diff --git a/llvm/lib/Target/X86/X86ISelLowering.cpp b/llvm/lib/Target/X86/X86ISelLowering.cpp
--- a/llvm/lib/Target/X86/X86ISelLowering.cpp
+++ b/llvm/lib/Target/X86/X86ISelLowering.cpp
@@ -29864,22 +29864,29 @@
 
     unsigned InBits = InVT.getSizeInBits();
     if (128 % InBits == 0) {
-      // 128 bit and smaller inputs should avoid truncate all together and
-      // just use a build_vector that will become a shuffle.
-      // TODO: Widen and use a shuffle directly?
+      // 128 bit and smaller inputs should avoid truncate altogether and just
+      // use a shuffle.
       MVT InEltVT = InVT.getSimpleVT().getVectorElementType();
-      EVT EltVT = VT.getVectorElementType();
-      unsigned WidenNumElts = WidenVT.getVectorNumElements();
-      SmallVector<SDValue, 16> Ops(WidenNumElts, DAG.getUNDEF(EltVT));
-      // Use the original element count so we don't do more scalar opts than
-      // necessary.
-      unsigned MinElts = VT.getVectorNumElements();
-      for (unsigned i=0; i < MinElts; ++i) {
-        SDValue Val = DAG.getNode(ISD::EXTRACT_VECTOR_ELT, dl, InEltVT, In,
-                                  DAG.getIntPtrConstant(i, dl));
-        Ops[i] = DAG.getNode(ISD::TRUNCATE, dl, EltVT, Val);
-      }
-      Results.push_back(DAG.getBuildVector(WidenVT, dl, Ops));
+      MVT EltVT = WidenVT.getVectorElementType();
+      // Bitcast to get the correct element type for the result.
+      MVT TempVT = MVT::getVectorVT(EltVT, InBits / EltVT.getSizeInBits());
+      SDValue Temp = DAG.getNode(ISD::BITCAST, dl, TempVT, In);
+      // Widen to the desired result type.
+      unsigned NumConcats =
+          WidenVT.getVectorNumElements() / TempVT.getVectorNumElements();
+      if (NumConcats != 1) {
+        SmallVector<SDValue, 8> ConcatOps(NumConcats, DAG.getUNDEF(TempVT));
+        ConcatOps[0] = Temp;
+        Temp = DAG.getNode(ISD::CONCAT_VECTORS, dl, WidenVT, ConcatOps);
+      }
+      // Shuffle to pick out the correct elements for the result.
+      unsigned Scale = InEltVT.getSizeInBits() / EltVT.getSizeInBits();
+      SmallVector<int, 16> Mask(WidenVT.getVectorNumElements(), -1);
+      for (unsigned i = 0; i < InVT.getVectorNumElements(); ++i)
+        Mask[i] = i * Scale;
+      SDValue Undef = DAG.getUNDEF(Temp.getSimpleValueType());
+      SDValue Res = DAG.getVectorShuffle(WidenVT, dl, Temp, Undef, Mask);
+      Results.push_back(Res);
       return;
     }
     // With AVX512 there are some cases that can use a target specific
diff --git a/llvm/test/CodeGen/X86/avx512-vec-cmp.ll b/llvm/test/CodeGen/X86/avx512-vec-cmp.ll
--- a/llvm/test/CodeGen/X86/avx512-vec-cmp.ll
+++ b/llvm/test/CodeGen/X86/avx512-vec-cmp.ll
@@ -1486,9 +1486,11 @@
 ; AVX512BW-NEXT:    kshiftlw $1, %k1, %k1 ## encoding: [0xc4,0xe3,0xf9,0x32,0xc9,0x01]
 ; AVX512BW-NEXT:    korw %k1, %k0, %k0 ## encoding: [0xc5,0xfc,0x45,0xc1]
 ; AVX512BW-NEXT:    vpmovm2w %k0, %zmm0 ## encoding: [0x62,0xf2,0xfe,0x48,0x28,0xc0]
-; AVX512BW-NEXT:    vpacksswb %xmm0, %xmm0, %xmm0 ## encoding: [0xc5,0xf9,0x63,0xc0]
+; AVX512BW-NEXT:    vpshufb {{.*#+}} xmm0 = xmm0[0,2,u,u,u,u,u,u,u,u,u,u,u,u,u,u]
+; AVX512BW-NEXT:    ## encoding: [0xc4,0xe2,0x79,0x00,0x05,A,A,A,A]
+; AVX512BW-NEXT:    ## fixup A - offset: 5, value: LCPI68_0-4, kind: reloc_riprel_4byte
 ; AVX512BW-NEXT:    vpand {{.*}}(%rip), %xmm0, %xmm0 ## encoding: [0xc5,0xf9,0xdb,0x05,A,A,A,A]
-; AVX512BW-NEXT:    ## fixup A - offset: 4, value: LCPI68_0-4, kind: reloc_riprel_4byte
+; AVX512BW-NEXT:    ## fixup A - offset: 4, value: LCPI68_1-4, kind: reloc_riprel_4byte
 ; AVX512BW-NEXT:    vpextrw $0, %xmm0, (%rsi) ## EVEX TO VEX Compression encoding: [0xc4,0xe3,0x79,0x15,0x06,0x00]
 ; AVX512BW-NEXT:    vzeroupper ## encoding: [0xc5,0xf8,0x77]
 ; AVX512BW-NEXT:    retq ## encoding: [0xc3]
@@ -1517,7 +1519,7 @@
 ; SKX-NEXT:    kshiftrb $7, %k1, %k1 ## encoding: [0xc4,0xe3,0x79,0x30,0xc9,0x07]
 ; SKX-NEXT:    korw %k0, %k1, %k0 ## encoding: [0xc5,0xf4,0x45,0xc0]
 ; SKX-NEXT:    vpmovm2w %k0, %xmm0 ## encoding: [0x62,0xf2,0xfe,0x08,0x28,0xc0]
-; SKX-NEXT:    vpacksswb %xmm0, %xmm0, %xmm0 ## EVEX TO VEX Compression encoding: [0xc5,0xf9,0x63,0xc0]
+; SKX-NEXT:    vpmovwb %xmm0, %xmm0 ## encoding: [0x62,0xf2,0x7e,0x08,0x30,0xc0]
 ; SKX-NEXT:    vpand {{.*}}(%rip), %xmm0, %xmm0 ## EVEX TO VEX Compression encoding: [0xc5,0xf9,0xdb,0x05,A,A,A,A]
 ; SKX-NEXT:    ## fixup A - offset: 4, value: LCPI68_0-4, kind: reloc_riprel_4byte
 ; SKX-NEXT:    vpextrw $0, %xmm0, (%rsi) ## EVEX TO VEX Compression encoding: [0xc4,0xe3,0x79,0x15,0x06,0x00]
diff --git a/llvm/test/CodeGen/X86/bitcast-and-setcc-256.ll b/llvm/test/CodeGen/X86/bitcast-and-setcc-256.ll
--- a/llvm/test/CodeGen/X86/bitcast-and-setcc-256.ll
+++ b/llvm/test/CodeGen/X86/bitcast-and-setcc-256.ll
@@ -380,6 +380,7 @@
 ; SSE2-SSSE3-NEXT:    pcmpeqd {{[0-9]+}}(%rsp), %xmm8
 ; SSE2-SSSE3-NEXT:    packssdw %xmm9, %xmm8
 ; SSE2-SSSE3-NEXT:    pand %xmm6, %xmm8
+; SSE2-SSSE3-NEXT:    psllw $15, %xmm8
 ; SSE2-SSSE3-NEXT:    packsswb %xmm8, %xmm8
 ; SSE2-SSSE3-NEXT:    pmovmskb %xmm8, %eax
 ; SSE2-SSSE3-NEXT:    # kill: def $al killed $al killed $eax
@@ -577,6 +578,7 @@
 ; SSE2-SSSE3-NEXT:    cmpltps {{[0-9]+}}(%rsp), %xmm8
 ; SSE2-SSSE3-NEXT:    packssdw %xmm9, %xmm8
 ; SSE2-SSSE3-NEXT:    pand %xmm4, %xmm8
+; SSE2-SSSE3-NEXT:    psllw $15, %xmm8
 ; SSE2-SSSE3-NEXT:    packsswb %xmm8, %xmm8
 ; SSE2-SSSE3-NEXT:    pmovmskb %xmm8, %eax
 ; SSE2-SSSE3-NEXT:    # kill: def $al killed $al killed $eax
diff --git a/llvm/test/CodeGen/X86/bitcast-and-setcc-512.ll b/llvm/test/CodeGen/X86/bitcast-and-setcc-512.ll
--- a/llvm/test/CodeGen/X86/bitcast-and-setcc-512.ll
+++ b/llvm/test/CodeGen/X86/bitcast-and-setcc-512.ll
@@ -14,19 +14,20 @@
 ; SSE-NEXT:    movdqa {{[0-9]+}}(%rsp), %xmm11
 ; SSE-NEXT:    pcmpgtq %xmm7, %xmm3
 ; SSE-NEXT:    pcmpgtq %xmm6, %xmm2
-; SSE-NEXT:    packssdw %xmm3, %xmm2
+; SSE-NEXT:    shufps {{.*#+}} xmm2 = xmm2[0,2],xmm3[0,2]
 ; SSE-NEXT:    pcmpgtq %xmm5, %xmm1
 ; SSE-NEXT:    pcmpgtq %xmm4, %xmm0
-; SSE-NEXT:    packssdw %xmm1, %xmm0
+; SSE-NEXT:    shufps {{.*#+}} xmm0 = xmm0[0,2],xmm1[0,2]
 ; SSE-NEXT:    packssdw %xmm2, %xmm0
 ; SSE-NEXT:    pcmpgtq {{[0-9]+}}(%rsp), %xmm11
 ; SSE-NEXT:    pcmpgtq {{[0-9]+}}(%rsp), %xmm10
-; SSE-NEXT:    packssdw %xmm11, %xmm10
+; SSE-NEXT:    shufps {{.*#+}} xmm10 = xmm10[0,2],xmm11[0,2]
 ; SSE-NEXT:    pcmpgtq {{[0-9]+}}(%rsp), %xmm9
 ; SSE-NEXT:    pcmpgtq {{[0-9]+}}(%rsp), %xmm8
-; SSE-NEXT:    packssdw %xmm9, %xmm8
+; SSE-NEXT:    shufps {{.*#+}} xmm8 = xmm8[0,2],xmm9[0,2]
 ; SSE-NEXT:    packssdw %xmm10, %xmm8
 ; SSE-NEXT:    pand %xmm0, %xmm8
+; SSE-NEXT:    psllw $15, %xmm8
 ; SSE-NEXT:    packsswb %xmm8, %xmm8
 ; SSE-NEXT:    pmovmskb %xmm8, %eax
 ; SSE-NEXT:    # kill: def $al killed $al killed $eax
@@ -110,19 +111,20 @@
 ; SSE-NEXT:    movapd {{[0-9]+}}(%rsp), %xmm11
 ; SSE-NEXT:    cmpltpd %xmm3, %xmm7
 ; SSE-NEXT:    cmpltpd %xmm2, %xmm6
-; SSE-NEXT:    packssdw %xmm7, %xmm6
+; SSE-NEXT:    shufps {{.*#+}} xmm6 = xmm6[0,2],xmm7[0,2]
 ; SSE-NEXT:    cmpltpd %xmm1, %xmm5
 ; SSE-NEXT:    cmpltpd %xmm0, %xmm4
-; SSE-NEXT:    packssdw %xmm5, %xmm4
+; SSE-NEXT:    shufps {{.*#+}} xmm4 = xmm4[0,2],xmm5[0,2]
 ; SSE-NEXT:    packssdw %xmm6, %xmm4
 ; SSE-NEXT:    cmpltpd {{[0-9]+}}(%rsp), %xmm11
 ; SSE-NEXT:    cmpltpd {{[0-9]+}}(%rsp), %xmm10
-; SSE-NEXT:    packssdw %xmm11, %xmm10
+; SSE-NEXT:    shufps {{.*#+}} xmm10 = xmm10[0,2],xmm11[0,2]
 ; SSE-NEXT:    cmpltpd {{[0-9]+}}(%rsp), %xmm9
 ; SSE-NEXT:    cmpltpd {{[0-9]+}}(%rsp), %xmm8
-; SSE-NEXT:    packssdw %xmm9, %xmm8
+; SSE-NEXT:    shufps {{.*#+}} xmm8 = xmm8[0,2],xmm9[0,2]
 ; SSE-NEXT:    packssdw %xmm10, %xmm8
 ; SSE-NEXT:    pand %xmm4, %xmm8
+; SSE-NEXT:    psllw $15, %xmm8
 ; SSE-NEXT:    packsswb %xmm8, %xmm8
 ; SSE-NEXT:    pmovmskb %xmm8, %eax
 ; SSE-NEXT:    # kill: def $al killed $al killed $eax
@@ -320,6 +322,7 @@
 ; SSE-NEXT:    packssdw %xmm9, %xmm8
 ; SSE-NEXT:    packsswb %xmm10, %xmm8
 ; SSE-NEXT:    pand %xmm0, %xmm8
+; SSE-NEXT:    psllw $7, %xmm8
 ; SSE-NEXT:    pmovmskb %xmm8, %eax
 ; SSE-NEXT:    # kill: def $ax killed $ax killed $eax
 ; SSE-NEXT:    retq
@@ -422,6 +425,7 @@
 ; SSE-NEXT:    packssdw %xmm9, %xmm8
 ; SSE-NEXT:    packsswb %xmm10, %xmm8
 ; SSE-NEXT:    pand %xmm4, %xmm8
+; SSE-NEXT:    psllw $7, %xmm8
 ; SSE-NEXT:    pmovmskb %xmm8, %eax
 ; SSE-NEXT:    # kill: def $ax killed $ax killed $eax
 ; SSE-NEXT:    retq
diff --git a/llvm/test/CodeGen/X86/extract-concat.ll b/llvm/test/CodeGen/X86/extract-concat.ll
--- a/llvm/test/CodeGen/X86/extract-concat.ll
+++ b/llvm/test/CodeGen/X86/extract-concat.ll
@@ -9,22 +9,18 @@
 ; SSE2-LABEL: foo:
 ; SSE2:       # %bb.0:
 ; SSE2-NEXT:    cvttps2dq %xmm0, %xmm0
-; SSE2-NEXT:    movaps %xmm0, -{{[0-9]+}}(%rsp)
-; SSE2-NEXT:    movzbl -{{[0-9]+}}(%rsp), %eax
-; SSE2-NEXT:    movl -{{[0-9]+}}(%rsp), %ecx
-; SSE2-NEXT:    shll $8, %ecx
-; SSE2-NEXT:    orl %eax, %ecx
-; SSE2-NEXT:    movd %ecx, %xmm0
-; SSE2-NEXT:    movl $65280, %eax # imm = 0xFF00
-; SSE2-NEXT:    orl -{{[0-9]+}}(%rsp), %eax
-; SSE2-NEXT:    pinsrw $1, %eax, %xmm0
+; SSE2-NEXT:    pand {{.*}}(%rip), %xmm0
+; SSE2-NEXT:    packuswb %xmm0, %xmm0
+; SSE2-NEXT:    packuswb %xmm0, %xmm0
+; SSE2-NEXT:    pand {{.*}}(%rip), %xmm0
+; SSE2-NEXT:    por {{.*}}(%rip), %xmm0
 ; SSE2-NEXT:    movd %xmm0, (%rdi)
 ; SSE2-NEXT:    retq
 ;
 ; SSE42-LABEL: foo:
 ; SSE42:       # %bb.0:
 ; SSE42-NEXT:    cvttps2dq %xmm0, %xmm0
-; SSE42-NEXT:    pshufb {{.*#+}} xmm0 = xmm0[0,4,8],zero,xmm0[u,u,u,u,u,u,u,u,u,u,u,u]
+; SSE42-NEXT:    pshufb {{.*#+}} xmm0 = xmm0[0,4,8,u,u,u,u,u,u,u,u,u,u,u,u,u]
 ; SSE42-NEXT:    movl $255, %eax
 ; SSE42-NEXT:    pinsrb $3, %eax, %xmm0
 ; SSE42-NEXT:    movd %xmm0, (%rdi)
@@ -33,7 +29,7 @@
 ; AVX-LABEL: foo:
 ; AVX:       # %bb.0:
 ; AVX-NEXT:    vcvttps2dq %xmm0, %xmm0
-; AVX-NEXT:    vpshufb {{.*#+}} xmm0 = xmm0[0,4,8],zero,xmm0[u,u,u,u,u,u,u,u,u,u,u,u]
+; AVX-NEXT:    vpshufb {{.*#+}} xmm0 = xmm0[0,4,8,u,u,u,u,u,u,u,u,u,u,u,u,u]
 ; AVX-NEXT:    movl $255, %eax
 ; AVX-NEXT:    vpinsrb $3, %eax, %xmm0, %xmm0
 ; AVX-NEXT:    vmovd %xmm0, (%rdi)
diff --git a/llvm/test/CodeGen/X86/masked_store_trunc.ll b/llvm/test/CodeGen/X86/masked_store_trunc.ll
--- a/llvm/test/CodeGen/X86/masked_store_trunc.ll
+++ b/llvm/test/CodeGen/X86/masked_store_trunc.ll
@@ -215,17 +215,15 @@
 ; SSE2-LABEL: truncstore_v8i64_v8i16:
 ; SSE2:       # %bb.0:
 ; SSE2-NEXT:    pxor %xmm6, %xmm6
-; SSE2-NEXT:    pshufd {{.*#+}} xmm1 = xmm1[0,2,2,3]
-; SSE2-NEXT:    pshuflw {{.*#+}} xmm1 = xmm1[0,2,2,3,4,5,6,7]
+; SSE2-NEXT:    shufps {{.*#+}} xmm2 = xmm2[0,2],xmm3[0,2]
+; SSE2-NEXT:    pshuflw {{.*#+}} xmm2 = xmm2[0,2,2,3,4,5,6,7]
+; SSE2-NEXT:    pshufhw {{.*#+}} xmm2 = xmm2[0,1,2,3,4,6,6,7]
+; SSE2-NEXT:    pshufd {{.*#+}} xmm2 = xmm2[0,2,2,3]
+; SSE2-NEXT:    shufps {{.*#+}} xmm0 = xmm0[0,2],xmm1[0,2]
+; SSE2-NEXT:    pshuflw {{.*#+}} xmm0 = xmm0[0,2,2,3,4,5,6,7]
+; SSE2-NEXT:    pshufhw {{.*#+}} xmm0 = xmm0[0,1,2,3,4,6,6,7]
 ; SSE2-NEXT:    pshufd {{.*#+}} xmm0 = xmm0[0,2,2,3]
-; SSE2-NEXT:    pshuflw {{.*#+}} xmm7 = xmm0[0,2,2,3,4,5,6,7]
-; SSE2-NEXT:    punpckldq {{.*#+}} xmm7 = xmm7[0],xmm1[0],xmm7[1],xmm1[1]
-; SSE2-NEXT:    pshufd {{.*#+}} xmm0 = xmm3[0,2,2,3]
-; SSE2-NEXT:    pshuflw {{.*#+}} xmm1 = xmm0[0,1,0,2,4,5,6,7]
-; SSE2-NEXT:    pshufd {{.*#+}} xmm0 = xmm2[0,2,2,3]
-; SSE2-NEXT:    pshuflw {{.*#+}} xmm0 = xmm0[0,1,0,2,4,5,6,7]
-; SSE2-NEXT:    punpckldq {{.*#+}} xmm0 = xmm0[0],xmm1[0],xmm0[1],xmm1[1]
-; SSE2-NEXT:    movsd {{.*#+}} xmm0 = xmm7[0],xmm0[1]
+; SSE2-NEXT:    punpcklqdq {{.*#+}} xmm0 = xmm0[0],xmm2[0]
 ; SSE2-NEXT:    pcmpeqd %xmm6, %xmm5
 ; SSE2-NEXT:    pcmpeqd %xmm1, %xmm1
 ; SSE2-NEXT:    pxor %xmm1, %xmm5
diff --git a/llvm/test/CodeGen/X86/masked_store_trunc_usat.ll b/llvm/test/CodeGen/X86/masked_store_trunc_usat.ll
--- a/llvm/test/CodeGen/X86/masked_store_trunc_usat.ll
+++ b/llvm/test/CodeGen/X86/masked_store_trunc_usat.ll
@@ -321,47 +321,48 @@
 ; SSE2-NEXT:    pxor %xmm8, %xmm8
 ; SSE2-NEXT:    movdqa {{.*#+}} xmm9 = [65535,65535]
 ; SSE2-NEXT:    movdqa {{.*#+}} xmm11 = [9223372039002259456,9223372039002259456]
-; SSE2-NEXT:    movdqa %xmm2, %xmm6
-; SSE2-NEXT:    pxor %xmm11, %xmm6
+; SSE2-NEXT:    movdqa %xmm1, %xmm7
+; SSE2-NEXT:    pxor %xmm11, %xmm7
 ; SSE2-NEXT:    movdqa {{.*#+}} xmm10 = [9223372039002324991,9223372039002324991]
-; SSE2-NEXT:    movdqa %xmm10, %xmm7
-; SSE2-NEXT:    pcmpgtd %xmm6, %xmm7
-; SSE2-NEXT:    pshufd {{.*#+}} xmm12 = xmm7[0,0,2,2]
-; SSE2-NEXT:    pcmpeqd %xmm10, %xmm6
-; SSE2-NEXT:    pshufd {{.*#+}} xmm6 = xmm6[1,1,3,3]
-; SSE2-NEXT:    pand %xmm12, %xmm6
-; SSE2-NEXT:    pshufd {{.*#+}} xmm13 = xmm7[1,1,3,3]
-; SSE2-NEXT:    por %xmm6, %xmm13
-; SSE2-NEXT:    pand %xmm13, %xmm2
-; SSE2-NEXT:    pandn %xmm9, %xmm13
-; SSE2-NEXT:    por %xmm2, %xmm13
-; SSE2-NEXT:    movdqa %xmm3, %xmm2
-; SSE2-NEXT:    pxor %xmm11, %xmm2
 ; SSE2-NEXT:    movdqa %xmm10, %xmm6
-; SSE2-NEXT:    pcmpgtd %xmm2, %xmm6
+; SSE2-NEXT:    pcmpgtd %xmm7, %xmm6
 ; SSE2-NEXT:    pshufd {{.*#+}} xmm12 = xmm6[0,0,2,2]
-; SSE2-NEXT:    pcmpeqd %xmm10, %xmm2
-; SSE2-NEXT:    pshufd {{.*#+}} xmm7 = xmm2[1,1,3,3]
+; SSE2-NEXT:    pcmpeqd %xmm10, %xmm7
+; SSE2-NEXT:    pshufd {{.*#+}} xmm7 = xmm7[1,1,3,3]
 ; SSE2-NEXT:    pand %xmm12, %xmm7
-; SSE2-NEXT:    pshufd {{.*#+}} xmm2 = xmm6[1,1,3,3]
-; SSE2-NEXT:    por %xmm7, %xmm2
-; SSE2-NEXT:    pand %xmm2, %xmm3
-; SSE2-NEXT:    pandn %xmm9, %xmm2
-; SSE2-NEXT:    por %xmm3, %xmm2
-; SSE2-NEXT:    movdqa %xmm0, %xmm3
-; SSE2-NEXT:    pxor %xmm11, %xmm3
+; SSE2-NEXT:    pshufd {{.*#+}} xmm6 = xmm6[1,1,3,3]
+; SSE2-NEXT:    por %xmm7, %xmm6
+; SSE2-NEXT:    pand %xmm6, %xmm1
+; SSE2-NEXT:    pandn %xmm9, %xmm6
+; SSE2-NEXT:    por %xmm1, %xmm6
+; SSE2-NEXT:    movdqa %xmm0, %xmm1
+; SSE2-NEXT:    pxor %xmm11, %xmm1
+; SSE2-NEXT:    movdqa %xmm10, %xmm7
+; SSE2-NEXT:    pcmpgtd %xmm1, %xmm7
+; SSE2-NEXT:    pshufd {{.*#+}} xmm12 = xmm7[0,0,2,2]
+; SSE2-NEXT:    pcmpeqd %xmm10, %xmm1
+; SSE2-NEXT:    pshufd {{.*#+}} xmm13 = xmm1[1,1,3,3]
+; SSE2-NEXT:    pand %xmm12, %xmm13
+; SSE2-NEXT:    pshufd {{.*#+}} xmm1 = xmm7[1,1,3,3]
+; SSE2-NEXT:    por %xmm13, %xmm1
+; SSE2-NEXT:    pand %xmm1, %xmm0
+; SSE2-NEXT:    pandn %xmm9, %xmm1
+; SSE2-NEXT:    por %xmm0, %xmm1
+; SSE2-NEXT:    shufps {{.*#+}} xmm1 = xmm1[0,2],xmm6[0,2]
+; SSE2-NEXT:    movdqa %xmm3, %xmm0
+; SSE2-NEXT:    pxor %xmm11, %xmm0
 ; SSE2-NEXT:    movdqa %xmm10, %xmm6
-; SSE2-NEXT:    pcmpgtd %xmm3, %xmm6
+; SSE2-NEXT:    pcmpgtd %xmm0, %xmm6
 ; SSE2-NEXT:    pshufd {{.*#+}} xmm7 = xmm6[0,0,2,2]
-; SSE2-NEXT:    pcmpeqd %xmm10, %xmm3
-; SSE2-NEXT:    pshufd {{.*#+}} xmm3 = xmm3[1,1,3,3]
-; SSE2-NEXT:    pand %xmm7, %xmm3
+; SSE2-NEXT:    pcmpeqd %xmm10, %xmm0
+; SSE2-NEXT:    pshufd {{.*#+}} xmm0 = xmm0[1,1,3,3]
+; SSE2-NEXT:    pand %xmm7, %xmm0
 ; SSE2-NEXT:    pshufd {{.*#+}} xmm6 = xmm6[1,1,3,3]
-; SSE2-NEXT:    por %xmm3, %xmm6
-; SSE2-NEXT:    pand %xmm6, %xmm0
-; SSE2-NEXT:    pandn %xmm9, %xmm6
 ; SSE2-NEXT:    por %xmm0, %xmm6
-; SSE2-NEXT:    pxor %xmm1, %xmm11
+; SSE2-NEXT:    pand %xmm6, %xmm3
+; SSE2-NEXT:    pandn %xmm9, %xmm6
+; SSE2-NEXT:    por %xmm3, %xmm6
+; SSE2-NEXT:    pxor %xmm2, %xmm11
 ; SSE2-NEXT:    movdqa %xmm10, %xmm0
 ; SSE2-NEXT:    pcmpgtd %xmm11, %xmm0
 ; SSE2-NEXT:    pshufd {{.*#+}} xmm3 = xmm0[0,0,2,2]
@@ -370,20 +371,17 @@
 ; SSE2-NEXT:    pand %xmm3, %xmm7
 ; SSE2-NEXT:    pshufd {{.*#+}} xmm0 = xmm0[1,1,3,3]
 ; SSE2-NEXT:    por %xmm7, %xmm0
-; SSE2-NEXT:    pand %xmm0, %xmm1
+; SSE2-NEXT:    pand %xmm0, %xmm2
 ; SSE2-NEXT:    pandn %xmm9, %xmm0
-; SSE2-NEXT:    por %xmm1, %xmm0
-; SSE2-NEXT:    pshufd {{.*#+}} xmm0 = xmm0[0,2,2,3]
+; SSE2-NEXT:    por %xmm2, %xmm0
+; SSE2-NEXT:    shufps {{.*#+}} xmm0 = xmm0[0,2],xmm6[0,2]
 ; SSE2-NEXT:    pshuflw {{.*#+}} xmm0 = xmm0[0,2,2,3,4,5,6,7]
-; SSE2-NEXT:    pshufd {{.*#+}} xmm1 = xmm6[0,2,2,3]
-; SSE2-NEXT:    pshuflw {{.*#+}} xmm1 = xmm1[0,2,2,3,4,5,6,7]
-; SSE2-NEXT:    punpckldq {{.*#+}} xmm1 = xmm1[0],xmm0[0],xmm1[1],xmm0[1]
-; SSE2-NEXT:    pshufd {{.*#+}} xmm0 = xmm2[0,2,2,3]
-; SSE2-NEXT:    pshuflw {{.*#+}} xmm2 = xmm0[0,1,0,2,4,5,6,7]
-; SSE2-NEXT:    pshufd {{.*#+}} xmm0 = xmm13[0,2,2,3]
-; SSE2-NEXT:    pshuflw {{.*#+}} xmm0 = xmm0[0,1,0,2,4,5,6,7]
-; SSE2-NEXT:    punpckldq {{.*#+}} xmm0 = xmm0[0],xmm2[0],xmm0[1],xmm2[1]
-; SSE2-NEXT:    movsd {{.*#+}} xmm0 = xmm1[0],xmm0[1]
+; SSE2-NEXT:    pshufhw {{.*#+}} xmm0 = xmm0[0,1,2,3,4,6,6,7]
+; SSE2-NEXT:    pshufd {{.*#+}} xmm2 = xmm0[0,2,2,3]
+; SSE2-NEXT:    pshuflw {{.*#+}} xmm0 = xmm1[0,2,2,3,4,5,6,7]
+; SSE2-NEXT:    pshufhw {{.*#+}} xmm0 = xmm0[0,1,2,3,4,6,6,7]
+; SSE2-NEXT:    pshufd {{.*#+}} xmm0 = xmm0[0,2,2,3]
+; SSE2-NEXT:    punpcklqdq {{.*#+}} xmm0 = xmm0[0],xmm2[0]
 ; SSE2-NEXT:    pcmpeqd %xmm8, %xmm5
 ; SSE2-NEXT:    pcmpeqd %xmm1, %xmm1
 ; SSE2-NEXT:    pxor %xmm1, %xmm5
diff --git a/llvm/test/CodeGen/X86/psubus.ll b/llvm/test/CodeGen/X86/psubus.ll
--- a/llvm/test/CodeGen/X86/psubus.ll
+++ b/llvm/test/CodeGen/X86/psubus.ll
@@ -1485,16 +1485,16 @@
 ; SSE2:       # %bb.0: # %vector.ph
 ; SSE2-NEXT:    pxor %xmm5, %xmm5
 ; SSE2-NEXT:    movdqa %xmm0, %xmm10
-; SSE2-NEXT:    punpckhwd {{.*#+}} xmm10 = xmm10[4],xmm5[4],xmm10[5],xmm5[5],xmm10[6],xmm5[6],xmm10[7],xmm5[7]
+; SSE2-NEXT:    punpcklwd {{.*#+}} xmm10 = xmm10[0],xmm5[0],xmm10[1],xmm5[1],xmm10[2],xmm5[2],xmm10[3],xmm5[3]
 ; SSE2-NEXT:    movdqa %xmm10, %xmm8
-; SSE2-NEXT:    punpckldq {{.*#+}} xmm8 = xmm8[0],xmm5[0],xmm8[1],xmm5[1]
-; SSE2-NEXT:    punpckhdq {{.*#+}} xmm10 = xmm10[2],xmm5[2],xmm10[3],xmm5[3]
-; SSE2-NEXT:    punpcklwd {{.*#+}} xmm0 = xmm0[0],xmm5[0],xmm0[1],xmm5[1],xmm0[2],xmm5[2],xmm0[3],xmm5[3]
+; SSE2-NEXT:    punpckhdq {{.*#+}} xmm8 = xmm8[2],xmm5[2],xmm8[3],xmm5[3]
+; SSE2-NEXT:    punpckldq {{.*#+}} xmm10 = xmm10[0],xmm5[0],xmm10[1],xmm5[1]
+; SSE2-NEXT:    punpckhwd {{.*#+}} xmm0 = xmm0[4],xmm5[4],xmm0[5],xmm5[5],xmm0[6],xmm5[6],xmm0[7],xmm5[7]
 ; SSE2-NEXT:    movdqa %xmm0, %xmm9
-; SSE2-NEXT:    punpckldq {{.*#+}} xmm9 = xmm9[0],xmm5[0],xmm9[1],xmm5[1]
-; SSE2-NEXT:    punpckhdq {{.*#+}} xmm0 = xmm0[2],xmm5[2],xmm0[3],xmm5[3]
+; SSE2-NEXT:    punpckhdq {{.*#+}} xmm9 = xmm9[2],xmm5[2],xmm9[3],xmm5[3]
+; SSE2-NEXT:    punpckldq {{.*#+}} xmm0 = xmm0[0],xmm5[0],xmm0[1],xmm5[1]
 ; SSE2-NEXT:    movdqa {{.*#+}} xmm11 = [9223372039002259456,9223372039002259456]
-; SSE2-NEXT:    movdqa %xmm2, %xmm6
+; SSE2-NEXT:    movdqa %xmm3, %xmm6
 ; SSE2-NEXT:    pxor %xmm11, %xmm6
 ; SSE2-NEXT:    movdqa %xmm0, %xmm7
 ; SSE2-NEXT:    por %xmm11, %xmm7
@@ -1507,9 +1507,9 @@
 ; SSE2-NEXT:    pshufd {{.*#+}} xmm13 = xmm5[1,1,3,3]
 ; SSE2-NEXT:    por %xmm7, %xmm13
 ; SSE2-NEXT:    pand %xmm13, %xmm0
-; SSE2-NEXT:    pandn %xmm2, %xmm13
+; SSE2-NEXT:    pandn %xmm3, %xmm13
 ; SSE2-NEXT:    por %xmm0, %xmm13
-; SSE2-NEXT:    movdqa %xmm1, %xmm0
+; SSE2-NEXT:    movdqa %xmm4, %xmm0
 ; SSE2-NEXT:    pxor %xmm11, %xmm0
 ; SSE2-NEXT:    movdqa %xmm9, %xmm5
 ; SSE2-NEXT:    por %xmm11, %xmm5
@@ -1522,9 +1522,9 @@
 ; SSE2-NEXT:    pshufd {{.*#+}} xmm0 = xmm7[1,1,3,3]
 ; SSE2-NEXT:    por %xmm5, %xmm0
 ; SSE2-NEXT:    pand %xmm0, %xmm9
-; SSE2-NEXT:    pandn %xmm1, %xmm0
+; SSE2-NEXT:    pandn %xmm4, %xmm0
 ; SSE2-NEXT:    por %xmm9, %xmm0
-; SSE2-NEXT:    movdqa %xmm4, %xmm5
+; SSE2-NEXT:    movdqa %xmm1, %xmm5
 ; SSE2-NEXT:    pxor %xmm11, %xmm5
 ; SSE2-NEXT:    movdqa %xmm10, %xmm7
 ; SSE2-NEXT:    por %xmm11, %xmm7
@@ -1537,9 +1537,9 @@
 ; SSE2-NEXT:    pshufd {{.*#+}} xmm7 = xmm6[1,1,3,3]
 ; SSE2-NEXT:    por %xmm5, %xmm7
 ; SSE2-NEXT:    pand %xmm7, %xmm10
-; SSE2-NEXT:    pandn %xmm4, %xmm7
+; SSE2-NEXT:    pandn %xmm1, %xmm7
 ; SSE2-NEXT:    por %xmm10, %xmm7
-; SSE2-NEXT:    movdqa %xmm3, %xmm5
+; SSE2-NEXT:    movdqa %xmm2, %xmm5
 ; SSE2-NEXT:    pxor %xmm11, %xmm5
 ; SSE2-NEXT:    por %xmm8, %xmm11
 ; SSE2-NEXT:    movdqa %xmm11, %xmm6
@@ -1551,30 +1551,28 @@
 ; SSE2-NEXT:    pshufd {{.*#+}} xmm6 = xmm6[1,1,3,3]
 ; SSE2-NEXT:    por %xmm5, %xmm6
 ; SSE2-NEXT:    pand %xmm6, %xmm8
-; SSE2-NEXT:    pandn %xmm3, %xmm6
+; SSE2-NEXT:    pandn %xmm2, %xmm6
 ; SSE2-NEXT:    por %xmm8, %xmm6
-; SSE2-NEXT:    psubq %xmm3, %xmm6
-; SSE2-NEXT:    psubq %xmm4, %xmm7
-; SSE2-NEXT:    psubq %xmm1, %xmm0
-; SSE2-NEXT:    psubq %xmm2, %xmm13
-; SSE2-NEXT:    pshufd {{.*#+}} xmm1 = xmm13[0,2,2,3]
-; SSE2-NEXT:    pshuflw {{.*#+}} xmm1 = xmm1[0,2,2,3,4,5,6,7]
+; SSE2-NEXT:    psubq %xmm2, %xmm6
+; SSE2-NEXT:    psubq %xmm1, %xmm7
+; SSE2-NEXT:    shufps {{.*#+}} xmm7 = xmm7[0,2],xmm6[0,2]
+; SSE2-NEXT:    psubq %xmm4, %xmm0
+; SSE2-NEXT:    psubq %xmm3, %xmm13
+; SSE2-NEXT:    shufps {{.*#+}} xmm13 = xmm13[0,2],xmm0[0,2]
+; SSE2-NEXT:    pshuflw {{.*#+}} xmm0 = xmm13[0,2,2,3,4,5,6,7]
+; SSE2-NEXT:    pshufhw {{.*#+}} xmm0 = xmm0[0,1,2,3,4,6,6,7]
+; SSE2-NEXT:    pshufd {{.*#+}} xmm1 = xmm0[0,2,2,3]
+; SSE2-NEXT:    pshuflw {{.*#+}} xmm0 = xmm7[0,2,2,3,4,5,6,7]
+; SSE2-NEXT:    pshufhw {{.*#+}} xmm0 = xmm0[0,1,2,3,4,6,6,7]
 ; SSE2-NEXT:    pshufd {{.*#+}} xmm0 = xmm0[0,2,2,3]
-; SSE2-NEXT:    pshuflw {{.*#+}} xmm2 = xmm0[0,2,2,3,4,5,6,7]
-; SSE2-NEXT:    punpckldq {{.*#+}} xmm2 = xmm2[0],xmm1[0],xmm2[1],xmm1[1]
-; SSE2-NEXT:    pshufd {{.*#+}} xmm0 = xmm7[0,2,2,3]
-; SSE2-NEXT:    pshuflw {{.*#+}} xmm1 = xmm0[0,1,0,2,4,5,6,7]
-; SSE2-NEXT:    pshufd {{.*#+}} xmm0 = xmm6[0,2,2,3]
-; SSE2-NEXT:    pshuflw {{.*#+}} xmm0 = xmm0[0,1,0,2,4,5,6,7]
-; SSE2-NEXT:    punpckldq {{.*#+}} xmm0 = xmm0[0],xmm1[0],xmm0[1],xmm1[1]
-; SSE2-NEXT:    movsd {{.*#+}} xmm0 = xmm2[0],xmm0[1]
+; SSE2-NEXT:    punpcklqdq {{.*#+}} xmm0 = xmm0[0],xmm1[0]
 ; SSE2-NEXT:    retq
 ;
 ; SSSE3-LABEL: psubus_8i64_max:
 ; SSSE3:       # %bb.0: # %vector.ph
-; SSSE3-NEXT:    movdqa {{.*#+}} xmm5 = [9223372039002259456,9223372039002259456]
-; SSSE3-NEXT:    movdqa %xmm2, %xmm7
-; SSSE3-NEXT:    pxor %xmm5, %xmm7
+; SSSE3-NEXT:    movdqa {{.*#+}} xmm11 = [9223372039002259456,9223372039002259456]
+; SSSE3-NEXT:    movdqa %xmm4, %xmm7
+; SSSE3-NEXT:    pxor %xmm11, %xmm7
 ; SSSE3-NEXT:    movdqa {{.*#+}} xmm8 = [9223372039002324991,9223372039002324991]
 ; SSSE3-NEXT:    movdqa %xmm8, %xmm6
 ; SSSE3-NEXT:    pcmpgtd %xmm7, %xmm6
@@ -1585,59 +1583,54 @@
 ; SSSE3-NEXT:    pshufd {{.*#+}} xmm6 = xmm6[1,1,3,3]
 ; SSSE3-NEXT:    por %xmm7, %xmm6
 ; SSSE3-NEXT:    movdqa {{.*#+}} xmm9 = [65535,65535]
-; SSSE3-NEXT:    pand %xmm6, %xmm2
-; SSSE3-NEXT:    pandn %xmm9, %xmm6
-; SSSE3-NEXT:    por %xmm2, %xmm6
-; SSSE3-NEXT:    pshufd {{.*#+}} xmm2 = xmm6[0,2,2,3]
-; SSSE3-NEXT:    pshuflw {{.*#+}} xmm10 = xmm2[0,2,2,3,4,5,6,7]
-; SSSE3-NEXT:    movdqa %xmm1, %xmm6
-; SSSE3-NEXT:    pxor %xmm5, %xmm6
-; SSSE3-NEXT:    movdqa %xmm8, %xmm7
-; SSSE3-NEXT:    pcmpgtd %xmm6, %xmm7
-; SSSE3-NEXT:    pshufd {{.*#+}} xmm2 = xmm7[0,0,2,2]
-; SSSE3-NEXT:    pcmpeqd %xmm8, %xmm6
-; SSSE3-NEXT:    pshufd {{.*#+}} xmm6 = xmm6[1,1,3,3]
-; SSSE3-NEXT:    pand %xmm2, %xmm6
-; SSSE3-NEXT:    pshufd {{.*#+}} xmm2 = xmm7[1,1,3,3]
-; SSSE3-NEXT:    por %xmm6, %xmm2
-; SSSE3-NEXT:    pand %xmm2, %xmm1
-; SSSE3-NEXT:    pandn %xmm9, %xmm2
-; SSSE3-NEXT:    por %xmm1, %xmm2
-; SSSE3-NEXT:    pshufd {{.*#+}} xmm1 = xmm2[0,2,2,3]
-; SSSE3-NEXT:    pshuflw {{.*#+}} xmm1 = xmm1[0,2,2,3,4,5,6,7]
-; SSSE3-NEXT:    punpckldq {{.*#+}} xmm1 = xmm1[0],xmm10[0],xmm1[1],xmm10[1]
-; SSSE3-NEXT:    movdqa %xmm4, %xmm2
-; SSSE3-NEXT:    pxor %xmm5, %xmm2
-; SSSE3-NEXT:    movdqa %xmm8, %xmm6
-; SSSE3-NEXT:    pcmpgtd %xmm2, %xmm6
-; SSSE3-NEXT:    pshufd {{.*#+}} xmm7 = xmm6[0,0,2,2]
-; SSSE3-NEXT:    pcmpeqd %xmm8, %xmm2
-; SSSE3-NEXT:    pshufd {{.*#+}} xmm2 = xmm2[1,1,3,3]
-; SSSE3-NEXT:    pand %xmm7, %xmm2
-; SSSE3-NEXT:    pshufd {{.*#+}} xmm6 = xmm6[1,1,3,3]
-; SSSE3-NEXT:    por %xmm2, %xmm6
 ; SSSE3-NEXT:    pand %xmm6, %xmm4
 ; SSSE3-NEXT:    pandn %xmm9, %xmm6
 ; SSSE3-NEXT:    por %xmm4, %xmm6
-; SSSE3-NEXT:    pshufd {{.*#+}} xmm2 = xmm6[0,2,2,3]
-; SSSE3-NEXT:    pshuflw {{.*#+}} xmm2 = xmm2[0,1,0,2,4,5,6,7]
-; SSSE3-NEXT:    pxor %xmm3, %xmm5
-; SSSE3-NEXT:    movdqa %xmm8, %xmm4
-; SSSE3-NEXT:    pcmpgtd %xmm5, %xmm4
-; SSSE3-NEXT:    pshufd {{.*#+}} xmm6 = xmm4[0,0,2,2]
-; SSSE3-NEXT:    pcmpeqd %xmm8, %xmm5
-; SSSE3-NEXT:    pshufd {{.*#+}} xmm5 = xmm5[1,1,3,3]
-; SSSE3-NEXT:    pand %xmm6, %xmm5
-; SSSE3-NEXT:    pshufd {{.*#+}} xmm4 = xmm4[1,1,3,3]
+; SSSE3-NEXT:    movdqa %xmm3, %xmm4
+; SSSE3-NEXT:    pxor %xmm11, %xmm4
+; SSSE3-NEXT:    movdqa %xmm8, %xmm7
+; SSSE3-NEXT:    pcmpgtd %xmm4, %xmm7
+; SSSE3-NEXT:    pshufd {{.*#+}} xmm10 = xmm7[0,0,2,2]
+; SSSE3-NEXT:    pcmpeqd %xmm8, %xmm4
+; SSSE3-NEXT:    pshufd {{.*#+}} xmm5 = xmm4[1,1,3,3]
+; SSSE3-NEXT:    pand %xmm10, %xmm5
+; SSSE3-NEXT:    pshufd {{.*#+}} xmm4 = xmm7[1,1,3,3]
 ; SSSE3-NEXT:    por %xmm5, %xmm4
 ; SSSE3-NEXT:    pand %xmm4, %xmm3
 ; SSSE3-NEXT:    pandn %xmm9, %xmm4
 ; SSSE3-NEXT:    por %xmm3, %xmm4
-; SSSE3-NEXT:    pshufd {{.*#+}} xmm3 = xmm4[0,2,2,3]
-; SSSE3-NEXT:    pshuflw {{.*#+}} xmm3 = xmm3[0,1,0,2,4,5,6,7]
-; SSSE3-NEXT:    punpckldq {{.*#+}} xmm3 = xmm3[0],xmm2[0],xmm3[1],xmm2[1]
-; SSSE3-NEXT:    movsd {{.*#+}} xmm3 = xmm1[0],xmm3[1]
-; SSSE3-NEXT:    psubusw %xmm3, %xmm0
+; SSSE3-NEXT:    shufps {{.*#+}} xmm4 = xmm4[0,2],xmm6[0,2]
+; SSSE3-NEXT:    movdqa {{.*#+}} xmm3 = [0,1,4,5,8,9,12,13,8,9,12,13,12,13,14,15]
+; SSSE3-NEXT:    pshufb %xmm3, %xmm4
+; SSSE3-NEXT:    movdqa %xmm2, %xmm5
+; SSSE3-NEXT:    pxor %xmm11, %xmm5
+; SSSE3-NEXT:    movdqa %xmm8, %xmm6
+; SSSE3-NEXT:    pcmpgtd %xmm5, %xmm6
+; SSSE3-NEXT:    pshufd {{.*#+}} xmm7 = xmm6[0,0,2,2]
+; SSSE3-NEXT:    pcmpeqd %xmm8, %xmm5
+; SSSE3-NEXT:    pshufd {{.*#+}} xmm5 = xmm5[1,1,3,3]
+; SSSE3-NEXT:    pand %xmm7, %xmm5
+; SSSE3-NEXT:    pshufd {{.*#+}} xmm6 = xmm6[1,1,3,3]
+; SSSE3-NEXT:    por %xmm5, %xmm6
+; SSSE3-NEXT:    pand %xmm6, %xmm2
+; SSSE3-NEXT:    pandn %xmm9, %xmm6
+; SSSE3-NEXT:    por %xmm2, %xmm6
+; SSSE3-NEXT:    pxor %xmm1, %xmm11
+; SSSE3-NEXT:    movdqa %xmm8, %xmm2
+; SSSE3-NEXT:    pcmpgtd %xmm11, %xmm2
+; SSSE3-NEXT:    pshufd {{.*#+}} xmm5 = xmm2[0,0,2,2]
+; SSSE3-NEXT:    pcmpeqd %xmm8, %xmm11
+; SSSE3-NEXT:    pshufd {{.*#+}} xmm7 = xmm11[1,1,3,3]
+; SSSE3-NEXT:    pand %xmm5, %xmm7
+; SSSE3-NEXT:    pshufd {{.*#+}} xmm2 = xmm2[1,1,3,3]
+; SSSE3-NEXT:    por %xmm7, %xmm2
+; SSSE3-NEXT:    pand %xmm2, %xmm1
+; SSSE3-NEXT:    pandn %xmm9, %xmm2
+; SSSE3-NEXT:    por %xmm1, %xmm2
+; SSSE3-NEXT:    shufps {{.*#+}} xmm2 = xmm2[0,2],xmm6[0,2]
+; SSSE3-NEXT:    pshufb %xmm3, %xmm2
+; SSSE3-NEXT:    punpcklqdq {{.*#+}} xmm2 = xmm2[0],xmm4[0]
+; SSSE3-NEXT:    psubusw %xmm2, %xmm0
 ; SSSE3-NEXT:    retq
 ;
 ; SSE41-LABEL: psubus_8i64_max:
diff --git a/llvm/test/CodeGen/X86/vec_cast.ll b/llvm/test/CodeGen/X86/vec_cast.ll
--- a/llvm/test/CodeGen/X86/vec_cast.ll
+++ b/llvm/test/CodeGen/X86/vec_cast.ll
@@ -131,10 +131,10 @@
 ;
 ; CHECK-WIN-LABEL: g:
 ; CHECK-WIN:       # %bb.0:
-; CHECK-WIN-NEXT:    movdqa (%rcx), %xmm0
 ; CHECK-WIN-NEXT:    movdqa (%rdx), %xmm1
 ; CHECK-WIN-NEXT:    movdqa {{.*#+}} xmm2 = [0,1,4,5,8,9,12,13,8,9,12,13,12,13,14,15]
 ; CHECK-WIN-NEXT:    pshufb %xmm2, %xmm1
+; CHECK-WIN-NEXT:    movdqa (%rcx), %xmm0
 ; CHECK-WIN-NEXT:    pshufb %xmm2, %xmm0
 ; CHECK-WIN-NEXT:    punpcklqdq {{.*#+}} xmm0 = xmm0[0],xmm1[0]
 ; CHECK-WIN-NEXT:    retq
diff --git a/llvm/test/CodeGen/X86/vector-reduce-and-bool.ll b/llvm/test/CodeGen/X86/vector-reduce-and-bool.ll
--- a/llvm/test/CodeGen/X86/vector-reduce-and-bool.ll
+++ b/llvm/test/CodeGen/X86/vector-reduce-and-bool.ll
@@ -499,20 +499,18 @@
 define i1 @trunc_v8i64_v8i1(<8 x i64>) {
 ; SSE2-LABEL: trunc_v8i64_v8i1:
 ; SSE2:       # %bb.0:
-; SSE2-NEXT:    pshufd {{.*#+}} xmm1 = xmm1[0,2,2,3]
-; SSE2-NEXT:    pshuflw {{.*#+}} xmm1 = xmm1[0,2,2,3,4,5,6,7]
-; SSE2-NEXT:    pshufd {{.*#+}} xmm0 = xmm0[0,2,2,3]
-; SSE2-NEXT:    pshuflw {{.*#+}} xmm0 = xmm0[0,2,2,3,4,5,6,7]
-; SSE2-NEXT:    punpckldq {{.*#+}} xmm0 = xmm0[0],xmm1[0],xmm0[1],xmm1[1]
-; SSE2-NEXT:    pshufd {{.*#+}} xmm1 = xmm3[0,2,2,3]
-; SSE2-NEXT:    pshuflw {{.*#+}} xmm1 = xmm1[0,1,0,2,4,5,6,7]
+; SSE2-NEXT:    shufps {{.*#+}} xmm2 = xmm2[0,2],xmm3[0,2]
+; SSE2-NEXT:    pshuflw {{.*#+}} xmm2 = xmm2[0,2,2,3,4,5,6,7]
+; SSE2-NEXT:    pshufhw {{.*#+}} xmm2 = xmm2[0,1,2,3,4,6,6,7]
 ; SSE2-NEXT:    pshufd {{.*#+}} xmm2 = xmm2[0,2,2,3]
-; SSE2-NEXT:    pshuflw {{.*#+}} xmm2 = xmm2[0,1,0,2,4,5,6,7]
-; SSE2-NEXT:    punpckldq {{.*#+}} xmm2 = xmm2[0],xmm1[0],xmm2[1],xmm1[1]
-; SSE2-NEXT:    movsd {{.*#+}} xmm2 = xmm0[0],xmm2[1]
-; SSE2-NEXT:    psllw $15, %xmm2
-; SSE2-NEXT:    packsswb %xmm2, %xmm2
-; SSE2-NEXT:    pmovmskb %xmm2, %eax
+; SSE2-NEXT:    shufps {{.*#+}} xmm0 = xmm0[0,2],xmm1[0,2]
+; SSE2-NEXT:    pshuflw {{.*#+}} xmm0 = xmm0[0,2,2,3,4,5,6,7]
+; SSE2-NEXT:    pshufhw {{.*#+}} xmm0 = xmm0[0,1,2,3,4,6,6,7]
+; SSE2-NEXT:    pshufd {{.*#+}} xmm0 = xmm0[0,2,2,3]
+; SSE2-NEXT:    punpcklqdq {{.*#+}} xmm0 = xmm0[0],xmm2[0]
+; SSE2-NEXT:    psllw $15, %xmm0
+; SSE2-NEXT:    packsswb %xmm0, %xmm0
+; SSE2-NEXT:    pmovmskb %xmm0, %eax
 ; SSE2-NEXT:    cmpb $-1, %al
 ; SSE2-NEXT:    sete %al
 ; SSE2-NEXT:    retq
diff --git a/llvm/test/CodeGen/X86/vector-reduce-mul.ll b/llvm/test/CodeGen/X86/vector-reduce-mul.ll
--- a/llvm/test/CodeGen/X86/vector-reduce-mul.ll
+++ b/llvm/test/CodeGen/X86/vector-reduce-mul.ll
@@ -1620,12 +1620,10 @@
 ; SSE2-NEXT:    pmullw %xmm0, %xmm1
 ; SSE2-NEXT:    pshufd {{.*#+}} xmm0 = xmm1[1,1,1,1]
 ; SSE2-NEXT:    pmullw %xmm1, %xmm0
-; SSE2-NEXT:    movdqa {{.*#+}} xmm1 = [255,255,255,255,255,255,255,255]
-; SSE2-NEXT:    pand %xmm0, %xmm1
-; SSE2-NEXT:    packuswb %xmm1, %xmm1
-; SSE2-NEXT:    psrld $16, %xmm0
-; SSE2-NEXT:    pmullw %xmm1, %xmm0
-; SSE2-NEXT:    movd %xmm0, %eax
+; SSE2-NEXT:    movdqa %xmm0, %xmm1
+; SSE2-NEXT:    psrld $16, %xmm1
+; SSE2-NEXT:    pmullw %xmm0, %xmm1
+; SSE2-NEXT:    movd %xmm1, %eax
 ; SSE2-NEXT:    # kill: def $al killed $al killed $eax
 ; SSE2-NEXT:    retq
 ;
diff --git a/llvm/test/CodeGen/X86/vector-reduce-or-bool.ll b/llvm/test/CodeGen/X86/vector-reduce-or-bool.ll
--- a/llvm/test/CodeGen/X86/vector-reduce-or-bool.ll
+++ b/llvm/test/CodeGen/X86/vector-reduce-or-bool.ll
@@ -486,19 +486,17 @@
 define i1 @trunc_v8i64_v8i1(<8 x i64>) {
 ; SSE2-LABEL: trunc_v8i64_v8i1:
 ; SSE2:       # %bb.0:
-; SSE2-NEXT:    pshufd {{.*#+}} xmm1 = xmm1[0,2,2,3]
-; SSE2-NEXT:    pshuflw {{.*#+}} xmm1 = xmm1[0,2,2,3,4,5,6,7]
-; SSE2-NEXT:    pshufd {{.*#+}} xmm0 = xmm0[0,2,2,3]
-; SSE2-NEXT:    pshuflw {{.*#+}} xmm0 = xmm0[0,2,2,3,4,5,6,7]
-; SSE2-NEXT:    punpckldq {{.*#+}} xmm0 = xmm0[0],xmm1[0],xmm0[1],xmm1[1]
-; SSE2-NEXT:    pshufd {{.*#+}} xmm1 = xmm3[0,2,2,3]
-; SSE2-NEXT:    pshuflw {{.*#+}} xmm1 = xmm1[0,1,0,2,4,5,6,7]
+; SSE2-NEXT:    shufps {{.*#+}} xmm2 = xmm2[0,2],xmm3[0,2]
+; SSE2-NEXT:    pshuflw {{.*#+}} xmm2 = xmm2[0,2,2,3,4,5,6,7]
+; SSE2-NEXT:    pshufhw {{.*#+}} xmm2 = xmm2[0,1,2,3,4,6,6,7]
 ; SSE2-NEXT:    pshufd {{.*#+}} xmm2 = xmm2[0,2,2,3]
-; SSE2-NEXT:    pshuflw {{.*#+}} xmm2 = xmm2[0,1,0,2,4,5,6,7]
-; SSE2-NEXT:    punpckldq {{.*#+}} xmm2 = xmm2[0],xmm1[0],xmm2[1],xmm1[1]
-; SSE2-NEXT:    movsd {{.*#+}} xmm2 = xmm0[0],xmm2[1]
-; SSE2-NEXT:    psllw $15, %xmm2
-; SSE2-NEXT:    pmovmskb %xmm2, %eax
+; SSE2-NEXT:    shufps {{.*#+}} xmm0 = xmm0[0,2],xmm1[0,2]
+; SSE2-NEXT:    pshuflw {{.*#+}} xmm0 = xmm0[0,2,2,3,4,5,6,7]
+; SSE2-NEXT:    pshufhw {{.*#+}} xmm0 = xmm0[0,1,2,3,4,6,6,7]
+; SSE2-NEXT:    pshufd {{.*#+}} xmm0 = xmm0[0,2,2,3]
+; SSE2-NEXT:    punpcklqdq {{.*#+}} xmm0 = xmm0[0],xmm2[0]
+; SSE2-NEXT:    psllw $15, %xmm0
+; SSE2-NEXT:    pmovmskb %xmm0, %eax
 ; SSE2-NEXT:    testl $43690, %eax # imm = 0xAAAA
 ; SSE2-NEXT:    setne %al
 ; SSE2-NEXT:    retq
diff --git a/llvm/test/CodeGen/X86/vector-reduce-xor-bool.ll b/llvm/test/CodeGen/X86/vector-reduce-xor-bool.ll
--- a/llvm/test/CodeGen/X86/vector-reduce-xor-bool.ll
+++ b/llvm/test/CodeGen/X86/vector-reduce-xor-bool.ll
@@ -511,20 +511,18 @@
 define i1 @trunc_v8i64_v8i1(<8 x i64>) {
 ; SSE2-LABEL: trunc_v8i64_v8i1:
 ; SSE2:       # %bb.0:
-; SSE2-NEXT:    pshufd {{.*#+}} xmm1 = xmm1[0,2,2,3]
-; SSE2-NEXT:    pshuflw {{.*#+}} xmm1 = xmm1[0,2,2,3,4,5,6,7]
-; SSE2-NEXT:    pshufd {{.*#+}} xmm0 = xmm0[0,2,2,3]
-; SSE2-NEXT:    pshuflw {{.*#+}} xmm0 = xmm0[0,2,2,3,4,5,6,7]
-; SSE2-NEXT:    punpckldq {{.*#+}} xmm0 = xmm0[0],xmm1[0],xmm0[1],xmm1[1]
-; SSE2-NEXT:    pshufd {{.*#+}} xmm1 = xmm3[0,2,2,3]
-; SSE2-NEXT:    pshuflw {{.*#+}} xmm1 = xmm1[0,1,0,2,4,5,6,7]
+; SSE2-NEXT:    shufps {{.*#+}} xmm2 = xmm2[0,2],xmm3[0,2]
+; SSE2-NEXT:    pshuflw {{.*#+}} xmm2 = xmm2[0,2,2,3,4,5,6,7]
+; SSE2-NEXT:    pshufhw {{.*#+}} xmm2 = xmm2[0,1,2,3,4,6,6,7]
 ; SSE2-NEXT:    pshufd {{.*#+}} xmm2 = xmm2[0,2,2,3]
-; SSE2-NEXT:    pshuflw {{.*#+}} xmm2 = xmm2[0,1,0,2,4,5,6,7]
-; SSE2-NEXT:    punpckldq {{.*#+}} xmm2 = xmm2[0],xmm1[0],xmm2[1],xmm1[1]
-; SSE2-NEXT:    movsd {{.*#+}} xmm2 = xmm0[0],xmm2[1]
-; SSE2-NEXT:    psllw $15, %xmm2
-; SSE2-NEXT:    packsswb %xmm2, %xmm2
-; SSE2-NEXT:    pmovmskb %xmm2, %eax
+; SSE2-NEXT:    shufps {{.*#+}} xmm0 = xmm0[0,2],xmm1[0,2]
+; SSE2-NEXT:    pshuflw {{.*#+}} xmm0 = xmm0[0,2,2,3,4,5,6,7]
+; SSE2-NEXT:    pshufhw {{.*#+}} xmm0 = xmm0[0,1,2,3,4,6,6,7]
+; SSE2-NEXT:    pshufd {{.*#+}} xmm0 = xmm0[0,2,2,3]
+; SSE2-NEXT:    punpcklqdq {{.*#+}} xmm0 = xmm0[0],xmm2[0]
+; SSE2-NEXT:    psllw $15, %xmm0
+; SSE2-NEXT:    packsswb %xmm0, %xmm0
+; SSE2-NEXT:    pmovmskb %xmm0, %eax
 ; SSE2-NEXT:    testb %al, %al
 ; SSE2-NEXT:    setnp %al
 ; SSE2-NEXT:    retq
diff --git a/llvm/test/CodeGen/X86/vector-trunc-math.ll b/llvm/test/CodeGen/X86/vector-trunc-math.ll
--- a/llvm/test/CodeGen/X86/vector-trunc-math.ll
+++ b/llvm/test/CodeGen/X86/vector-trunc-math.ll
@@ -61,21 +61,19 @@
 define <8 x i16> @trunc_add_v8i64_v8i16(<8 x i64> %a0, <8 x i64> %a1) nounwind {
 ; SSE-LABEL: trunc_add_v8i64_v8i16:
 ; SSE:       # %bb.0:
-; SSE-NEXT:    paddq %xmm6, %xmm2
-; SSE-NEXT:    paddq %xmm7, %xmm3
-; SSE-NEXT:    paddq %xmm4, %xmm0
 ; SSE-NEXT:    paddq %xmm5, %xmm1
+; SSE-NEXT:    paddq %xmm4, %xmm0
+; SSE-NEXT:    shufps {{.*#+}} xmm0 = xmm0[0,2],xmm1[0,2]
+; SSE-NEXT:    paddq %xmm7, %xmm3
+; SSE-NEXT:    paddq %xmm6, %xmm2
+; SSE-NEXT:    shufps {{.*#+}} xmm2 = xmm2[0,2],xmm3[0,2]
+; SSE-NEXT:    pshuflw {{.*#+}} xmm1 = xmm2[0,2,2,3,4,5,6,7]
+; SSE-NEXT:    pshufhw {{.*#+}} xmm1 = xmm1[0,1,2,3,4,6,6,7]
 ; SSE-NEXT:    pshufd {{.*#+}} xmm1 = xmm1[0,2,2,3]
-; SSE-NEXT:    pshuflw {{.*#+}} xmm1 = xmm1[0,2,2,3,4,5,6,7]
+; SSE-NEXT:    pshuflw {{.*#+}} xmm0 = xmm0[0,2,2,3,4,5,6,7]
+; SSE-NEXT:    pshufhw {{.*#+}} xmm0 = xmm0[0,1,2,3,4,6,6,7]
 ; SSE-NEXT:    pshufd {{.*#+}} xmm0 = xmm0[0,2,2,3]
-; SSE-NEXT:    pshuflw {{.*#+}} xmm4 = xmm0[0,2,2,3,4,5,6,7]
-; SSE-NEXT:    punpckldq {{.*#+}} xmm4 = xmm4[0],xmm1[0],xmm4[1],xmm1[1]
-; SSE-NEXT:    pshufd {{.*#+}} xmm0 = xmm3[0,2,2,3]
-; SSE-NEXT:    pshuflw {{.*#+}} xmm1 = xmm0[0,1,0,2,4,5,6,7]
-; SSE-NEXT:    pshufd {{.*#+}} xmm0 = xmm2[0,2,2,3]
-; SSE-NEXT:    pshuflw {{.*#+}} xmm0 = xmm0[0,1,0,2,4,5,6,7]
-; SSE-NEXT:    punpckldq {{.*#+}} xmm0 = xmm0[0],xmm1[0],xmm0[1],xmm1[1]
-; SSE-NEXT:    movsd {{.*#+}} xmm0 = xmm4[0],xmm0[1]
+; SSE-NEXT:    punpcklqdq {{.*#+}} xmm0 = xmm0[0],xmm1[0]
 ; SSE-NEXT:    retq
 ;
 ; AVX1-LABEL: trunc_add_v8i64_v8i16:
@@ -537,17 +535,15 @@
 define <8 x i16> @trunc_add_const_v8i64_v8i16(<8 x i64> %a0) nounwind {
 ; SSE-LABEL: trunc_add_const_v8i64_v8i16:
 ; SSE:       # %bb.0:
-; SSE-NEXT:    pshufd {{.*#+}} xmm1 = xmm1[0,2,2,3]
-; SSE-NEXT:    pshuflw {{.*#+}} xmm1 = xmm1[0,2,2,3,4,5,6,7]
+; SSE-NEXT:    shufps {{.*#+}} xmm2 = xmm2[0,2],xmm3[0,2]
+; SSE-NEXT:    pshuflw {{.*#+}} xmm2 = xmm2[0,2,2,3,4,5,6,7]
+; SSE-NEXT:    pshufhw {{.*#+}} xmm2 = xmm2[0,1,2,3,4,6,6,7]
+; SSE-NEXT:    pshufd {{.*#+}} xmm2 = xmm2[0,2,2,3]
+; SSE-NEXT:    shufps {{.*#+}} xmm0 = xmm0[0,2],xmm1[0,2]
+; SSE-NEXT:    pshuflw {{.*#+}} xmm0 = xmm0[0,2,2,3,4,5,6,7]
+; SSE-NEXT:    pshufhw {{.*#+}} xmm0 = xmm0[0,1,2,3,4,6,6,7]
 ; SSE-NEXT:    pshufd {{.*#+}} xmm0 = xmm0[0,2,2,3]
-; SSE-NEXT:    pshuflw {{.*#+}} xmm4 = xmm0[0,2,2,3,4,5,6,7]
-; SSE-NEXT:    punpckldq {{.*#+}} xmm4 = xmm4[0],xmm1[0],xmm4[1],xmm1[1]
-; SSE-NEXT:    pshufd {{.*#+}} xmm0 = xmm3[0,2,2,3]
-; SSE-NEXT:    pshuflw {{.*#+}} xmm1 = xmm0[0,1,0,2,4,5,6,7]
-; SSE-NEXT:    pshufd {{.*#+}} xmm0 = xmm2[0,2,2,3]
-; SSE-NEXT:    pshuflw {{.*#+}} xmm0 = xmm0[0,1,0,2,4,5,6,7]
-; SSE-NEXT:    punpckldq {{.*#+}} xmm0 = xmm0[0],xmm1[0],xmm0[1],xmm1[1]
-; SSE-NEXT:    movsd {{.*#+}} xmm0 = xmm4[0],xmm0[1]
+; SSE-NEXT:    punpcklqdq {{.*#+}} xmm0 = xmm0[0],xmm2[0]
 ; SSE-NEXT:    paddw {{.*}}(%rip), %xmm0
 ; SSE-NEXT:    retq
 ;
@@ -905,21 +901,19 @@
 define <8 x i16> @trunc_sub_v8i64_v8i16(<8 x i64> %a0, <8 x i64> %a1) nounwind {
 ; SSE-LABEL: trunc_sub_v8i64_v8i16:
 ; SSE:       # %bb.0:
-; SSE-NEXT:    psubq %xmm6, %xmm2
-; SSE-NEXT:    psubq %xmm7, %xmm3
-; SSE-NEXT:    psubq %xmm4, %xmm0
 ; SSE-NEXT:    psubq %xmm5, %xmm1
+; SSE-NEXT:    psubq %xmm4, %xmm0
+; SSE-NEXT:    shufps {{.*#+}} xmm0 = xmm0[0,2],xmm1[0,2]
+; SSE-NEXT:    psubq %xmm7, %xmm3
+; SSE-NEXT:    psubq %xmm6, %xmm2
+; SSE-NEXT:    shufps {{.*#+}} xmm2 = xmm2[0,2],xmm3[0,2]
+; SSE-NEXT:    pshuflw {{.*#+}} xmm1 = xmm2[0,2,2,3,4,5,6,7]
+; SSE-NEXT:    pshufhw {{.*#+}} xmm1 = xmm1[0,1,2,3,4,6,6,7]
 ; SSE-NEXT:    pshufd {{.*#+}} xmm1 = xmm1[0,2,2,3]
-; SSE-NEXT:    pshuflw {{.*#+}} xmm1 = xmm1[0,2,2,3,4,5,6,7]
+; SSE-NEXT:    pshuflw {{.*#+}} xmm0 = xmm0[0,2,2,3,4,5,6,7]
+; SSE-NEXT:    pshufhw {{.*#+}} xmm0 = xmm0[0,1,2,3,4,6,6,7]
 ; SSE-NEXT:    pshufd {{.*#+}} xmm0 = xmm0[0,2,2,3]
-; SSE-NEXT:    pshuflw {{.*#+}} xmm4 = xmm0[0,2,2,3,4,5,6,7]
-; SSE-NEXT:    punpckldq {{.*#+}} xmm4 = xmm4[0],xmm1[0],xmm4[1],xmm1[1]
-; SSE-NEXT:    pshufd {{.*#+}} xmm0 = xmm3[0,2,2,3]
-; SSE-NEXT:    pshuflw {{.*#+}} xmm1 = xmm0[0,1,0,2,4,5,6,7]
-; SSE-NEXT:    pshufd {{.*#+}} xmm0 = xmm2[0,2,2,3]
-; SSE-NEXT:    pshuflw {{.*#+}} xmm0 = xmm0[0,1,0,2,4,5,6,7]
-; SSE-NEXT:    punpckldq {{.*#+}} xmm0 = xmm0[0],xmm1[0],xmm0[1],xmm1[1]
-; SSE-NEXT:    movsd {{.*#+}} xmm0 = xmm4[0],xmm0[1]
+; SSE-NEXT:    punpcklqdq {{.*#+}} xmm0 = xmm0[0],xmm1[0]
 ; SSE-NEXT:    retq
 ;
 ; AVX1-LABEL: trunc_sub_v8i64_v8i16:
@@ -1349,17 +1343,15 @@
 define <8 x i16> @trunc_sub_const_v8i64_v8i16(<8 x i64> %a0) nounwind {
 ; SSE-LABEL: trunc_sub_const_v8i64_v8i16:
 ; SSE:       # %bb.0:
-; SSE-NEXT:    pshufd {{.*#+}} xmm1 = xmm1[0,2,2,3]
-; SSE-NEXT:    pshuflw {{.*#+}} xmm1 = xmm1[0,2,2,3,4,5,6,7]
+; SSE-NEXT:    shufps {{.*#+}} xmm2 = xmm2[0,2],xmm3[0,2]
+; SSE-NEXT:    pshuflw {{.*#+}} xmm2 = xmm2[0,2,2,3,4,5,6,7]
+; SSE-NEXT:    pshufhw {{.*#+}} xmm2 = xmm2[0,1,2,3,4,6,6,7]
+; SSE-NEXT:    pshufd {{.*#+}} xmm2 = xmm2[0,2,2,3]
+; SSE-NEXT:    shufps {{.*#+}} xmm0 = xmm0[0,2],xmm1[0,2]
+; SSE-NEXT:    pshuflw {{.*#+}} xmm0 = xmm0[0,2,2,3,4,5,6,7]
+; SSE-NEXT:    pshufhw {{.*#+}} xmm0 = xmm0[0,1,2,3,4,6,6,7]
 ; SSE-NEXT:    pshufd {{.*#+}} xmm0 = xmm0[0,2,2,3]
-; SSE-NEXT:    pshuflw {{.*#+}} xmm4 = xmm0[0,2,2,3,4,5,6,7]
-; SSE-NEXT:    punpckldq {{.*#+}} xmm4 = xmm4[0],xmm1[0],xmm4[1],xmm1[1]
-; SSE-NEXT:    pshufd {{.*#+}} xmm0 = xmm3[0,2,2,3]
-; SSE-NEXT:    pshuflw {{.*#+}} xmm1 = xmm0[0,1,0,2,4,5,6,7]
-; SSE-NEXT:    pshufd {{.*#+}} xmm0 = xmm2[0,2,2,3]
-; SSE-NEXT:    pshuflw {{.*#+}} xmm0 = xmm0[0,1,0,2,4,5,6,7]
-; SSE-NEXT:    punpckldq {{.*#+}} xmm0 = xmm0[0],xmm1[0],xmm0[1],xmm1[1]
-; SSE-NEXT:    movsd {{.*#+}} xmm0 = xmm4[0],xmm0[1]
+; SSE-NEXT:    punpcklqdq {{.*#+}} xmm0 = xmm0[0],xmm2[0]
 ; SSE-NEXT:    psubw {{.*}}(%rip), %xmm0
 ; SSE-NEXT:    retq
 ;
@@ -1776,29 +1768,25 @@
 define <8 x i16> @trunc_mul_v8i64_v8i16(<8 x i64> %a0, <8 x i64> %a1) nounwind {
 ; SSE-LABEL: trunc_mul_v8i64_v8i16:
 ; SSE:       # %bb.0:
-; SSE-NEXT:    pshufd {{.*#+}} xmm5 = xmm5[0,2,2,3]
-; SSE-NEXT:    pshuflw {{.*#+}} xmm5 = xmm5[0,2,2,3,4,5,6,7]
-; SSE-NEXT:    pshufd {{.*#+}} xmm4 = xmm4[0,2,2,3]
-; SSE-NEXT:    pshuflw {{.*#+}} xmm4 = xmm4[0,2,2,3,4,5,6,7]
-; SSE-NEXT:    punpckldq {{.*#+}} xmm4 = xmm4[0],xmm5[0],xmm4[1],xmm5[1]
-; SSE-NEXT:    pshufd {{.*#+}} xmm5 = xmm7[0,2,2,3]
-; SSE-NEXT:    pshuflw {{.*#+}} xmm5 = xmm5[0,1,0,2,4,5,6,7]
+; SSE-NEXT:    shufps {{.*#+}} xmm6 = xmm6[0,2],xmm7[0,2]
+; SSE-NEXT:    pshuflw {{.*#+}} xmm6 = xmm6[0,2,2,3,4,5,6,7]
+; SSE-NEXT:    pshufhw {{.*#+}} xmm6 = xmm6[0,1,2,3,4,6,6,7]
 ; SSE-NEXT:    pshufd {{.*#+}} xmm6 = xmm6[0,2,2,3]
-; SSE-NEXT:    pshuflw {{.*#+}} xmm6 = xmm6[0,1,0,2,4,5,6,7]
-; SSE-NEXT:    punpckldq {{.*#+}} xmm6 = xmm6[0],xmm5[0],xmm6[1],xmm5[1]
-; SSE-NEXT:    movsd {{.*#+}} xmm6 = xmm4[0],xmm6[1]
-; SSE-NEXT:    pshufd {{.*#+}} xmm1 = xmm1[0,2,2,3]
-; SSE-NEXT:    pshuflw {{.*#+}} xmm1 = xmm1[0,2,2,3,4,5,6,7]
+; SSE-NEXT:    shufps {{.*#+}} xmm4 = xmm4[0,2],xmm5[0,2]
+; SSE-NEXT:    pshuflw {{.*#+}} xmm4 = xmm4[0,2,2,3,4,5,6,7]
+; SSE-NEXT:    pshufhw {{.*#+}} xmm4 = xmm4[0,1,2,3,4,6,6,7]
+; SSE-NEXT:    pshufd {{.*#+}} xmm4 = xmm4[0,2,2,3]
+; SSE-NEXT:    punpcklqdq {{.*#+}} xmm4 = xmm4[0],xmm6[0]
+; SSE-NEXT:    shufps {{.*#+}} xmm2 = xmm2[0,2],xmm3[0,2]
+; SSE-NEXT:    pshuflw {{.*#+}} xmm2 = xmm2[0,2,2,3,4,5,6,7]
+; SSE-NEXT:    pshufhw {{.*#+}} xmm2 = xmm2[0,1,2,3,4,6,6,7]
+; SSE-NEXT:    pshufd {{.*#+}} xmm2 = xmm2[0,2,2,3]
+; SSE-NEXT:    shufps {{.*#+}} xmm0 = xmm0[0,2],xmm1[0,2]
+; SSE-NEXT:    pshuflw {{.*#+}} xmm0 = xmm0[0,2,2,3,4,5,6,7]
+; SSE-NEXT:    pshufhw {{.*#+}} xmm0 = xmm0[0,1,2,3,4,6,6,7]
 ; SSE-NEXT:    pshufd {{.*#+}} xmm0 = xmm0[0,2,2,3]
-; SSE-NEXT:    pshuflw {{.*#+}} xmm4 = xmm0[0,2,2,3,4,5,6,7]
-; SSE-NEXT:    punpckldq {{.*#+}} xmm4 = xmm4[0],xmm1[0],xmm4[1],xmm1[1]
-; SSE-NEXT:    pshufd {{.*#+}} xmm0 = xmm3[0,2,2,3]
-; SSE-NEXT:    pshuflw {{.*#+}} xmm1 = xmm0[0,1,0,2,4,5,6,7]
-; SSE-NEXT:    pshufd {{.*#+}} xmm0 = xmm2[0,2,2,3]
-; SSE-NEXT:    pshuflw {{.*#+}} xmm0 = xmm0[0,1,0,2,4,5,6,7]
-; SSE-NEXT:    punpckldq {{.*#+}} xmm0 = xmm0[0],xmm1[0],xmm0[1],xmm1[1]
-; SSE-NEXT:    movsd {{.*#+}} xmm0 = xmm4[0],xmm0[1]
-; SSE-NEXT:    pmullw %xmm6, %xmm0
+; SSE-NEXT:    punpcklqdq {{.*#+}} xmm0 = xmm0[0],xmm2[0]
+; SSE-NEXT:    pmullw %xmm4, %xmm0
 ; SSE-NEXT:    retq
 ;
 ; AVX1-LABEL: trunc_mul_v8i64_v8i16:
@@ -2357,17 +2345,15 @@
 define <8 x i16> @trunc_mul_const_v8i64_v8i16(<8 x i64> %a0) nounwind {
 ; SSE-LABEL: trunc_mul_const_v8i64_v8i16:
 ; SSE:       # %bb.0:
-; SSE-NEXT:    pshufd {{.*#+}} xmm1 = xmm1[0,2,2,3]
-; SSE-NEXT:    pshuflw {{.*#+}} xmm1 = xmm1[0,2,2,3,4,5,6,7]
+; SSE-NEXT:    shufps {{.*#+}} xmm2 = xmm2[0,2],xmm3[0,2]
+; SSE-NEXT:    pshuflw {{.*#+}} xmm2 = xmm2[0,2,2,3,4,5,6,7]
+; SSE-NEXT:    pshufhw {{.*#+}} xmm2 = xmm2[0,1,2,3,4,6,6,7]
+; SSE-NEXT:    pshufd {{.*#+}} xmm2 = xmm2[0,2,2,3]
+; SSE-NEXT:    shufps {{.*#+}} xmm0 = xmm0[0,2],xmm1[0,2]
+; SSE-NEXT:    pshuflw {{.*#+}} xmm0 = xmm0[0,2,2,3,4,5,6,7]
+; SSE-NEXT:    pshufhw {{.*#+}} xmm0 = xmm0[0,1,2,3,4,6,6,7]
 ; SSE-NEXT:    pshufd {{.*#+}} xmm0 = xmm0[0,2,2,3]
-; SSE-NEXT:    pshuflw {{.*#+}} xmm4 = xmm0[0,2,2,3,4,5,6,7]
-; SSE-NEXT:    punpckldq {{.*#+}} xmm4 = xmm4[0],xmm1[0],xmm4[1],xmm1[1]
-; SSE-NEXT:    pshufd {{.*#+}} xmm0 = xmm3[0,2,2,3]
-; SSE-NEXT:    pshuflw {{.*#+}} xmm1 = xmm0[0,1,0,2,4,5,6,7]
-; SSE-NEXT:    pshufd {{.*#+}} xmm0 = xmm2[0,2,2,3]
-; SSE-NEXT:    pshuflw {{.*#+}} xmm0 = xmm0[0,1,0,2,4,5,6,7]
-; SSE-NEXT:    punpckldq {{.*#+}} xmm0 = xmm0[0],xmm1[0],xmm0[1],xmm1[1]
-; SSE-NEXT:    movsd {{.*#+}} xmm0 = xmm4[0],xmm0[1]
+; SSE-NEXT:    punpcklqdq {{.*#+}} xmm0 = xmm0[0],xmm2[0]
 ; SSE-NEXT:    pmullw {{.*}}(%rip), %xmm0
 ; SSE-NEXT:    retq
 ;
@@ -2813,21 +2799,19 @@
 define <8 x i16> @trunc_and_v8i64_v8i16(<8 x i64> %a0, <8 x i64> %a1) nounwind {
 ; SSE-LABEL: trunc_and_v8i64_v8i16:
 ; SSE:       # %bb.0:
-; SSE-NEXT:    pand %xmm6, %xmm2
-; SSE-NEXT:    pand %xmm7, %xmm3
-; SSE-NEXT:    pand %xmm4, %xmm0
-; SSE-NEXT:    pand %xmm5, %xmm1
+; SSE-NEXT:    andps %xmm5, %xmm1
+; SSE-NEXT:    andps %xmm4, %xmm0
+; SSE-NEXT:    shufps {{.*#+}} xmm0 = xmm0[0,2],xmm1[0,2]
+; SSE-NEXT:    andps %xmm7, %xmm3
+; SSE-NEXT:    andps %xmm6, %xmm2
+; SSE-NEXT:    shufps {{.*#+}} xmm2 = xmm2[0,2],xmm3[0,2]
+; SSE-NEXT:    pshuflw {{.*#+}} xmm1 = xmm2[0,2,2,3,4,5,6,7]
+; SSE-NEXT:    pshufhw {{.*#+}} xmm1 = xmm1[0,1,2,3,4,6,6,7]
 ; SSE-NEXT:    pshufd {{.*#+}} xmm1 = xmm1[0,2,2,3]
-; SSE-NEXT:    pshuflw {{.*#+}} xmm1 = xmm1[0,2,2,3,4,5,6,7]
+; SSE-NEXT:    pshuflw {{.*#+}} xmm0 = xmm0[0,2,2,3,4,5,6,7]
+; SSE-NEXT:    pshufhw {{.*#+}} xmm0 = xmm0[0,1,2,3,4,6,6,7]
 ; SSE-NEXT:    pshufd {{.*#+}} xmm0 = xmm0[0,2,2,3]
-; SSE-NEXT:    pshuflw {{.*#+}} xmm4 = xmm0[0,2,2,3,4,5,6,7]
-; SSE-NEXT:    punpckldq {{.*#+}} xmm4 = xmm4[0],xmm1[0],xmm4[1],xmm1[1]
-; SSE-NEXT:    pshufd {{.*#+}} xmm0 = xmm3[0,2,2,3]
-; SSE-NEXT:    pshuflw {{.*#+}} xmm1 = xmm0[0,1,0,2,4,5,6,7]
-; SSE-NEXT:    pshufd {{.*#+}} xmm0 = xmm2[0,2,2,3]
-; SSE-NEXT:    pshuflw {{.*#+}} xmm0 = xmm0[0,1,0,2,4,5,6,7]
-; SSE-NEXT:    punpckldq {{.*#+}} xmm0 = xmm0[0],xmm1[0],xmm0[1],xmm1[1]
-; SSE-NEXT:    movsd {{.*#+}} xmm0 = xmm4[0],xmm0[1]
+; SSE-NEXT:    punpcklqdq {{.*#+}} xmm0 = xmm0[0],xmm1[0]
 ; SSE-NEXT:    retq
 ;
 ; AVX1-LABEL: trunc_and_v8i64_v8i16:
@@ -3210,18 +3194,16 @@
 define <8 x i16> @trunc_and_const_v8i64_v8i16(<8 x i64> %a0) nounwind {
 ; SSE-LABEL: trunc_and_const_v8i64_v8i16:
 ; SSE:       # %bb.0:
-; SSE-NEXT:    pshufd {{.*#+}} xmm1 = xmm1[0,2,2,3]
-; SSE-NEXT:    pshuflw {{.*#+}} xmm1 = xmm1[0,2,2,3,4,5,6,7]
+; SSE-NEXT:    shufps {{.*#+}} xmm2 = xmm2[0,2],xmm3[0,2]
+; SSE-NEXT:    pshuflw {{.*#+}} xmm2 = xmm2[0,2,2,3,4,5,6,7]
+; SSE-NEXT:    pshufhw {{.*#+}} xmm2 = xmm2[0,1,2,3,4,6,6,7]
+; SSE-NEXT:    pshufd {{.*#+}} xmm2 = xmm2[0,2,2,3]
+; SSE-NEXT:    shufps {{.*#+}} xmm0 = xmm0[0,2],xmm1[0,2]
+; SSE-NEXT:    pshuflw {{.*#+}} xmm0 = xmm0[0,2,2,3,4,5,6,7]
+; SSE-NEXT:    pshufhw {{.*#+}} xmm0 = xmm0[0,1,2,3,4,6,6,7]
 ; SSE-NEXT:    pshufd {{.*#+}} xmm0 = xmm0[0,2,2,3]
-; SSE-NEXT:    pshuflw {{.*#+}} xmm4 = xmm0[0,2,2,3,4,5,6,7]
-; SSE-NEXT:    punpckldq {{.*#+}} xmm4 = xmm4[0],xmm1[0],xmm4[1],xmm1[1]
-; SSE-NEXT:    pshufd {{.*#+}} xmm0 = xmm3[0,2,2,3]
-; SSE-NEXT:    pshuflw {{.*#+}} xmm1 = xmm0[0,1,0,2,4,5,6,7]
-; SSE-NEXT:    pshufd {{.*#+}} xmm0 = xmm2[0,2,2,3]
-; SSE-NEXT:    pshuflw {{.*#+}} xmm0 = xmm0[0,1,0,2,4,5,6,7]
-; SSE-NEXT:    punpckldq {{.*#+}} xmm0 = xmm0[0],xmm1[0],xmm0[1],xmm1[1]
-; SSE-NEXT:    movsd {{.*#+}} xmm0 = xmm4[0],xmm0[1]
-; SSE-NEXT:    andpd {{.*}}(%rip), %xmm0
+; SSE-NEXT:    punpcklqdq {{.*#+}} xmm0 = xmm0[0],xmm2[0]
+; SSE-NEXT:    pand {{.*}}(%rip), %xmm0
 ; SSE-NEXT:    retq
 ;
 ; AVX1-LABEL: trunc_and_const_v8i64_v8i16:
@@ -3576,21 +3558,19 @@
 define <8 x i16> @trunc_xor_v8i64_v8i16(<8 x i64> %a0, <8 x i64> %a1) nounwind {
 ; SSE-LABEL: trunc_xor_v8i64_v8i16:
 ; SSE:       # %bb.0:
-; SSE-NEXT:    pxor %xmm6, %xmm2
-; SSE-NEXT:    pxor %xmm7, %xmm3
-; SSE-NEXT:    pxor %xmm4, %xmm0
-; SSE-NEXT:    pxor %xmm5, %xmm1
+; SSE-NEXT:    xorps %xmm5, %xmm1
+; SSE-NEXT:    xorps %xmm4, %xmm0
+; SSE-NEXT:    shufps {{.*#+}} xmm0 = xmm0[0,2],xmm1[0,2]
+; SSE-NEXT:    xorps %xmm7, %xmm3
+; SSE-NEXT:    xorps %xmm6, %xmm2
+; SSE-NEXT:    shufps {{.*#+}} xmm2 = xmm2[0,2],xmm3[0,2]
+; SSE-NEXT:    pshuflw {{.*#+}} xmm1 = xmm2[0,2,2,3,4,5,6,7]
+; SSE-NEXT:    pshufhw {{.*#+}} xmm1 = xmm1[0,1,2,3,4,6,6,7]
 ; SSE-NEXT:    pshufd {{.*#+}} xmm1 = xmm1[0,2,2,3]
-; SSE-NEXT:    pshuflw {{.*#+}} xmm1 = xmm1[0,2,2,3,4,5,6,7]
+; SSE-NEXT:    pshuflw {{.*#+}} xmm0 = xmm0[0,2,2,3,4,5,6,7]
+; SSE-NEXT:    pshufhw {{.*#+}} xmm0 = xmm0[0,1,2,3,4,6,6,7]
 ; SSE-NEXT:    pshufd {{.*#+}} xmm0 = xmm0[0,2,2,3]
-; SSE-NEXT:    pshuflw {{.*#+}} xmm4 = xmm0[0,2,2,3,4,5,6,7]
-; SSE-NEXT:    punpckldq {{.*#+}} xmm4 = xmm4[0],xmm1[0],xmm4[1],xmm1[1]
-; SSE-NEXT:    pshufd {{.*#+}} xmm0 = xmm3[0,2,2,3]
-; SSE-NEXT:    pshuflw {{.*#+}} xmm1 = xmm0[0,1,0,2,4,5,6,7]
-; SSE-NEXT:    pshufd {{.*#+}} xmm0 = xmm2[0,2,2,3]
-; SSE-NEXT:    pshuflw {{.*#+}} xmm0 = xmm0[0,1,0,2,4,5,6,7]
-; SSE-NEXT:    punpckldq {{.*#+}} xmm0 = xmm0[0],xmm1[0],xmm0[1],xmm1[1]
-; SSE-NEXT:    movsd {{.*#+}} xmm0 = xmm4[0],xmm0[1]
+; SSE-NEXT:    punpcklqdq {{.*#+}} xmm0 = xmm0[0],xmm1[0]
 ; SSE-NEXT:    retq
 ;
 ; AVX1-LABEL: trunc_xor_v8i64_v8i16:
@@ -3973,18 +3953,16 @@
 define <8 x i16> @trunc_xor_const_v8i64_v8i16(<8 x i64> %a0) nounwind {
 ; SSE-LABEL: trunc_xor_const_v8i64_v8i16:
 ; SSE:       # %bb.0:
-; SSE-NEXT:    pshufd {{.*#+}} xmm1 = xmm1[0,2,2,3]
-; SSE-NEXT:    pshuflw {{.*#+}} xmm1 = xmm1[0,2,2,3,4,5,6,7]
+; SSE-NEXT:    shufps {{.*#+}} xmm2 = xmm2[0,2],xmm3[0,2]
+; SSE-NEXT:    pshuflw {{.*#+}} xmm2 = xmm2[0,2,2,3,4,5,6,7]
+; SSE-NEXT:    pshufhw {{.*#+}} xmm2 = xmm2[0,1,2,3,4,6,6,7]
+; SSE-NEXT:    pshufd {{.*#+}} xmm2 = xmm2[0,2,2,3]
+; SSE-NEXT:    shufps {{.*#+}} xmm0 = xmm0[0,2],xmm1[0,2]
+; SSE-NEXT:    pshuflw {{.*#+}} xmm0 = xmm0[0,2,2,3,4,5,6,7]
+; SSE-NEXT:    pshufhw {{.*#+}} xmm0 = xmm0[0,1,2,3,4,6,6,7]
 ; SSE-NEXT:    pshufd {{.*#+}} xmm0 = xmm0[0,2,2,3]
-; SSE-NEXT:    pshuflw {{.*#+}} xmm4 = xmm0[0,2,2,3,4,5,6,7]
-; SSE-NEXT:    punpckldq {{.*#+}} xmm4 = xmm4[0],xmm1[0],xmm4[1],xmm1[1]
-; SSE-NEXT:    pshufd {{.*#+}} xmm0 = xmm3[0,2,2,3]
-; SSE-NEXT:    pshuflw {{.*#+}} xmm1 = xmm0[0,1,0,2,4,5,6,7]
-; SSE-NEXT:    pshufd {{.*#+}} xmm0 = xmm2[0,2,2,3]
-; SSE-NEXT:    pshuflw {{.*#+}} xmm0 = xmm0[0,1,0,2,4,5,6,7]
-; SSE-NEXT:    punpckldq {{.*#+}} xmm0 = xmm0[0],xmm1[0],xmm0[1],xmm1[1]
-; SSE-NEXT:    movsd {{.*#+}} xmm0 = xmm4[0],xmm0[1]
-; SSE-NEXT:    xorpd {{.*}}(%rip), %xmm0
+; SSE-NEXT:    punpcklqdq {{.*#+}} xmm0 = xmm0[0],xmm2[0]
+; SSE-NEXT:    pxor {{.*}}(%rip), %xmm0
 ; SSE-NEXT:    retq
 ;
 ; AVX1-LABEL: trunc_xor_const_v8i64_v8i16:
@@ -4339,21 +4317,19 @@
 define <8 x i16> @trunc_or_v8i64_v8i16(<8 x i64> %a0, <8 x i64> %a1) nounwind {
 ; SSE-LABEL: trunc_or_v8i64_v8i16:
 ; SSE:       # %bb.0:
-; SSE-NEXT:    por %xmm6, %xmm2
-; SSE-NEXT:    por %xmm7, %xmm3
-; SSE-NEXT:    por %xmm4, %xmm0
-; SSE-NEXT:    por %xmm5, %xmm1
+; SSE-NEXT:    orps %xmm5, %xmm1
+; SSE-NEXT:    orps %xmm4, %xmm0
+; SSE-NEXT:    shufps {{.*#+}} xmm0 = xmm0[0,2],xmm1[0,2]
+; SSE-NEXT:    orps %xmm7, %xmm3
+; SSE-NEXT:    orps %xmm6, %xmm2
+; SSE-NEXT:    shufps {{.*#+}} xmm2 = xmm2[0,2],xmm3[0,2]
+; SSE-NEXT:    pshuflw {{.*#+}} xmm1 = xmm2[0,2,2,3,4,5,6,7]
+; SSE-NEXT:    pshufhw {{.*#+}} xmm1 = xmm1[0,1,2,3,4,6,6,7]
 ; SSE-NEXT:    pshufd {{.*#+}} xmm1 = xmm1[0,2,2,3]
-; SSE-NEXT:    pshuflw {{.*#+}} xmm1 = xmm1[0,2,2,3,4,5,6,7]
+; SSE-NEXT:    pshuflw {{.*#+}} xmm0 = xmm0[0,2,2,3,4,5,6,7]
+; SSE-NEXT:    pshufhw {{.*#+}} xmm0 = xmm0[0,1,2,3,4,6,6,7]
 ; SSE-NEXT:    pshufd {{.*#+}} xmm0 = xmm0[0,2,2,3]
-; SSE-NEXT:    pshuflw {{.*#+}} xmm4 = xmm0[0,2,2,3,4,5,6,7]
-; SSE-NEXT:    punpckldq {{.*#+}} xmm4 = xmm4[0],xmm1[0],xmm4[1],xmm1[1]
-; SSE-NEXT:    pshufd {{.*#+}} xmm0 = xmm3[0,2,2,3]
-; SSE-NEXT:    pshuflw {{.*#+}} xmm1 = xmm0[0,1,0,2,4,5,6,7]
-; SSE-NEXT:    pshufd {{.*#+}} xmm0 = xmm2[0,2,2,3]
-; SSE-NEXT:    pshuflw {{.*#+}} xmm0 = xmm0[0,1,0,2,4,5,6,7]
-; SSE-NEXT:    punpckldq {{.*#+}} xmm0 = xmm0[0],xmm1[0],xmm0[1],xmm1[1]
-; SSE-NEXT:    movsd {{.*#+}} xmm0 = xmm4[0],xmm0[1]
+; SSE-NEXT:    punpcklqdq {{.*#+}} xmm0 = xmm0[0],xmm1[0]
 ; SSE-NEXT:    retq
 ;
 ; AVX1-LABEL: trunc_or_v8i64_v8i16:
@@ -4736,18 +4712,16 @@
 define <8 x i16> @trunc_or_const_v8i64_v8i16(<8 x i64> %a0) nounwind {
 ; SSE-LABEL: trunc_or_const_v8i64_v8i16:
 ; SSE:       # %bb.0:
-; SSE-NEXT:    pshufd {{.*#+}} xmm1 = xmm1[0,2,2,3]
-; SSE-NEXT:    pshuflw {{.*#+}} xmm1 = xmm1[0,2,2,3,4,5,6,7]
+; SSE-NEXT:    shufps {{.*#+}} xmm2 = xmm2[0,2],xmm3[0,2]
+; SSE-NEXT:    pshuflw {{.*#+}} xmm2 = xmm2[0,2,2,3,4,5,6,7]
+; SSE-NEXT:    pshufhw {{.*#+}} xmm2 = xmm2[0,1,2,3,4,6,6,7]
+; SSE-NEXT:    pshufd {{.*#+}} xmm2 = xmm2[0,2,2,3]
+; SSE-NEXT:    shufps {{.*#+}} xmm0 = xmm0[0,2],xmm1[0,2]
+; SSE-NEXT:    pshuflw {{.*#+}} xmm0 = xmm0[0,2,2,3,4,5,6,7]
+; SSE-NEXT:    pshufhw {{.*#+}} xmm0 = xmm0[0,1,2,3,4,6,6,7]
 ; SSE-NEXT:    pshufd {{.*#+}} xmm0 = xmm0[0,2,2,3]
-; SSE-NEXT:    pshuflw {{.*#+}} xmm4 = xmm0[0,2,2,3,4,5,6,7]
-; SSE-NEXT:    punpckldq {{.*#+}} xmm4 = xmm4[0],xmm1[0],xmm4[1],xmm1[1]
-; SSE-NEXT:    pshufd {{.*#+}} xmm0 = xmm3[0,2,2,3]
-; SSE-NEXT:    pshuflw {{.*#+}} xmm1 = xmm0[0,1,0,2,4,5,6,7]
-; SSE-NEXT:    pshufd {{.*#+}} xmm0 = xmm2[0,2,2,3]
-; SSE-NEXT:    pshuflw {{.*#+}} xmm0 = xmm0[0,1,0,2,4,5,6,7]
-; SSE-NEXT:    punpckldq {{.*#+}} xmm0 = xmm0[0],xmm1[0],xmm0[1],xmm1[1]
-; SSE-NEXT:    movsd {{.*#+}} xmm0 = xmm4[0],xmm0[1]
-; SSE-NEXT:    orpd {{.*}}(%rip), %xmm0
+; SSE-NEXT:    punpcklqdq {{.*#+}} xmm0 = xmm0[0],xmm2[0]
+; SSE-NEXT:    por {{.*}}(%rip), %xmm0
 ; SSE-NEXT:    retq
 ;
 ; AVX1-LABEL: trunc_or_const_v8i64_v8i16:
diff --git a/llvm/test/CodeGen/X86/vector-trunc-packus.ll b/llvm/test/CodeGen/X86/vector-trunc-packus.ll
--- a/llvm/test/CodeGen/X86/vector-trunc-packus.ll
+++ b/llvm/test/CodeGen/X86/vector-trunc-packus.ll
@@ -1859,22 +1859,22 @@
 define <8 x i16> @trunc_packus_v8i64_v8i16(<8 x i64>* %p0) "min-legal-vector-width"="256" {
 ; SSE2-LABEL: trunc_packus_v8i64_v8i16:
 ; SSE2:       # %bb.0:
-; SSE2-NEXT:    movdqa (%rdi), %xmm7
-; SSE2-NEXT:    movdqa 16(%rdi), %xmm2
-; SSE2-NEXT:    movdqa 32(%rdi), %xmm9
-; SSE2-NEXT:    movdqa 48(%rdi), %xmm6
+; SSE2-NEXT:    movdqa (%rdi), %xmm5
+; SSE2-NEXT:    movdqa 16(%rdi), %xmm9
+; SSE2-NEXT:    movdqa 32(%rdi), %xmm2
+; SSE2-NEXT:    movdqa 48(%rdi), %xmm7
 ; SSE2-NEXT:    movdqa {{.*#+}} xmm8 = [65535,65535]
 ; SSE2-NEXT:    movdqa {{.*#+}} xmm11 = [2147483648,2147483648]
 ; SSE2-NEXT:    movdqa %xmm2, %xmm1
 ; SSE2-NEXT:    pxor %xmm11, %xmm1
 ; SSE2-NEXT:    movdqa {{.*#+}} xmm10 = [2147549183,2147549183]
-; SSE2-NEXT:    movdqa %xmm10, %xmm5
-; SSE2-NEXT:    pcmpgtd %xmm1, %xmm5
-; SSE2-NEXT:    pshufd {{.*#+}} xmm3 = xmm5[0,0,2,2]
+; SSE2-NEXT:    movdqa %xmm10, %xmm6
+; SSE2-NEXT:    pcmpgtd %xmm1, %xmm6
+; SSE2-NEXT:    pshufd {{.*#+}} xmm3 = xmm6[0,0,2,2]
 ; SSE2-NEXT:    pcmpeqd %xmm10, %xmm1
 ; SSE2-NEXT:    pshufd {{.*#+}} xmm4 = xmm1[1,1,3,3]
 ; SSE2-NEXT:    pand %xmm3, %xmm4
-; SSE2-NEXT:    pshufd {{.*#+}} xmm1 = xmm5[1,1,3,3]
+; SSE2-NEXT:    pshufd {{.*#+}} xmm1 = xmm6[1,1,3,3]
 ; SSE2-NEXT:    por %xmm4, %xmm1
 ; SSE2-NEXT:    pand %xmm1, %xmm2
 ; SSE2-NEXT:    pandn %xmm8, %xmm1
@@ -1885,26 +1885,26 @@
 ; SSE2-NEXT:    pcmpgtd %xmm2, %xmm3
 ; SSE2-NEXT:    pshufd {{.*#+}} xmm4 = xmm3[0,0,2,2]
 ; SSE2-NEXT:    pcmpeqd %xmm10, %xmm2
-; SSE2-NEXT:    pshufd {{.*#+}} xmm5 = xmm2[1,1,3,3]
-; SSE2-NEXT:    pand %xmm4, %xmm5
+; SSE2-NEXT:    pshufd {{.*#+}} xmm6 = xmm2[1,1,3,3]
+; SSE2-NEXT:    pand %xmm4, %xmm6
 ; SSE2-NEXT:    pshufd {{.*#+}} xmm2 = xmm3[1,1,3,3]
-; SSE2-NEXT:    por %xmm5, %xmm2
+; SSE2-NEXT:    por %xmm6, %xmm2
 ; SSE2-NEXT:    pand %xmm2, %xmm7
 ; SSE2-NEXT:    pandn %xmm8, %xmm2
 ; SSE2-NEXT:    por %xmm7, %xmm2
-; SSE2-NEXT:    movdqa %xmm6, %xmm3
+; SSE2-NEXT:    movdqa %xmm5, %xmm3
 ; SSE2-NEXT:    pxor %xmm11, %xmm3
 ; SSE2-NEXT:    movdqa %xmm10, %xmm4
 ; SSE2-NEXT:    pcmpgtd %xmm3, %xmm4
-; SSE2-NEXT:    pshufd {{.*#+}} xmm5 = xmm4[0,0,2,2]
+; SSE2-NEXT:    pshufd {{.*#+}} xmm6 = xmm4[0,0,2,2]
 ; SSE2-NEXT:    pcmpeqd %xmm10, %xmm3
 ; SSE2-NEXT:    pshufd {{.*#+}} xmm3 = xmm3[1,1,3,3]
-; SSE2-NEXT:    pand %xmm5, %xmm3
+; SSE2-NEXT:    pand %xmm6, %xmm3
 ; SSE2-NEXT:    pshufd {{.*#+}} xmm7 = xmm4[1,1,3,3]
 ; SSE2-NEXT:    por %xmm3, %xmm7
-; SSE2-NEXT:    pand %xmm7, %xmm6
+; SSE2-NEXT:    pand %xmm7, %xmm5
 ; SSE2-NEXT:    pandn %xmm8, %xmm7
-; SSE2-NEXT:    por %xmm6, %xmm7
+; SSE2-NEXT:    por %xmm5, %xmm7
 ; SSE2-NEXT:    movdqa %xmm9, %xmm3
 ; SSE2-NEXT:    pxor %xmm11, %xmm3
 ; SSE2-NEXT:    movdqa %xmm10, %xmm4
@@ -1924,12 +1924,24 @@
 ; SSE2-NEXT:    pcmpgtd %xmm11, %xmm5
 ; SSE2-NEXT:    pshufd {{.*#+}} xmm6 = xmm5[0,0,2,2]
 ; SSE2-NEXT:    pcmpeqd %xmm11, %xmm3
+; SSE2-NEXT:    pshufd {{.*#+}} xmm3 = xmm3[1,1,3,3]
+; SSE2-NEXT:    pand %xmm6, %xmm3
+; SSE2-NEXT:    pshufd {{.*#+}} xmm5 = xmm5[1,1,3,3]
+; SSE2-NEXT:    por %xmm3, %xmm5
+; SSE2-NEXT:    pand %xmm4, %xmm5
+; SSE2-NEXT:    movdqa %xmm7, %xmm3
+; SSE2-NEXT:    pxor %xmm11, %xmm3
+; SSE2-NEXT:    movdqa %xmm3, %xmm4
+; SSE2-NEXT:    pcmpgtd %xmm11, %xmm4
+; SSE2-NEXT:    pshufd {{.*#+}} xmm6 = xmm4[0,0,2,2]
+; SSE2-NEXT:    pcmpeqd %xmm11, %xmm3
 ; SSE2-NEXT:    pshufd {{.*#+}} xmm0 = xmm3[1,1,3,3]
 ; SSE2-NEXT:    pand %xmm6, %xmm0
-; SSE2-NEXT:    pshufd {{.*#+}} xmm3 = xmm5[1,1,3,3]
+; SSE2-NEXT:    pshufd {{.*#+}} xmm3 = xmm4[1,1,3,3]
 ; SSE2-NEXT:    por %xmm0, %xmm3
-; SSE2-NEXT:    pand %xmm4, %xmm3
-; SSE2-NEXT:    movdqa %xmm7, %xmm0
+; SSE2-NEXT:    pand %xmm7, %xmm3
+; SSE2-NEXT:    shufps {{.*#+}} xmm3 = xmm3[0,2],xmm5[0,2]
+; SSE2-NEXT:    movdqa %xmm2, %xmm0
 ; SSE2-NEXT:    pxor %xmm11, %xmm0
 ; SSE2-NEXT:    movdqa %xmm0, %xmm4
 ; SSE2-NEXT:    pcmpgtd %xmm11, %xmm4
@@ -1939,126 +1951,90 @@
 ; SSE2-NEXT:    pand %xmm5, %xmm0
 ; SSE2-NEXT:    pshufd {{.*#+}} xmm4 = xmm4[1,1,3,3]
 ; SSE2-NEXT:    por %xmm0, %xmm4
-; SSE2-NEXT:    pand %xmm7, %xmm4
-; SSE2-NEXT:    movdqa %xmm2, %xmm0
-; SSE2-NEXT:    pxor %xmm11, %xmm0
-; SSE2-NEXT:    movdqa %xmm0, %xmm5
-; SSE2-NEXT:    pcmpgtd %xmm11, %xmm5
-; SSE2-NEXT:    pshufd {{.*#+}} xmm6 = xmm5[0,0,2,2]
-; SSE2-NEXT:    pcmpeqd %xmm11, %xmm0
-; SSE2-NEXT:    pshufd {{.*#+}} xmm0 = xmm0[1,1,3,3]
-; SSE2-NEXT:    pand %xmm6, %xmm0
-; SSE2-NEXT:    pshufd {{.*#+}} xmm5 = xmm5[1,1,3,3]
-; SSE2-NEXT:    por %xmm0, %xmm5
-; SSE2-NEXT:    pand %xmm2, %xmm5
+; SSE2-NEXT:    pand %xmm2, %xmm4
 ; SSE2-NEXT:    movdqa %xmm1, %xmm0
 ; SSE2-NEXT:    pxor %xmm11, %xmm0
 ; SSE2-NEXT:    movdqa %xmm0, %xmm2
 ; SSE2-NEXT:    pcmpgtd %xmm11, %xmm2
-; SSE2-NEXT:    pshufd {{.*#+}} xmm6 = xmm2[0,0,2,2]
+; SSE2-NEXT:    pshufd {{.*#+}} xmm5 = xmm2[0,0,2,2]
 ; SSE2-NEXT:    pcmpeqd %xmm11, %xmm0
 ; SSE2-NEXT:    pshufd {{.*#+}} xmm0 = xmm0[1,1,3,3]
-; SSE2-NEXT:    pand %xmm6, %xmm0
+; SSE2-NEXT:    pand %xmm5, %xmm0
 ; SSE2-NEXT:    pshufd {{.*#+}} xmm2 = xmm2[1,1,3,3]
 ; SSE2-NEXT:    por %xmm0, %xmm2
 ; SSE2-NEXT:    pand %xmm1, %xmm2
-; SSE2-NEXT:    pshufd {{.*#+}} xmm0 = xmm2[0,2,2,3]
-; SSE2-NEXT:    pshuflw {{.*#+}} xmm0 = xmm0[0,2,2,3,4,5,6,7]
-; SSE2-NEXT:    pshufd {{.*#+}} xmm1 = xmm5[0,2,2,3]
-; SSE2-NEXT:    pshuflw {{.*#+}} xmm1 = xmm1[0,2,2,3,4,5,6,7]
-; SSE2-NEXT:    punpckldq {{.*#+}} xmm1 = xmm1[0],xmm0[0],xmm1[1],xmm0[1]
-; SSE2-NEXT:    pshufd {{.*#+}} xmm0 = xmm4[0,2,2,3]
-; SSE2-NEXT:    pshuflw {{.*#+}} xmm2 = xmm0[0,1,0,2,4,5,6,7]
-; SSE2-NEXT:    pshufd {{.*#+}} xmm0 = xmm3[0,2,2,3]
-; SSE2-NEXT:    pshuflw {{.*#+}} xmm0 = xmm0[0,1,0,2,4,5,6,7]
-; SSE2-NEXT:    punpckldq {{.*#+}} xmm0 = xmm0[0],xmm2[0],xmm0[1],xmm2[1]
-; SSE2-NEXT:    movsd {{.*#+}} xmm0 = xmm1[0],xmm0[1]
+; SSE2-NEXT:    shufps {{.*#+}} xmm2 = xmm2[0,2],xmm4[0,2]
+; SSE2-NEXT:    pshuflw {{.*#+}} xmm0 = xmm2[0,2,2,3,4,5,6,7]
+; SSE2-NEXT:    pshufhw {{.*#+}} xmm0 = xmm0[0,1,2,3,4,6,6,7]
+; SSE2-NEXT:    pshufd {{.*#+}} xmm1 = xmm0[0,2,2,3]
+; SSE2-NEXT:    pshuflw {{.*#+}} xmm0 = xmm3[0,2,2,3,4,5,6,7]
+; SSE2-NEXT:    pshufhw {{.*#+}} xmm0 = xmm0[0,1,2,3,4,6,6,7]
+; SSE2-NEXT:    pshufd {{.*#+}} xmm0 = xmm0[0,2,2,3]
+; SSE2-NEXT:    punpcklqdq {{.*#+}} xmm0 = xmm0[0],xmm1[0]
 ; SSE2-NEXT:    retq
 ;
 ; SSSE3-LABEL: trunc_packus_v8i64_v8i16:
 ; SSSE3:       # %bb.0:
-; SSSE3-NEXT:    movdqa (%rdi), %xmm7
-; SSSE3-NEXT:    movdqa 16(%rdi), %xmm2
-; SSSE3-NEXT:    movdqa 32(%rdi), %xmm9
-; SSSE3-NEXT:    movdqa 48(%rdi), %xmm6
+; SSSE3-NEXT:    movdqa (%rdi), %xmm5
+; SSSE3-NEXT:    movdqa 16(%rdi), %xmm9
+; SSSE3-NEXT:    movdqa 32(%rdi), %xmm3
+; SSSE3-NEXT:    movdqa 48(%rdi), %xmm7
 ; SSSE3-NEXT:    movdqa {{.*#+}} xmm8 = [65535,65535]
 ; SSSE3-NEXT:    movdqa {{.*#+}} xmm11 = [2147483648,2147483648]
-; SSSE3-NEXT:    movdqa %xmm2, %xmm1
-; SSSE3-NEXT:    pxor %xmm11, %xmm1
-; SSSE3-NEXT:    movdqa {{.*#+}} xmm10 = [2147549183,2147549183]
-; SSSE3-NEXT:    movdqa %xmm10, %xmm5
-; SSSE3-NEXT:    pcmpgtd %xmm1, %xmm5
-; SSSE3-NEXT:    pshufd {{.*#+}} xmm3 = xmm5[0,0,2,2]
-; SSSE3-NEXT:    pcmpeqd %xmm10, %xmm1
-; SSSE3-NEXT:    pshufd {{.*#+}} xmm4 = xmm1[1,1,3,3]
-; SSSE3-NEXT:    pand %xmm3, %xmm4
-; SSSE3-NEXT:    pshufd {{.*#+}} xmm1 = xmm5[1,1,3,3]
-; SSSE3-NEXT:    por %xmm4, %xmm1
-; SSSE3-NEXT:    pand %xmm1, %xmm2
-; SSSE3-NEXT:    pandn %xmm8, %xmm1
-; SSSE3-NEXT:    por %xmm2, %xmm1
-; SSSE3-NEXT:    movdqa %xmm7, %xmm2
+; SSSE3-NEXT:    movdqa %xmm3, %xmm2
 ; SSSE3-NEXT:    pxor %xmm11, %xmm2
-; SSSE3-NEXT:    movdqa %xmm10, %xmm3
-; SSSE3-NEXT:    pcmpgtd %xmm2, %xmm3
-; SSSE3-NEXT:    pshufd {{.*#+}} xmm4 = xmm3[0,0,2,2]
+; SSSE3-NEXT:    movdqa {{.*#+}} xmm10 = [2147549183,2147549183]
+; SSSE3-NEXT:    movdqa %xmm10, %xmm6
+; SSSE3-NEXT:    pcmpgtd %xmm2, %xmm6
+; SSSE3-NEXT:    pshufd {{.*#+}} xmm0 = xmm6[0,0,2,2]
 ; SSSE3-NEXT:    pcmpeqd %xmm10, %xmm2
-; SSSE3-NEXT:    pshufd {{.*#+}} xmm5 = xmm2[1,1,3,3]
-; SSSE3-NEXT:    pand %xmm4, %xmm5
-; SSSE3-NEXT:    pshufd {{.*#+}} xmm2 = xmm3[1,1,3,3]
-; SSSE3-NEXT:    por %xmm5, %xmm2
-; SSSE3-NEXT:    pand %xmm2, %xmm7
+; SSSE3-NEXT:    pshufd {{.*#+}} xmm4 = xmm2[1,1,3,3]
+; SSSE3-NEXT:    pand %xmm0, %xmm4
+; SSSE3-NEXT:    pshufd {{.*#+}} xmm2 = xmm6[1,1,3,3]
+; SSSE3-NEXT:    por %xmm4, %xmm2
+; SSSE3-NEXT:    pand %xmm2, %xmm3
 ; SSSE3-NEXT:    pandn %xmm8, %xmm2
-; SSSE3-NEXT:    por %xmm7, %xmm2
-; SSSE3-NEXT:    movdqa %xmm6, %xmm3
-; SSSE3-NEXT:    pxor %xmm11, %xmm3
-; SSSE3-NEXT:    movdqa %xmm10, %xmm4
-; SSSE3-NEXT:    pcmpgtd %xmm3, %xmm4
-; SSSE3-NEXT:    pshufd {{.*#+}} xmm5 = xmm4[0,0,2,2]
-; SSSE3-NEXT:    pcmpeqd %xmm10, %xmm3
+; SSSE3-NEXT:    por %xmm3, %xmm2
+; SSSE3-NEXT:    movdqa %xmm7, %xmm0
+; SSSE3-NEXT:    pxor %xmm11, %xmm0
+; SSSE3-NEXT:    movdqa %xmm10, %xmm3
+; SSSE3-NEXT:    pcmpgtd %xmm0, %xmm3
+; SSSE3-NEXT:    pshufd {{.*#+}} xmm4 = xmm3[0,0,2,2]
+; SSSE3-NEXT:    pcmpeqd %xmm10, %xmm0
+; SSSE3-NEXT:    pshufd {{.*#+}} xmm0 = xmm0[1,1,3,3]
+; SSSE3-NEXT:    pand %xmm4, %xmm0
 ; SSSE3-NEXT:    pshufd {{.*#+}} xmm3 = xmm3[1,1,3,3]
-; SSSE3-NEXT:    pand %xmm5, %xmm3
-; SSSE3-NEXT:    pshufd {{.*#+}} xmm7 = xmm4[1,1,3,3]
-; SSSE3-NEXT:    por %xmm3, %xmm7
-; SSSE3-NEXT:    pand %xmm7, %xmm6
-; SSSE3-NEXT:    pandn %xmm8, %xmm7
-; SSSE3-NEXT:    por %xmm6, %xmm7
-; SSSE3-NEXT:    movdqa %xmm9, %xmm3
-; SSSE3-NEXT:    pxor %xmm11, %xmm3
+; SSSE3-NEXT:    por %xmm0, %xmm3
+; SSSE3-NEXT:    pand %xmm3, %xmm7
+; SSSE3-NEXT:    pandn %xmm8, %xmm3
+; SSSE3-NEXT:    por %xmm7, %xmm3
+; SSSE3-NEXT:    movdqa %xmm5, %xmm0
+; SSSE3-NEXT:    pxor %xmm11, %xmm0
 ; SSSE3-NEXT:    movdqa %xmm10, %xmm4
-; SSSE3-NEXT:    pcmpgtd %xmm3, %xmm4
-; SSSE3-NEXT:    pshufd {{.*#+}} xmm5 = xmm4[0,0,2,2]
-; SSSE3-NEXT:    pcmpeqd %xmm10, %xmm3
-; SSSE3-NEXT:    pshufd {{.*#+}} xmm3 = xmm3[1,1,3,3]
-; SSSE3-NEXT:    pand %xmm5, %xmm3
-; SSSE3-NEXT:    pshufd {{.*#+}} xmm4 = xmm4[1,1,3,3]
-; SSSE3-NEXT:    por %xmm3, %xmm4
-; SSSE3-NEXT:    pand %xmm4, %xmm9
-; SSSE3-NEXT:    pandn %xmm8, %xmm4
-; SSSE3-NEXT:    por %xmm9, %xmm4
-; SSSE3-NEXT:    movdqa %xmm4, %xmm3
-; SSSE3-NEXT:    pxor %xmm11, %xmm3
-; SSSE3-NEXT:    movdqa %xmm3, %xmm5
-; SSSE3-NEXT:    pcmpgtd %xmm11, %xmm5
-; SSSE3-NEXT:    pshufd {{.*#+}} xmm6 = xmm5[0,0,2,2]
-; SSSE3-NEXT:    pcmpeqd %xmm11, %xmm3
-; SSSE3-NEXT:    pshufd {{.*#+}} xmm0 = xmm3[1,1,3,3]
+; SSSE3-NEXT:    pcmpgtd %xmm0, %xmm4
+; SSSE3-NEXT:    pshufd {{.*#+}} xmm6 = xmm4[0,0,2,2]
+; SSSE3-NEXT:    pcmpeqd %xmm10, %xmm0
+; SSSE3-NEXT:    pshufd {{.*#+}} xmm0 = xmm0[1,1,3,3]
 ; SSSE3-NEXT:    pand %xmm6, %xmm0
-; SSSE3-NEXT:    pshufd {{.*#+}} xmm3 = xmm5[1,1,3,3]
-; SSSE3-NEXT:    por %xmm0, %xmm3
-; SSSE3-NEXT:    pand %xmm4, %xmm3
-; SSSE3-NEXT:    movdqa %xmm7, %xmm0
+; SSSE3-NEXT:    pshufd {{.*#+}} xmm7 = xmm4[1,1,3,3]
+; SSSE3-NEXT:    por %xmm0, %xmm7
+; SSSE3-NEXT:    pand %xmm7, %xmm5
+; SSSE3-NEXT:    pandn %xmm8, %xmm7
+; SSSE3-NEXT:    por %xmm5, %xmm7
+; SSSE3-NEXT:    movdqa %xmm9, %xmm0
 ; SSSE3-NEXT:    pxor %xmm11, %xmm0
-; SSSE3-NEXT:    movdqa %xmm0, %xmm4
-; SSSE3-NEXT:    pcmpgtd %xmm11, %xmm4
+; SSSE3-NEXT:    movdqa %xmm10, %xmm4
+; SSSE3-NEXT:    pcmpgtd %xmm0, %xmm4
 ; SSSE3-NEXT:    pshufd {{.*#+}} xmm5 = xmm4[0,0,2,2]
-; SSSE3-NEXT:    pcmpeqd %xmm11, %xmm0
+; SSSE3-NEXT:    pcmpeqd %xmm10, %xmm0
 ; SSSE3-NEXT:    pshufd {{.*#+}} xmm0 = xmm0[1,1,3,3]
 ; SSSE3-NEXT:    pand %xmm5, %xmm0
 ; SSSE3-NEXT:    pshufd {{.*#+}} xmm4 = xmm4[1,1,3,3]
 ; SSSE3-NEXT:    por %xmm0, %xmm4
-; SSSE3-NEXT:    pand %xmm7, %xmm4
-; SSSE3-NEXT:    movdqa %xmm2, %xmm0
+; SSSE3-NEXT:    pand %xmm4, %xmm9
+; SSSE3-NEXT:    pandn %xmm8, %xmm4
+; SSSE3-NEXT:    por %xmm9, %xmm4
+; SSSE3-NEXT:    movdqa %xmm4, %xmm0
 ; SSSE3-NEXT:    pxor %xmm11, %xmm0
 ; SSSE3-NEXT:    movdqa %xmm0, %xmm5
 ; SSSE3-NEXT:    pcmpgtd %xmm11, %xmm5
@@ -2068,29 +2044,46 @@
 ; SSSE3-NEXT:    pand %xmm6, %xmm0
 ; SSSE3-NEXT:    pshufd {{.*#+}} xmm5 = xmm5[1,1,3,3]
 ; SSSE3-NEXT:    por %xmm0, %xmm5
-; SSSE3-NEXT:    pand %xmm2, %xmm5
-; SSSE3-NEXT:    movdqa %xmm1, %xmm0
+; SSSE3-NEXT:    pand %xmm4, %xmm5
+; SSSE3-NEXT:    movdqa %xmm7, %xmm0
 ; SSSE3-NEXT:    pxor %xmm11, %xmm0
-; SSSE3-NEXT:    movdqa %xmm0, %xmm2
-; SSSE3-NEXT:    pcmpgtd %xmm11, %xmm2
-; SSSE3-NEXT:    pshufd {{.*#+}} xmm6 = xmm2[0,0,2,2]
+; SSSE3-NEXT:    movdqa %xmm0, %xmm4
+; SSSE3-NEXT:    pcmpgtd %xmm11, %xmm4
+; SSSE3-NEXT:    pshufd {{.*#+}} xmm6 = xmm4[0,0,2,2]
 ; SSSE3-NEXT:    pcmpeqd %xmm11, %xmm0
-; SSSE3-NEXT:    pshufd {{.*#+}} xmm0 = xmm0[1,1,3,3]
-; SSSE3-NEXT:    pand %xmm6, %xmm0
-; SSSE3-NEXT:    pshufd {{.*#+}} xmm2 = xmm2[1,1,3,3]
-; SSSE3-NEXT:    por %xmm0, %xmm2
-; SSSE3-NEXT:    pand %xmm1, %xmm2
-; SSSE3-NEXT:    pshufd {{.*#+}} xmm0 = xmm2[0,2,2,3]
-; SSSE3-NEXT:    pshuflw {{.*#+}} xmm0 = xmm0[0,2,2,3,4,5,6,7]
-; SSSE3-NEXT:    pshufd {{.*#+}} xmm1 = xmm5[0,2,2,3]
-; SSSE3-NEXT:    pshuflw {{.*#+}} xmm1 = xmm1[0,2,2,3,4,5,6,7]
-; SSSE3-NEXT:    punpckldq {{.*#+}} xmm1 = xmm1[0],xmm0[0],xmm1[1],xmm0[1]
-; SSSE3-NEXT:    pshufd {{.*#+}} xmm0 = xmm4[0,2,2,3]
-; SSSE3-NEXT:    pshuflw {{.*#+}} xmm2 = xmm0[0,1,0,2,4,5,6,7]
-; SSSE3-NEXT:    pshufd {{.*#+}} xmm0 = xmm3[0,2,2,3]
-; SSSE3-NEXT:    pshuflw {{.*#+}} xmm0 = xmm0[0,1,0,2,4,5,6,7]
-; SSSE3-NEXT:    punpckldq {{.*#+}} xmm0 = xmm0[0],xmm2[0],xmm0[1],xmm2[1]
-; SSSE3-NEXT:    movsd {{.*#+}} xmm0 = xmm1[0],xmm0[1]
+; SSSE3-NEXT:    pshufd {{.*#+}} xmm1 = xmm0[1,1,3,3]
+; SSSE3-NEXT:    pand %xmm6, %xmm1
+; SSSE3-NEXT:    pshufd {{.*#+}} xmm0 = xmm4[1,1,3,3]
+; SSSE3-NEXT:    por %xmm1, %xmm0
+; SSSE3-NEXT:    pand %xmm7, %xmm0
+; SSSE3-NEXT:    shufps {{.*#+}} xmm0 = xmm0[0,2],xmm5[0,2]
+; SSSE3-NEXT:    movdqa %xmm3, %xmm1
+; SSSE3-NEXT:    pxor %xmm11, %xmm1
+; SSSE3-NEXT:    movdqa %xmm1, %xmm4
+; SSSE3-NEXT:    pcmpgtd %xmm11, %xmm4
+; SSSE3-NEXT:    pshufd {{.*#+}} xmm5 = xmm4[0,0,2,2]
+; SSSE3-NEXT:    pcmpeqd %xmm11, %xmm1
+; SSSE3-NEXT:    pshufd {{.*#+}} xmm1 = xmm1[1,1,3,3]
+; SSSE3-NEXT:    pand %xmm5, %xmm1
+; SSSE3-NEXT:    pshufd {{.*#+}} xmm4 = xmm4[1,1,3,3]
+; SSSE3-NEXT:    por %xmm1, %xmm4
+; SSSE3-NEXT:    pand %xmm3, %xmm4
+; SSSE3-NEXT:    movdqa %xmm2, %xmm1
+; SSSE3-NEXT:    pxor %xmm11, %xmm1
+; SSSE3-NEXT:    movdqa %xmm1, %xmm3
+; SSSE3-NEXT:    pcmpgtd %xmm11, %xmm3
+; SSSE3-NEXT:    pshufd {{.*#+}} xmm5 = xmm3[0,0,2,2]
+; SSSE3-NEXT:    pcmpeqd %xmm11, %xmm1
+; SSSE3-NEXT:    pshufd {{.*#+}} xmm1 = xmm1[1,1,3,3]
+; SSSE3-NEXT:    pand %xmm5, %xmm1
+; SSSE3-NEXT:    pshufd {{.*#+}} xmm3 = xmm3[1,1,3,3]
+; SSSE3-NEXT:    por %xmm1, %xmm3
+; SSSE3-NEXT:    pand %xmm2, %xmm3
+; SSSE3-NEXT:    shufps {{.*#+}} xmm3 = xmm3[0,2],xmm4[0,2]
+; SSSE3-NEXT:    movdqa {{.*#+}} xmm1 = [0,1,4,5,8,9,12,13,8,9,12,13,12,13,14,15]
+; SSSE3-NEXT:    pshufb %xmm1, %xmm3
+; SSSE3-NEXT:    pshufb %xmm1, %xmm0
+; SSSE3-NEXT:    punpcklqdq {{.*#+}} xmm0 = xmm0[0],xmm3[0]
 ; SSSE3-NEXT:    retq
 ;
 ; SSE41-LABEL: trunc_packus_v8i64_v8i16:
diff --git a/llvm/test/CodeGen/X86/vector-trunc-usat.ll b/llvm/test/CodeGen/X86/vector-trunc-usat.ll
--- a/llvm/test/CodeGen/X86/vector-trunc-usat.ll
+++ b/llvm/test/CodeGen/X86/vector-trunc-usat.ll
@@ -1301,148 +1301,141 @@
 define <8 x i16> @trunc_usat_v8i64_v8i16(<8 x i64>* %p0) {
 ; SSE2-LABEL: trunc_usat_v8i64_v8i16:
 ; SSE2:       # %bb.0:
-; SSE2-NEXT:    movdqa (%rdi), %xmm4
-; SSE2-NEXT:    movdqa 16(%rdi), %xmm9
-; SSE2-NEXT:    movdqa 32(%rdi), %xmm6
-; SSE2-NEXT:    movdqa 48(%rdi), %xmm7
+; SSE2-NEXT:    movdqa (%rdi), %xmm6
+; SSE2-NEXT:    movdqa 16(%rdi), %xmm5
+; SSE2-NEXT:    movdqa 32(%rdi), %xmm9
+; SSE2-NEXT:    movdqa 48(%rdi), %xmm4
 ; SSE2-NEXT:    movdqa {{.*#+}} xmm8 = [65535,65535]
-; SSE2-NEXT:    movdqa {{.*#+}} xmm3 = [9223372039002259456,9223372039002259456]
-; SSE2-NEXT:    movdqa %xmm6, %xmm2
-; SSE2-NEXT:    pxor %xmm3, %xmm2
-; SSE2-NEXT:    movdqa {{.*#+}} xmm10 = [9223372039002324991,9223372039002324991]
-; SSE2-NEXT:    movdqa %xmm10, %xmm5
-; SSE2-NEXT:    pcmpgtd %xmm2, %xmm5
-; SSE2-NEXT:    pshufd {{.*#+}} xmm0 = xmm5[0,0,2,2]
-; SSE2-NEXT:    pcmpeqd %xmm10, %xmm2
-; SSE2-NEXT:    pshufd {{.*#+}} xmm1 = xmm2[1,1,3,3]
-; SSE2-NEXT:    pand %xmm0, %xmm1
-; SSE2-NEXT:    pshufd {{.*#+}} xmm2 = xmm5[1,1,3,3]
-; SSE2-NEXT:    por %xmm1, %xmm2
-; SSE2-NEXT:    pand %xmm2, %xmm6
-; SSE2-NEXT:    pandn %xmm8, %xmm2
-; SSE2-NEXT:    por %xmm6, %xmm2
-; SSE2-NEXT:    movdqa %xmm7, %xmm0
-; SSE2-NEXT:    pxor %xmm3, %xmm0
-; SSE2-NEXT:    movdqa %xmm10, %xmm1
-; SSE2-NEXT:    pcmpgtd %xmm0, %xmm1
-; SSE2-NEXT:    pshufd {{.*#+}} xmm5 = xmm1[0,0,2,2]
-; SSE2-NEXT:    pcmpeqd %xmm10, %xmm0
-; SSE2-NEXT:    pshufd {{.*#+}} xmm0 = xmm0[1,1,3,3]
-; SSE2-NEXT:    pand %xmm5, %xmm0
-; SSE2-NEXT:    pshufd {{.*#+}} xmm6 = xmm1[1,1,3,3]
-; SSE2-NEXT:    por %xmm0, %xmm6
-; SSE2-NEXT:    pand %xmm6, %xmm7
-; SSE2-NEXT:    pandn %xmm8, %xmm6
-; SSE2-NEXT:    por %xmm7, %xmm6
+; SSE2-NEXT:    movdqa {{.*#+}} xmm2 = [9223372039002259456,9223372039002259456]
+; SSE2-NEXT:    movdqa %xmm5, %xmm7
+; SSE2-NEXT:    pxor %xmm2, %xmm7
+; SSE2-NEXT:    movdqa {{.*#+}} xmm1 = [9223372039002324991,9223372039002324991]
+; SSE2-NEXT:    movdqa %xmm1, %xmm3
+; SSE2-NEXT:    pcmpgtd %xmm7, %xmm3
+; SSE2-NEXT:    pshufd {{.*#+}} xmm0 = xmm3[0,0,2,2]
+; SSE2-NEXT:    pcmpeqd %xmm1, %xmm7
+; SSE2-NEXT:    pshufd {{.*#+}} xmm7 = xmm7[1,1,3,3]
+; SSE2-NEXT:    pand %xmm0, %xmm7
+; SSE2-NEXT:    pshufd {{.*#+}} xmm0 = xmm3[1,1,3,3]
+; SSE2-NEXT:    por %xmm7, %xmm0
+; SSE2-NEXT:    pand %xmm0, %xmm5
+; SSE2-NEXT:    pandn %xmm8, %xmm0
+; SSE2-NEXT:    por %xmm5, %xmm0
+; SSE2-NEXT:    movdqa %xmm6, %xmm3
+; SSE2-NEXT:    pxor %xmm2, %xmm3
+; SSE2-NEXT:    movdqa %xmm1, %xmm5
+; SSE2-NEXT:    pcmpgtd %xmm3, %xmm5
+; SSE2-NEXT:    pshufd {{.*#+}} xmm7 = xmm5[0,0,2,2]
+; SSE2-NEXT:    pcmpeqd %xmm1, %xmm3
+; SSE2-NEXT:    pshufd {{.*#+}} xmm3 = xmm3[1,1,3,3]
+; SSE2-NEXT:    pand %xmm7, %xmm3
+; SSE2-NEXT:    pshufd {{.*#+}} xmm5 = xmm5[1,1,3,3]
+; SSE2-NEXT:    por %xmm3, %xmm5
+; SSE2-NEXT:    pand %xmm5, %xmm6
+; SSE2-NEXT:    pandn %xmm8, %xmm5
+; SSE2-NEXT:    por %xmm6, %xmm5
+; SSE2-NEXT:    shufps {{.*#+}} xmm5 = xmm5[0,2],xmm0[0,2]
 ; SSE2-NEXT:    movdqa %xmm4, %xmm0
-; SSE2-NEXT:    pxor %xmm3, %xmm0
-; SSE2-NEXT:    movdqa %xmm10, %xmm1
-; SSE2-NEXT:    pcmpgtd %xmm0, %xmm1
-; SSE2-NEXT:    pshufd {{.*#+}} xmm5 = xmm1[0,0,2,2]
-; SSE2-NEXT:    pcmpeqd %xmm10, %xmm0
+; SSE2-NEXT:    pxor %xmm2, %xmm0
+; SSE2-NEXT:    movdqa %xmm1, %xmm3
+; SSE2-NEXT:    pcmpgtd %xmm0, %xmm3
+; SSE2-NEXT:    pshufd {{.*#+}} xmm6 = xmm3[0,0,2,2]
+; SSE2-NEXT:    pcmpeqd %xmm1, %xmm0
 ; SSE2-NEXT:    pshufd {{.*#+}} xmm0 = xmm0[1,1,3,3]
-; SSE2-NEXT:    pand %xmm5, %xmm0
-; SSE2-NEXT:    pshufd {{.*#+}} xmm1 = xmm1[1,1,3,3]
-; SSE2-NEXT:    por %xmm0, %xmm1
-; SSE2-NEXT:    pand %xmm1, %xmm4
-; SSE2-NEXT:    pandn %xmm8, %xmm1
-; SSE2-NEXT:    por %xmm4, %xmm1
-; SSE2-NEXT:    pxor %xmm9, %xmm3
-; SSE2-NEXT:    movdqa %xmm10, %xmm0
-; SSE2-NEXT:    pcmpgtd %xmm3, %xmm0
-; SSE2-NEXT:    pshufd {{.*#+}} xmm4 = xmm0[0,0,2,2]
-; SSE2-NEXT:    pcmpeqd %xmm10, %xmm3
+; SSE2-NEXT:    pand %xmm6, %xmm0
 ; SSE2-NEXT:    pshufd {{.*#+}} xmm3 = xmm3[1,1,3,3]
-; SSE2-NEXT:    pand %xmm4, %xmm3
+; SSE2-NEXT:    por %xmm0, %xmm3
+; SSE2-NEXT:    pand %xmm3, %xmm4
+; SSE2-NEXT:    pandn %xmm8, %xmm3
+; SSE2-NEXT:    por %xmm4, %xmm3
+; SSE2-NEXT:    pxor %xmm9, %xmm2
+; SSE2-NEXT:    movdqa %xmm1, %xmm0
+; SSE2-NEXT:    pcmpgtd %xmm2, %xmm0
+; SSE2-NEXT:    pshufd {{.*#+}} xmm4 = xmm0[0,0,2,2]
+; SSE2-NEXT:    pcmpeqd %xmm1, %xmm2
+; SSE2-NEXT:    pshufd {{.*#+}} xmm1 = xmm2[1,1,3,3]
+; SSE2-NEXT:    pand %xmm4, %xmm1
 ; SSE2-NEXT:    pshufd {{.*#+}} xmm0 = xmm0[1,1,3,3]
-; SSE2-NEXT:    por %xmm3, %xmm0
+; SSE2-NEXT:    por %xmm1, %xmm0
 ; SSE2-NEXT:    pand %xmm0, %xmm9
 ; SSE2-NEXT:    pandn %xmm8, %xmm0
 ; SSE2-NEXT:    por %xmm9, %xmm0
-; SSE2-NEXT:    pshufd {{.*#+}} xmm0 = xmm0[0,2,2,3]
+; SSE2-NEXT:    shufps {{.*#+}} xmm0 = xmm0[0,2],xmm3[0,2]
 ; SSE2-NEXT:    pshuflw {{.*#+}} xmm0 = xmm0[0,2,2,3,4,5,6,7]
-; SSE2-NEXT:    pshufd {{.*#+}} xmm1 = xmm1[0,2,2,3]
-; SSE2-NEXT:    pshuflw {{.*#+}} xmm1 = xmm1[0,2,2,3,4,5,6,7]
-; SSE2-NEXT:    punpckldq {{.*#+}} xmm1 = xmm1[0],xmm0[0],xmm1[1],xmm0[1]
-; SSE2-NEXT:    pshufd {{.*#+}} xmm0 = xmm6[0,2,2,3]
-; SSE2-NEXT:    pshuflw {{.*#+}} xmm3 = xmm0[0,1,0,2,4,5,6,7]
-; SSE2-NEXT:    pshufd {{.*#+}} xmm0 = xmm2[0,2,2,3]
-; SSE2-NEXT:    pshuflw {{.*#+}} xmm0 = xmm0[0,1,0,2,4,5,6,7]
-; SSE2-NEXT:    punpckldq {{.*#+}} xmm0 = xmm0[0],xmm3[0],xmm0[1],xmm3[1]
-; SSE2-NEXT:    movsd {{.*#+}} xmm0 = xmm1[0],xmm0[1]
+; SSE2-NEXT:    pshufhw {{.*#+}} xmm0 = xmm0[0,1,2,3,4,6,6,7]
+; SSE2-NEXT:    pshufd {{.*#+}} xmm1 = xmm0[0,2,2,3]
+; SSE2-NEXT:    pshuflw {{.*#+}} xmm0 = xmm5[0,2,2,3,4,5,6,7]
+; SSE2-NEXT:    pshufhw {{.*#+}} xmm0 = xmm0[0,1,2,3,4,6,6,7]
+; SSE2-NEXT:    pshufd {{.*#+}} xmm0 = xmm0[0,2,2,3]
+; SSE2-NEXT:    punpcklqdq {{.*#+}} xmm0 = xmm0[0],xmm1[0]
 ; SSE2-NEXT:    retq
 ;
 ; SSSE3-LABEL: trunc_usat_v8i64_v8i16:
 ; SSSE3:       # %bb.0:
-; SSSE3-NEXT:    movdqa (%rdi), %xmm4
-; SSSE3-NEXT:    movdqa 16(%rdi), %xmm9
-; SSSE3-NEXT:    movdqa 32(%rdi), %xmm6
-; SSSE3-NEXT:    movdqa 48(%rdi), %xmm7
+; SSSE3-NEXT:    movdqa (%rdi), %xmm6
+; SSSE3-NEXT:    movdqa 16(%rdi), %xmm0
+; SSSE3-NEXT:    movdqa 32(%rdi), %xmm9
+; SSSE3-NEXT:    movdqa 48(%rdi), %xmm5
 ; SSSE3-NEXT:    movdqa {{.*#+}} xmm8 = [65535,65535]
 ; SSSE3-NEXT:    movdqa {{.*#+}} xmm3 = [9223372039002259456,9223372039002259456]
-; SSSE3-NEXT:    movdqa %xmm6, %xmm2
-; SSSE3-NEXT:    pxor %xmm3, %xmm2
+; SSSE3-NEXT:    movdqa %xmm0, %xmm7
+; SSSE3-NEXT:    pxor %xmm3, %xmm7
 ; SSSE3-NEXT:    movdqa {{.*#+}} xmm10 = [9223372039002324991,9223372039002324991]
-; SSSE3-NEXT:    movdqa %xmm10, %xmm5
-; SSSE3-NEXT:    pcmpgtd %xmm2, %xmm5
-; SSSE3-NEXT:    pshufd {{.*#+}} xmm0 = xmm5[0,0,2,2]
-; SSSE3-NEXT:    pcmpeqd %xmm10, %xmm2
-; SSSE3-NEXT:    pshufd {{.*#+}} xmm1 = xmm2[1,1,3,3]
-; SSSE3-NEXT:    pand %xmm0, %xmm1
-; SSSE3-NEXT:    pshufd {{.*#+}} xmm2 = xmm5[1,1,3,3]
-; SSSE3-NEXT:    por %xmm1, %xmm2
-; SSSE3-NEXT:    pand %xmm2, %xmm6
-; SSSE3-NEXT:    pandn %xmm8, %xmm2
-; SSSE3-NEXT:    por %xmm6, %xmm2
-; SSSE3-NEXT:    movdqa %xmm7, %xmm0
-; SSSE3-NEXT:    pxor %xmm3, %xmm0
-; SSSE3-NEXT:    movdqa %xmm10, %xmm1
-; SSSE3-NEXT:    pcmpgtd %xmm0, %xmm1
-; SSSE3-NEXT:    pshufd {{.*#+}} xmm5 = xmm1[0,0,2,2]
-; SSSE3-NEXT:    pcmpeqd %xmm10, %xmm0
-; SSSE3-NEXT:    pshufd {{.*#+}} xmm0 = xmm0[1,1,3,3]
-; SSSE3-NEXT:    pand %xmm5, %xmm0
-; SSSE3-NEXT:    pshufd {{.*#+}} xmm6 = xmm1[1,1,3,3]
-; SSSE3-NEXT:    por %xmm0, %xmm6
-; SSSE3-NEXT:    pand %xmm6, %xmm7
-; SSSE3-NEXT:    pandn %xmm8, %xmm6
-; SSSE3-NEXT:    por %xmm7, %xmm6
-; SSSE3-NEXT:    movdqa %xmm4, %xmm0
+; SSSE3-NEXT:    movdqa %xmm10, %xmm4
+; SSSE3-NEXT:    pcmpgtd %xmm7, %xmm4
+; SSSE3-NEXT:    pshufd {{.*#+}} xmm1 = xmm4[0,0,2,2]
+; SSSE3-NEXT:    pcmpeqd %xmm10, %xmm7
+; SSSE3-NEXT:    pshufd {{.*#+}} xmm7 = xmm7[1,1,3,3]
+; SSSE3-NEXT:    pand %xmm1, %xmm7
+; SSSE3-NEXT:    pshufd {{.*#+}} xmm1 = xmm4[1,1,3,3]
+; SSSE3-NEXT:    por %xmm7, %xmm1
+; SSSE3-NEXT:    pand %xmm1, %xmm0
+; SSSE3-NEXT:    pandn %xmm8, %xmm1
+; SSSE3-NEXT:    por %xmm0, %xmm1
+; SSSE3-NEXT:    movdqa %xmm6, %xmm0
 ; SSSE3-NEXT:    pxor %xmm3, %xmm0
-; SSSE3-NEXT:    movdqa %xmm10, %xmm1
-; SSSE3-NEXT:    pcmpgtd %xmm0, %xmm1
-; SSSE3-NEXT:    pshufd {{.*#+}} xmm5 = xmm1[0,0,2,2]
+; SSSE3-NEXT:    movdqa %xmm10, %xmm4
+; SSSE3-NEXT:    pcmpgtd %xmm0, %xmm4
+; SSSE3-NEXT:    pshufd {{.*#+}} xmm7 = xmm4[0,0,2,2]
 ; SSSE3-NEXT:    pcmpeqd %xmm10, %xmm0
-; SSSE3-NEXT:    pshufd {{.*#+}} xmm0 = xmm0[1,1,3,3]
-; SSSE3-NEXT:    pand %xmm5, %xmm0
+; SSSE3-NEXT:    pshufd {{.*#+}} xmm2 = xmm0[1,1,3,3]
+; SSSE3-NEXT:    pand %xmm7, %xmm2
+; SSSE3-NEXT:    pshufd {{.*#+}} xmm0 = xmm4[1,1,3,3]
+; SSSE3-NEXT:    por %xmm2, %xmm0
+; SSSE3-NEXT:    pand %xmm0, %xmm6
+; SSSE3-NEXT:    pandn %xmm8, %xmm0
+; SSSE3-NEXT:    por %xmm6, %xmm0
+; SSSE3-NEXT:    shufps {{.*#+}} xmm0 = xmm0[0,2],xmm1[0,2]
+; SSSE3-NEXT:    movdqa %xmm5, %xmm1
+; SSSE3-NEXT:    pxor %xmm3, %xmm1
+; SSSE3-NEXT:    movdqa %xmm10, %xmm2
+; SSSE3-NEXT:    pcmpgtd %xmm1, %xmm2
+; SSSE3-NEXT:    pshufd {{.*#+}} xmm4 = xmm2[0,0,2,2]
+; SSSE3-NEXT:    pcmpeqd %xmm10, %xmm1
 ; SSSE3-NEXT:    pshufd {{.*#+}} xmm1 = xmm1[1,1,3,3]
-; SSSE3-NEXT:    por %xmm0, %xmm1
-; SSSE3-NEXT:    pand %xmm1, %xmm4
-; SSSE3-NEXT:    pandn %xmm8, %xmm1
-; SSSE3-NEXT:    por %xmm4, %xmm1
+; SSSE3-NEXT:    pand %xmm4, %xmm1
+; SSSE3-NEXT:    pshufd {{.*#+}} xmm2 = xmm2[1,1,3,3]
+; SSSE3-NEXT:    por %xmm1, %xmm2
+; SSSE3-NEXT:    pand %xmm2, %xmm5
+; SSSE3-NEXT:    pandn %xmm8, %xmm2
+; SSSE3-NEXT:    por %xmm5, %xmm2
 ; SSSE3-NEXT:    pxor %xmm9, %xmm3
-; SSSE3-NEXT:    movdqa %xmm10, %xmm0
-; SSSE3-NEXT:    pcmpgtd %xmm3, %xmm0
-; SSSE3-NEXT:    pshufd {{.*#+}} xmm4 = xmm0[0,0,2,2]
+; SSSE3-NEXT:    movdqa %xmm10, %xmm1
+; SSSE3-NEXT:    pcmpgtd %xmm3, %xmm1
+; SSSE3-NEXT:    pshufd {{.*#+}} xmm4 = xmm1[0,0,2,2]
 ; SSSE3-NEXT:    pcmpeqd %xmm10, %xmm3
 ; SSSE3-NEXT:    pshufd {{.*#+}} xmm3 = xmm3[1,1,3,3]
 ; SSSE3-NEXT:    pand %xmm4, %xmm3
-; SSSE3-NEXT:    pshufd {{.*#+}} xmm0 = xmm0[1,1,3,3]
-; SSSE3-NEXT:    por %xmm3, %xmm0
-; SSSE3-NEXT:    pand %xmm0, %xmm9
-; SSSE3-NEXT:    pandn %xmm8, %xmm0
-; SSSE3-NEXT:    por %xmm9, %xmm0
-; SSSE3-NEXT:    pshufd {{.*#+}} xmm0 = xmm0[0,2,2,3]
-; SSSE3-NEXT:    pshuflw {{.*#+}} xmm0 = xmm0[0,2,2,3,4,5,6,7]
-; SSSE3-NEXT:    pshufd {{.*#+}} xmm1 = xmm1[0,2,2,3]
-; SSSE3-NEXT:    pshuflw {{.*#+}} xmm1 = xmm1[0,2,2,3,4,5,6,7]
-; SSSE3-NEXT:    punpckldq {{.*#+}} xmm1 = xmm1[0],xmm0[0],xmm1[1],xmm0[1]
-; SSSE3-NEXT:    pshufd {{.*#+}} xmm0 = xmm6[0,2,2,3]
-; SSSE3-NEXT:    pshuflw {{.*#+}} xmm3 = xmm0[0,1,0,2,4,5,6,7]
-; SSSE3-NEXT:    pshufd {{.*#+}} xmm0 = xmm2[0,2,2,3]
-; SSSE3-NEXT:    pshuflw {{.*#+}} xmm0 = xmm0[0,1,0,2,4,5,6,7]
-; SSSE3-NEXT:    punpckldq {{.*#+}} xmm0 = xmm0[0],xmm3[0],xmm0[1],xmm3[1]
-; SSSE3-NEXT:    movsd {{.*#+}} xmm0 = xmm1[0],xmm0[1]
+; SSSE3-NEXT:    pshufd {{.*#+}} xmm1 = xmm1[1,1,3,3]
+; SSSE3-NEXT:    por %xmm3, %xmm1
+; SSSE3-NEXT:    pand %xmm1, %xmm9
+; SSSE3-NEXT:    pandn %xmm8, %xmm1
+; SSSE3-NEXT:    por %xmm9, %xmm1
+; SSSE3-NEXT:    shufps {{.*#+}} xmm1 = xmm1[0,2],xmm2[0,2]
+; SSSE3-NEXT:    movdqa {{.*#+}} xmm2 = [0,1,4,5,8,9,12,13,8,9,12,13,12,13,14,15]
+; SSSE3-NEXT:    pshufb %xmm2, %xmm1
+; SSSE3-NEXT:    pshufb %xmm2, %xmm0
+; SSSE3-NEXT:    punpcklqdq {{.*#+}} xmm0 = xmm0[0],xmm1[0]
 ; SSSE3-NEXT:    retq
 ;
 ; SSE41-LABEL: trunc_usat_v8i64_v8i16:
diff --git a/llvm/test/CodeGen/X86/vector-trunc.ll b/llvm/test/CodeGen/X86/vector-trunc.ll
--- a/llvm/test/CodeGen/X86/vector-trunc.ll
+++ b/llvm/test/CodeGen/X86/vector-trunc.ll
@@ -136,32 +136,25 @@
 define <8 x i16> @trunc8i64_8i16(<8 x i64> %a) {
 ; SSE2-LABEL: trunc8i64_8i16:
 ; SSE2:       # %bb.0: # %entry
-; SSE2-NEXT:    pshufd {{.*#+}} xmm1 = xmm1[0,2,2,3]
-; SSE2-NEXT:    pshuflw {{.*#+}} xmm1 = xmm1[0,2,2,3,4,5,6,7]
+; SSE2-NEXT:    shufps {{.*#+}} xmm2 = xmm2[0,2],xmm3[0,2]
+; SSE2-NEXT:    pshuflw {{.*#+}} xmm2 = xmm2[0,2,2,3,4,5,6,7]
+; SSE2-NEXT:    pshufhw {{.*#+}} xmm2 = xmm2[0,1,2,3,4,6,6,7]
+; SSE2-NEXT:    pshufd {{.*#+}} xmm2 = xmm2[0,2,2,3]
+; SSE2-NEXT:    shufps {{.*#+}} xmm0 = xmm0[0,2],xmm1[0,2]
+; SSE2-NEXT:    pshuflw {{.*#+}} xmm0 = xmm0[0,2,2,3,4,5,6,7]
+; SSE2-NEXT:    pshufhw {{.*#+}} xmm0 = xmm0[0,1,2,3,4,6,6,7]
 ; SSE2-NEXT:    pshufd {{.*#+}} xmm0 = xmm0[0,2,2,3]
-; SSE2-NEXT:    pshuflw {{.*#+}} xmm4 = xmm0[0,2,2,3,4,5,6,7]
-; SSE2-NEXT:    punpckldq {{.*#+}} xmm4 = xmm4[0],xmm1[0],xmm4[1],xmm1[1]
-; SSE2-NEXT:    pshufd {{.*#+}} xmm0 = xmm3[0,2,2,3]
-; SSE2-NEXT:    pshuflw {{.*#+}} xmm1 = xmm0[0,1,0,2,4,5,6,7]
-; SSE2-NEXT:    pshufd {{.*#+}} xmm0 = xmm2[0,2,2,3]
-; SSE2-NEXT:    pshuflw {{.*#+}} xmm0 = xmm0[0,1,0,2,4,5,6,7]
-; SSE2-NEXT:    punpckldq {{.*#+}} xmm0 = xmm0[0],xmm1[0],xmm0[1],xmm1[1]
-; SSE2-NEXT:    movsd {{.*#+}} xmm0 = xmm4[0],xmm0[1]
+; SSE2-NEXT:    punpcklqdq {{.*#+}} xmm0 = xmm0[0],xmm2[0]
 ; SSE2-NEXT:    retq
 ;
 ; SSSE3-LABEL: trunc8i64_8i16:
 ; SSSE3:       # %bb.0: # %entry
-; SSSE3-NEXT:    pshufd {{.*#+}} xmm1 = xmm1[0,2,2,3]
-; SSSE3-NEXT:    pshuflw {{.*#+}} xmm1 = xmm1[0,2,2,3,4,5,6,7]
-; SSSE3-NEXT:    pshufd {{.*#+}} xmm0 = xmm0[0,2,2,3]
-; SSSE3-NEXT:    pshuflw {{.*#+}} xmm4 = xmm0[0,2,2,3,4,5,6,7]
-; SSSE3-NEXT:    punpckldq {{.*#+}} xmm4 = xmm4[0],xmm1[0],xmm4[1],xmm1[1]
-; SSSE3-NEXT:    pshufd {{.*#+}} xmm0 = xmm3[0,2,2,3]
-; SSSE3-NEXT:    pshuflw {{.*#+}} xmm1 = xmm0[0,1,0,2,4,5,6,7]
-; SSSE3-NEXT:    pshufd {{.*#+}} xmm0 = xmm2[0,2,2,3]
-; SSSE3-NEXT:    pshuflw {{.*#+}} xmm0 = xmm0[0,1,0,2,4,5,6,7]
-; SSSE3-NEXT:    punpckldq {{.*#+}} xmm0 = xmm0[0],xmm1[0],xmm0[1],xmm1[1]
-; SSSE3-NEXT:    movsd {{.*#+}} xmm0 = xmm4[0],xmm0[1]
+; SSSE3-NEXT:    shufps {{.*#+}} xmm2 = xmm2[0,2],xmm3[0,2]
+; SSSE3-NEXT:    movdqa {{.*#+}} xmm3 = [0,1,4,5,8,9,12,13,8,9,12,13,12,13,14,15]
+; SSSE3-NEXT:    pshufb %xmm3, %xmm2
+; SSSE3-NEXT:    shufps {{.*#+}} xmm0 = xmm0[0,2],xmm1[0,2]
+; SSSE3-NEXT:    pshufb %xmm3, %xmm0
+; SSSE3-NEXT:    punpcklqdq {{.*#+}} xmm0 = xmm0[0],xmm2[0]
 ; SSSE3-NEXT:    retq
 ;
 ; SSE41-LABEL: trunc8i64_8i16: