8 x i16> %x, i
 ; AVX2-NEXT:    movzwl -24(%rsp,%rax,2), %eax
 ; AVX2-NEXT:    vmovd %eax, %xmm0
 ; AVX2-NEXT:    movslq {{[0-9]+}}(%rsp), %rax
-; AVX2-NEXT:    movzwl -24(%rsp,%rax,2), %eax
-; AVX2-NEXT:    vpinsrw $1, %eax, %xmm0, %xmm0
+; AVX2-NEXT:    vpinsrw $1, -24(%rsp,%rax,2), %xmm0, %xmm0
 ; AVX2-NEXT:    movslq {{[0-9]+}}(%rsp), %rax
-; AVX2-NEXT:    movzwl -24(%rsp,%rax,2), %eax
-; AVX2-NEXT:    vpinsrw $2, %eax, %xmm0, %xmm0
+; AVX2-NEXT:    vpinsrw $2, -24(%rsp,%rax,2), %xmm0, %xmm0
 ; AVX2-NEXT:    movslq {{[0-9]+}}(%rsp), %rax
-; AVX2-NEXT:    movzwl -24(%rsp,%rax,2), %eax
-; AVX2-NEXT:    vpinsrw $3, %eax, %xmm0, %xmm0
+; AVX2-NEXT:    vpinsrw $3, -24(%rsp,%rax,2), %xmm0, %xmm0
 ; AVX2-NEXT:    movslq {{[0-9]+}}(%rsp), %rax
-; AVX2-NEXT:    movzwl -24(%rsp,%rax,2), %eax
-; AVX2-NEXT:    vpinsrw $4, %eax, %xmm0, %xmm0
+; AVX2-NEXT:    vpinsrw $4, -24(%rsp,%rax,2), %xmm0, %xmm0
 ; AVX2-NEXT:    movslq {{[0-9]+}}(%rsp), %rax
-; AVX2-NEXT:    movzwl -24(%rsp,%rax,2), %eax
-; AVX2-NEXT:    vpinsrw $5, %eax, %xmm0, %xmm0
+; AVX2-NEXT:    vpinsrw $5, -24(%rsp,%rax,2), %xmm0, %xmm0
 ; AVX2-NEXT:    movslq {{[0-9]+}}(%rsp), %rax
-; AVX2-NEXT:    movzwl -24(%rsp,%rax,2), %eax
-; AVX2-NEXT:    vpinsrw $6, %eax, %xmm0, %xmm0
+; AVX2-NEXT:    vpinsrw $6, -24(%rsp,%rax,2), %xmm0, %xmm0
 ; AVX2-NEXT:    movslq {{[0-9]+}}(%rsp), %rax
-; AVX2-NEXT:    movzwl -24(%rsp,%rax,2), %eax
-; AVX2-NEXT:    vpinsrw $7, %eax, %xmm0, %xmm0
+; AVX2-NEXT:    vpinsrw $7, -24(%rsp,%rax,2), %xmm0, %xmm0
 ; AVX2-NEXT:    movslq %edi, %rax
 ; AVX2-NEXT:    movzwl -24(%rsp,%rax,2), %eax
 ; AVX2-NEXT:    vmovd %eax, %xmm1
@@ -560,11 +526,9 @@ define <16 x i16> @var_shuffle_v16i16_v8i16_xxxxxxxxxxxxxxxx_i16(<8 x i16> %x, i
 ; AVX2-NEXT:    movslq %r9d, %rax
 ; AVX2-NEXT:    vpinsrw $5, -24(%rsp,%rax,2), %xmm1, %xmm1
 ; AVX2-NEXT:    movslq {{[0-9]+}}(%rsp), %rax
-; AVX2-NEXT:    movzwl -24(%rsp,%rax,2), %eax
-; AVX2-NEXT:    vpinsrw $6, %eax, %xmm1, %xmm1
+; AVX2-NEXT:    vpinsrw $6, -24(%rsp,%rax,2), %xmm1, %xmm1
 ; AVX2-NEXT:    movslq {{[0-9]+}}(%rsp), %rax
-; AVX2-NEXT:    movzwl -24(%rsp,%rax,2), %eax
-; AVX2-NEXT:    vpinsrw $7, %eax, %xmm1, %xmm1
+; AVX2-NEXT:    vpinsrw $7, -24(%rsp,%rax,2), %xmm1, %xmm1
 ; AVX2-NEXT:    vinserti128 $1, %xmm0, %ymm1, %ymm0
 ; AVX2-NEXT:    retq
   %x0  = extractelement <8 x i16> %x, i32 %i0
@@ -620,11 +584,11 @@ define <4 x i64> @mem_shuffle_v4i64_v4i64_xxxx_i64(<4 x i64> %x, i64* %i) nounwi
 ; AVX1-NEXT:    vmovaps %ymm0, (%rsp)
 ; AVX1-NEXT:    vmovq {{.*#+}} xmm0 = mem[0],zero
 ; AVX1-NEXT:    vmovq {{.*#+}} xmm1 = mem[0],zero
+; AVX1-NEXT:    vpunpcklqdq {{.*#+}} xmm0 = xmm1[0],xmm0[0]
+; AVX1-NEXT:    vmovq {{.*#+}} xmm1 = mem[0],zero
 ; AVX1-NEXT:    vmovq {{.*#+}} xmm2 = mem[0],zero
 ; AVX1-NEXT:    vpunpcklqdq {{.*#+}} xmm1 = xmm2[0],xmm1[0]
-; AVX1-NEXT:    vmovq {{.*#+}} xmm2 = mem[0],zero
-; AVX1-NEXT:    vpunpcklqdq {{.*#+}} xmm0 = xmm0[0],xmm2[0]
-; AVX1-NEXT:    vinsertf128 $1, %xmm1, %ymm0, %ymm0
+; AVX1-NEXT:    vinsertf128 $1, %xmm0, %ymm1, %ymm0
 ; AVX1-NEXT:    movq %rbp, %rsp
 ; AVX1-NEXT:    popq %rbp
 ; AVX1-NEXT:    retq
@@ -642,11 +606,11 @@ define <4 x i64> @mem_shuffle_v4i64_v4i64_xxxx_i64(<4 x i64> %x, i64* %i) nounwi
 ; AVX2-NEXT:    vmovaps %ymm0, (%rsp)
 ; AVX2-NEXT:    vmovq {{.*#+}} xmm0 = mem[0],zero
 ; AVX2-NEXT:    vmovq {{.*#+}} xmm1 = mem[0],zero
+; AVX2-NEXT:    vpunpcklqdq {{.*#+}} xmm0 = xmm1[0],xmm0[0]
+; AVX2-NEXT:    vmovq {{.*#+}} xmm1 = mem[0],zero
 ; AVX2-NEXT:    vmovq {{.*#+}} xmm2 = mem[0],zero
 ; AVX2-NEXT:    vpunpcklqdq {{.*#+}} xmm1 = xmm2[0],xmm1[0]
-; AVX2-NEXT:    vmovq {{.*#+}} xmm2 = mem[0],zero
-; AVX2-NEXT:    vpunpcklqdq {{.*#+}} xmm0 = xmm0[0],xmm2[0]
-; AVX2-NEXT:    vinserti128 $1, %xmm1, %ymm0, %ymm0
+; AVX2-NEXT:    vinserti128 $1, %xmm0, %ymm1, %ymm0
 ; AVX2-NEXT:    movq %rbp, %rsp
 ; AVX2-NEXT:    popq %rbp
 ; AVX2-NEXT:    retq
@@ -679,11 +643,11 @@ define <4 x i64> @mem_shuffle_v4i64_v2i64_xxxx_i64(<2 x i64> %x, i64* %i) nounwi
 ; AVX1-NEXT:    vmovaps %xmm0, -{{[0-9]+}}(%rsp)
 ; AVX1-NEXT:    vmovq {{.*#+}} xmm0 = mem[0],zero
 ; AVX1-NEXT:    vmovq {{.*#+}} xmm1 = mem[0],zero
+; AVX1-NEXT:    vpunpcklqdq {{.*#+}} xmm0 = xmm1[0],xmm0[0]
+; AVX1-NEXT:    vmovq {{.*#+}} xmm1 = mem[0],zero
 ; AVX1-NEXT:    vmovq {{.*#+}} xmm2 = mem[0],zero
 ; AVX1-NEXT:    vpunpcklqdq {{.*#+}} xmm1 = xmm2[0],xmm1[0]
-; AVX1-NEXT:    vmovq {{.*#+}} xmm2 = mem[0],zero
-; AVX1-NEXT:    vpunpcklqdq {{.*#+}} xmm0 = xmm0[0],xmm2[0]
-; AVX1-NEXT:    vinsertf128 $1, %xmm1, %ymm0, %ymm0
+; AVX1-NEXT:    vinsertf128 $1, %xmm0, %ymm1, %ymm0
 ; AVX1-NEXT:    retq
 ;
 ; AVX2-LABEL: mem_shuffle_v4i64_v2i64_xxxx_i64:
@@ -695,11 +659,11 @@ define <4 x i64> @mem_shuffle_v4i64_v2i64_xxxx_i64(<2 x i64> %x, i64* %i) nounwi
 ; AVX2-NEXT:    vmovaps %xmm0, -{{[0-9]+}}(%rsp)
 ; AVX2-NEXT:    vmovq {{.*#+}} xmm0 = mem[0],zero
 ; AVX2-NEXT:    vmovq {{.*#+}} xmm1 = mem[0],zero
+; AVX2-NEXT:    vpunpcklqdq {{.*#+}} xmm0 = xmm1[0],xmm0[0]
+; AVX2-NEXT:    vmovq {{.*#+}} xmm1 = mem[0],zero
 ; AVX2-NEXT:    vmovq {{.*#+}} xmm2 = mem[0],zero
 ; AVX2-NEXT:    vpunpcklqdq {{.*#+}} xmm1 = xmm2[0],xmm1[0]
-; AVX2-NEXT:    vmovq {{.*#+}} xmm2 = mem[0],zero
-; AVX2-NEXT:    vpunpcklqdq {{.*#+}} xmm0 = xmm0[0],xmm2[0]
-; AVX2-NEXT:    vinserti128 $1, %xmm1, %ymm0, %ymm0
+; AVX2-NEXT:    vinserti128 $1, %xmm0, %ymm1, %ymm0
 ; AVX2-NEXT:    retq
   %p0  = getelementptr inbounds i64, i64* %i, i32 0
   %p1  = getelementptr inbounds i64, i64* %i, i32 1
diff --git a/llvm/test/CodeGen/X86/win32-eh.ll b/llvm/test/CodeGen/X86/win32-eh.ll
index 88403c68740..de8464e4f8b 100644
--- a/llvm/test/CodeGen/X86/win32-eh.ll
+++ b/llvm/test/CodeGen/X86/win32-eh.ll
@@ -27,23 +27,26 @@ catch:
 
 ; CHECK-LABEL: _use_except_handler3:
 ; CHECK: pushl %ebp
-; CHECK: movl %esp, %ebp
-; CHECK: pushl %ebx
-; CHECK: pushl %edi
-; CHECK: pushl %esi
-; CHECK: subl ${{[0-9]+}}, %esp
-; CHECK: movl $-1, -16(%ebp)
-; CHECK: movl $L__ehtable$use_except_handler3, -20(%ebp)
-; CHECK: leal -28(%ebp), %[[node:[^ ,]*]]
-; CHECK: movl $__except_handler3, -24(%ebp)
-; CHECK: movl %fs:0, %[[next:[^ ,]*]]
-; CHECK: movl %[[next]], -28(%ebp)
-; CHECK: movl %[[node]], %fs:0
-; CHECK: calll _may_throw_or_crash
+; CHECK-NEXT: movl %esp, %ebp
+; CHECK-NEXT: pushl %ebx
+; CHECK-NEXT: pushl %edi
+; CHECK-NEXT: pushl %esi
+; CHECK-NEXT: subl ${{[0-9]+}}, %esp
+; CHECK-NEXT: movl %esp, -36(%ebp)
+; CHECK-NEXT: movl $-1, -16(%ebp)
+; CHECK-NEXT: movl $L__ehtable$use_except_handler3, -20(%ebp)
+; CHECK-NEXT: leal -28(%ebp), %[[node:[^ ,]*]]
+; CHECK-NEXT: movl $__except_handler3, -24(%ebp)
+; CHECK-NEXT: movl %fs:0, %[[next:[^ ,]*]]
+; CHECK-NEXT: movl %[[next]], -28(%ebp)
+; CHECK-NEXT: movl %[[node]], %fs:0
+; CHECK-NEXT: movl $0, -16(%ebp)
+; CHECK-NEXT: calll _may_throw_or_crash
+
 ; CHECK: movl -28(%ebp), %[[next:[^ ,]*]]
-; CHECK: movl %[[next]], %fs:0
+; CHECK-NEXT: movl %[[next]], %fs:0
 ; CHECK: retl
-; CHECK: LBB1_2: # %catch{{$}}
+; CHECK-NEXT: LBB1_2: # %catch{{$}}
 
 ; CHECK: .section .xdata,"dr"
 ; CHECK-LABEL: L__ehtable$use_except_handler3:
@@ -66,23 +69,37 @@ catch:
 
 ; CHECK-LABEL: _use_except_handler4:
 ; CHECK: pushl %ebp
-; CHECK: movl %esp, %ebp
-; CHECK: subl ${{[0-9]+}}, %esp
-; CHECK: movl %esp, -36(%ebp)
-; CHECK: movl $-2, -16(%ebp)
-; CHECK: movl $L__ehtable$use_except_handler4, %[[lsda:[^ ,]*]]
-; CHECK: xorl ___security_cookie, %[[lsda]]
-; CHECK: movl %[[lsda]], -20(%ebp)
-; CHECK: leal -28(%ebp), %[[node:[^ ,]*]]
-; CHECK: movl $__except_handler4, -24(%ebp)
-; CHECK: movl %fs:0, %[[next:[^ ,]*]]
-; CHECK: movl %[[next]], -28(%ebp)
-; CHECK: movl %[[node]], %fs:0
-; CHECK: calll _may_throw_or_crash
+; CHECK-NEXT: movl %esp, %ebp
+; CHECK-NEXT: pushl %ebx
+; CHECK-NEXT: pushl %edi
+; CHECK-NEXT: pushl %esi
+; CHECK-NEXT: subl ${{[0-9]+}}, %esp
+; CHECK-NEXT: movl %ebp, %eax
+; CHECK-NEXT: movl %esp, -36(%ebp)
+; CHECK-NEXT: movl $-2, -16(%ebp)
+; CHECK-NEXT: movl $L__ehtable$use_except_handler4, %[[lsda:[^ ,]*]]
+; CHECK-NEXT: movl ___security_cookie, %[[seccookie:[^ ,]*]]
+; CHECK-NEXT: xorl %[[seccookie]], %[[lsda]]
+; CHECK-NEXT: movl %[[lsda]], -20(%ebp)
+; CHECK-NEXT: xorl %[[seccookie]], %[[tmp1:[^ ,]*]]
+; CHECK-NEXT: movl %[[tmp1]], -40(%ebp)
+; CHECK-NEXT: leal -28(%ebp), %[[node:[^ ,]*]]
+; CHECK-NEXT: movl $__except_handler4, -24(%ebp)
+; CHECK-NEXT: movl %fs:0, %[[next:[^ ,]*]]
+; CHECK-NEXT: movl %[[next]], -28(%ebp)
+; CHECK-NEXT: movl %[[node]], %fs:0
+; CHECK-NEXT: movl $0, -16(%ebp)
+; CHECK-NEXT: calll _may_throw_or_crash
+
 ; CHECK: movl -28(%ebp), %[[next:[^ ,]*]]
-; CHECK: movl %[[next]], %fs:0
-; CHECK: retl
-; CHECK: LBB2_2: # %catch{{$}}
+; CHECK-NEXT: movl %[[next]], %fs:0
+; CHECK-NEXT: addl $28, %esp
+; CHECK-NEXT: popl %esi
+; CHECK-NEXT: popl %edi
+; CHECK-NEXT: popl %ebx
+; CHECK-NEXT: popl %ebp
+; CHECK-NEXT: retl
+; CHECK-NEXT: LBB2_2: # %catch{{$}}
 
 ; CHECK: .section .xdata,"dr"
 ; CHECK-LABEL: L__ehtable$use_except_handler4:
@@ -109,26 +126,33 @@ catch:
 
 ; CHECK-LABEL: _use_except_handler4_ssp:
 ; CHECK: pushl %ebp
-; CHECK: movl %esp, %ebp
-; CHECK: subl ${{[0-9]+}}, %esp
-; CHECK: movl %ebp, %[[ehguard:[^ ,]*]]
-; CHECK: movl %esp, -36(%ebp)
-; CHECK: movl $-2, -16(%ebp)
-; CHECK: movl $L__ehtable$use_except_handler4_ssp, %[[lsda:[^ ,]*]]
-; CHECK: xorl ___security_cookie, %[[lsda]]
-; CHECK: movl %[[lsda]], -20(%ebp)
-; CHECK: xorl ___security_cookie, %[[ehguard]]
-; CHECK: movl %[[ehguard]], -40(%ebp)
-; CHECK: leal -28(%ebp), %[[node:[^ ,]*]]
-; CHECK: movl $__except_handler4, -24(%ebp)
-; CHECK: movl %fs:0, %[[next:[^ ,]*]]
-; CHECK: movl %[[next]], -28(%ebp)
-; CHECK: movl %[[node]], %fs:0
-; CHECK: calll _may_throw_or_crash
+; CHECK-NEXT: movl %esp, %ebp
+; CHECK-NEXT: pushl %ebx
+; CHECK-NEXT: pushl %edi
+; CHECK-NEXT: pushl %esi
+; CHECK-NEXT: subl ${{[0-9]+}}, %esp
+; CHECK-NEXT: movl %ebp, %[[ehguard:[^ ,]*]]
+; CHECK-NEXT: movl %esp, -36(%ebp)
+; CHECK-NEXT: movl $-2, -16(%ebp)
+; CHECK-NEXT: movl $L__ehtable$use_except_handler4_ssp, %[[lsda:[^ ,]*]]
+; CHECK-NEXT: movl ___security_cookie, %[[seccookie:[^ ,]*]]
+; CHECK-NEXT: xorl %[[seccookie]], %[[lsda]]
+; CHECK-NEXT: movl %[[lsda]], -20(%ebp)
+; CHECK-NEXT: xorl %[[seccookie]], %[[ehguard]]
+; CHECK-NEXT: movl %[[ehguard]], -40(%ebp)
+; CHECK-NEXT: leal -28(%ebp), %[[node:[^ ,]*]]
+; CHECK-NEXT: movl $__except_handler4, -24(%ebp)
+; CHECK-NEXT: movl %fs:0, %[[next:[^ ,]*]]
+; CHECK-NEXT: movl %[[next]], -28(%ebp)
+; CHECK-NEXT: movl %[[node]], %fs:0
+; CHECK-NEXT: movl $0, -16(%ebp)
+; CHECK-NEXT: calll _may_throw_or_crash
 ; CHECK: movl -28(%ebp), %[[next:[^ ,]*]]
-; CHECK: movl %[[next]], %fs:0   
+; CHECK-NEXT: movl %[[next]], %fs:0   
 ; CHECK: retl
-; CHECK: [[catch:[^ ,]*]]: # %catch{{$}}
+; CHECK-NEXT: [[catch:[^ ,]*]]: # %catch{{$}}
+
+
 
 ; CHECK: .section .xdata,"dr"
 ; CHECK-LABEL: L__ehtable$use_except_handler4_ssp:
@@ -155,23 +179,26 @@ catch:
 
 ; CHECK-LABEL: _use_CxxFrameHandler3:
 ; CHECK: pushl %ebp
-; CHECK: movl %esp, %ebp
-; CHECK: subl ${{[0-9]+}}, %esp
-; CHECK: movl %esp, -28(%ebp)
-; CHECK: movl $-1, -16(%ebp)
-; CHECK: leal -24(%ebp), %[[node:[^ ,]*]]
-; CHECK: movl $___ehhandler$use_CxxFrameHandler3, -20(%ebp)
-; CHECK: movl %fs:0, %[[next:[^ ,]*]]
-; CHECK: movl %[[next]], -24(%ebp)
-; CHECK: movl %[[node]], %fs:0
-; CHECK: movl $0, -16(%ebp)
-; CHECK: calll _may_throw_or_crash
+; CHECK-NEXT: movl %esp, %ebp
+; CHECK-NEXT: pushl %ebx
+; CHECK-NEXT: pushl %edi
+; CHECK-NEXT: pushl %esi
+; CHECK-NEXT: subl ${{[0-9]+}}, %esp
+; CHECK-NEXT: movl %esp, -28(%ebp)
+; CHECK-NEXT: movl $-1, -16(%ebp)
+; CHECK-NEXT: leal -24(%ebp), %[[node:[^ ,]*]]
+; CHECK-NEXT: movl $___ehhandler$use_CxxFrameHandler3, -20(%ebp)
+; CHECK-NEXT: movl %fs:0, %[[next:[^ ,]*]]
+; CHECK-NEXT: movl %[[next]], -24(%ebp)
+; CHECK-NEXT: movl %[[node]], %fs:0
+; CHECK-NEXT: movl $0, -16(%ebp)
+; CHECK-NEXT: calll _may_throw_or_crash
 ; CHECK: movl -24(%ebp), %[[next:[^ ,]*]]
-; CHECK: movl %[[next]], %fs:0
+; CHECK-NEXT: movl %[[next]], %fs:0
 ; CHECK: retl
 
 ; CHECK: .section .xdata,"dr"
-; CHECK: .p2align 2
+; CHECK-NEXT: .p2align 2
 ; CHECK-LABEL: L__ehtable$use_CxxFrameHandler3:
 ; CHECK-NEXT:  .long   429065506
 ; CHECK-NEXT:  .long   2
@@ -185,8 +212,8 @@ catch:
 
 ; CHECK-LABEL: ___ehhandler$use_CxxFrameHandler3:
 ; CHECK: movl $L__ehtable$use_CxxFrameHandler3, %eax
-; CHECK: jmp  ___CxxFrameHandler3 # TAILCALL
+; CHECK-NEXT: jmp  ___CxxFrameHandler3 # TAILCALL
 
 ; CHECK: .safeseh __except_handler3
-; CHECK: .safeseh __except_handler4
-; CHECK: .safeseh ___ehhandler$use_CxxFrameHandler3
+; CHECK-NEXT: .safeseh __except_handler4
+; CHECK-NEXT: .safeseh ___ehhandler$use_CxxFrameHandler3