From: Craig Topper <craig.topper@gmail.com>
Date: Tue, 2 Aug 2016 05:11:15 +0000 (+0000)
Subject: [AVX-512] Correct ExeDomain for many AVX-512 instructions.
X-Git-Tag: llvmorg-4.0.0-rc1~13605
X-Git-Url: http://review.tizen.org/git/?a=commitdiff_plain;h=05948fb36c5f61554844ba3f31a0674a655549a6;p=platform%2Fupstream%2Fllvm.git

[AVX-512] Correct ExeDomain for many AVX-512 instructions.

llvm-svn: 277416
---

diff --git a/llvm/lib/Target/X86/X86InstrAVX512.td b/llvm/lib/Target/X86/X86InstrAVX512.td
index 98527cd..3830491 100644
--- a/llvm/lib/Target/X86/X86InstrAVX512.td
+++ b/llvm/lib/Target/X86/X86InstrAVX512.td
@@ -4224,6 +4224,7 @@ defm VPTESTNM  : avx512_vptest_all_forms<0x26, 0x27, "vptestnm", X86testnm>, T8X
 //===----------------------------------------------------------------------===//
 multiclass avx512_shift_rmi<bits<8> opc, Format ImmFormR, Format ImmFormM,
                          string OpcodeStr, SDNode OpNode, X86VectorVTInfo _> {
+  let ExeDomain = _.ExeDomain in {
   defm ri : AVX512_maskable<opc, ImmFormR, _, (outs _.RC:$dst),
                    (ins _.RC:$src1, u8imm:$src2), OpcodeStr,
                       "$src2, $src1", "$src1, $src2",
@@ -4235,10 +4236,12 @@ multiclass avx512_shift_rmi<bits<8> opc, Format ImmFormR, Format ImmFormM,
                    (_.VT (OpNode (_.VT (bitconvert (_.LdFrag addr:$src1))),
                           (i8 imm:$src2))),
                    SSE_INTSHIFT_ITINS_P.rm>;
+  }
 }
 
 multiclass avx512_shift_rmbi<bits<8> opc, Format ImmFormM,
                          string OpcodeStr, SDNode OpNode, X86VectorVTInfo _> {
+  let ExeDomain = _.ExeDomain in
   defm mbi : AVX512_maskable<opc, ImmFormM, _, (outs _.RC:$dst),
                    (ins _.ScalarMemOp:$src1, u8imm:$src2), OpcodeStr,
       "$src2, ${src1}"##_.BroadcastStr, "${src1}"##_.BroadcastStr##", $src2",
@@ -4249,6 +4252,7 @@ multiclass avx512_shift_rmbi<bits<8> opc, Format ImmFormM,
 multiclass avx512_shift_rrm<bits<8> opc, string OpcodeStr, SDNode OpNode,
                          ValueType SrcVT, PatFrag bc_frag, X86VectorVTInfo _> {
    // src2 is always 128-bit
+  let ExeDomain = _.ExeDomain in {
   defm rr : AVX512_maskable<opc, MRMSrcReg, _, (outs _.RC:$dst),
                    (ins _.RC:$src1, VR128X:$src2), OpcodeStr,
                       "$src2, $src1", "$src1, $src2",
@@ -4260,6 +4264,7 @@ multiclass avx512_shift_rrm<bits<8> opc, string OpcodeStr, SDNode OpNode,
                    (_.VT (OpNode _.RC:$src1, (bc_frag (loadv2i64 addr:$src2)))),
                    SSE_INTSHIFT_ITINS_P.rm>, AVX512BIBase,
                    EVEX_4V;
+  }
 }
 
 multiclass avx512_shift_sizes<bits<8> opc, string OpcodeStr, SDNode OpNode,
@@ -4353,6 +4358,7 @@ defm VPSRL : avx512_shift_types<0xD2, 0xD3, 0xD1, "vpsrl", X86vsrl>;
 //===-------------------------------------------------------------------===//
 multiclass avx512_var_shift<bits<8> opc, string OpcodeStr, SDNode OpNode,
                             X86VectorVTInfo _> {
+  let ExeDomain = _.ExeDomain in {
   defm rr : AVX512_maskable<opc, MRMSrcReg, _, (outs _.RC:$dst),
                    (ins _.RC:$src1, _.RC:$src2), OpcodeStr,
                       "$src2, $src1", "$src1, $src2",
@@ -4365,10 +4371,12 @@ multiclass avx512_var_shift<bits<8> opc, string OpcodeStr, SDNode OpNode,
                    (_.VT (bitconvert (_.LdFrag addr:$src2))))),
                    SSE_INTSHIFT_ITINS_P.rm>, AVX5128IBase, EVEX_4V,
                    EVEX_CD8<_.EltSize, CD8VF>;
+  }
 }
 
 multiclass avx512_var_shift_mb<bits<8> opc, string OpcodeStr, SDNode OpNode,
                             X86VectorVTInfo _> {
+  let ExeDomain = _.ExeDomain in
   defm rmb : AVX512_maskable<opc, MRMSrcMem, _, (outs _.RC:$dst),
                     (ins _.RC:$src1, _.ScalarMemOp:$src2), OpcodeStr,
                     "${src2}"##_.BroadcastStr##", $src1",
@@ -4635,8 +4643,10 @@ multiclass avx512_permil<string OpcodeStr, bits<8> OpcImm, bits<8> OpcVar,
                     EVEX, AVX512AIi8Base, EVEX_CD8<_.info128.EltSize, CD8VF>;
 }
 
+let ExeDomain = SSEPackedSingle in
 defm VPERMILPS : avx512_permil<"vpermilps", 0x04, 0x0C, avx512vl_f32_info,
                                avx512vl_i32_info>;
+let ExeDomain = SSEPackedDouble in
 defm VPERMILPD : avx512_permil<"vpermilpd", 0x05, 0x0D, avx512vl_f64_info,
                                avx512vl_i64_info>, VEX_W;
 //===----------------------------------------------------------------------===//
@@ -7196,7 +7206,8 @@ defm VEXPANDPD : expand_by_elt_width <0x88, "vexpandpd", avx512vl_f64_info>,
 //                               op(broadcast(eltVt),imm)
 //all instruction created with FROUND_CURRENT
 multiclass avx512_unary_fp_packed_imm<bits<8> opc, string OpcodeStr, SDNode OpNode,
-                                                            X86VectorVTInfo _>{
+                                      X86VectorVTInfo _>{
+  let ExeDomain = _.ExeDomain in {
   defm rri : AVX512_maskable<opc, MRMSrcReg, _, (outs _.RC:$dst),
                       (ins _.RC:$src1, i32u8imm:$src2),
                       OpcodeStr##_.Suffix, "$src2, $src1", "$src1, $src2",
@@ -7216,11 +7227,13 @@ multiclass avx512_unary_fp_packed_imm<bits<8> opc, string OpcodeStr, SDNode OpNo
                     (OpNode (_.VT (X86VBroadcast(_.ScalarLdFrag addr:$src1))),
                             (i32 imm:$src2),
                             (i32 FROUND_CURRENT))>, EVEX_B;
+  }
 }
 
 //handle instruction  reg_vec1 = op(reg_vec2,reg_vec3,imm),{sae}
 multiclass avx512_unary_fp_sae_packed_imm<bits<8> opc, string OpcodeStr,
                                              SDNode OpNode, X86VectorVTInfo _>{
+  let ExeDomain = _.ExeDomain in
   defm rrib : AVX512_maskable<opc, MRMSrcReg, _, (outs _.RC:$dst),
                       (ins _.RC:$src1, i32u8imm:$src2),
                       OpcodeStr##_.Suffix, "$src2, {sae}, $src1",
@@ -7250,7 +7263,8 @@ multiclass avx512_common_unary_fp_sae_packed_imm<string OpcodeStr,
 //                               op(reg_vec2,broadcast(eltVt),imm)
 //all instruction created with FROUND_CURRENT
 multiclass avx512_fp_packed_imm<bits<8> opc, string OpcodeStr, SDNode OpNode,
-                                                            X86VectorVTInfo _>{
+                                X86VectorVTInfo _>{
+  let ExeDomain = _.ExeDomain in {
   defm rri : AVX512_maskable<opc, MRMSrcReg, _, (outs _.RC:$dst),
                       (ins _.RC:$src1, _.RC:$src2, i32u8imm:$src3),
                       OpcodeStr, "$src3, $src2, $src1", "$src1, $src2, $src3",
@@ -7273,13 +7287,14 @@ multiclass avx512_fp_packed_imm<bits<8> opc, string OpcodeStr, SDNode OpNode,
                             (_.VT (X86VBroadcast(_.ScalarLdFrag addr:$src2))),
                             (i32 imm:$src3),
                             (i32 FROUND_CURRENT))>, EVEX_B;
+  }
 }
 
 //handle instruction  reg_vec1 = op(reg_vec2,reg_vec3,imm)
 //                               op(reg_vec2,mem_vec,imm)
 multiclass avx512_3Op_rm_imm8<bits<8> opc, string OpcodeStr, SDNode OpNode,
                              X86VectorVTInfo DestInfo, X86VectorVTInfo SrcInfo>{
-
+  let ExeDomain = DestInfo.ExeDomain in {
   defm rri : AVX512_maskable<opc, MRMSrcReg, DestInfo, (outs DestInfo.RC:$dst),
                   (ins SrcInfo.RC:$src1, SrcInfo.RC:$src2, u8imm:$src3),
                   OpcodeStr, "$src3, $src2, $src1", "$src1, $src2, $src3",
@@ -7293,6 +7308,7 @@ multiclass avx512_3Op_rm_imm8<bits<8> opc, string OpcodeStr, SDNode OpNode,
                              (SrcInfo.VT (bitconvert
                                                 (SrcInfo.LdFrag addr:$src2))),
                              (i8 imm:$src3)))>;
+  }
 }
 
 //handle instruction  reg_vec1 = op(reg_vec2,reg_vec3,imm)
@@ -7302,6 +7318,7 @@ multiclass avx512_3Op_imm8<bits<8> opc, string OpcodeStr, SDNode OpNode,
                            X86VectorVTInfo _>:
   avx512_3Op_rm_imm8<opc, OpcodeStr, OpNode, _, _>{
 
+  let ExeDomain = _.ExeDomain in
   defm rmbi : AVX512_maskable<opc, MRMSrcMem, _, (outs _.RC:$dst),
                     (ins _.RC:$src1, _.ScalarMemOp:$src2, u8imm:$src3),
                     OpcodeStr, "$src3, ${src2}"##_.BroadcastStr##", $src1",
@@ -7315,8 +7332,8 @@ multiclass avx512_3Op_imm8<bits<8> opc, string OpcodeStr, SDNode OpNode,
 //                                      op(reg_vec2,mem_scalar,imm)
 //all instruction created with FROUND_CURRENT
 multiclass avx512_fp_scalar_imm<bits<8> opc, string OpcodeStr, SDNode OpNode,
-                                                           X86VectorVTInfo _> {
-
+                                X86VectorVTInfo _> {
+  let ExeDomain = _.ExeDomain in {
   defm rri : AVX512_maskable_scalar<opc, MRMSrcReg, _, (outs _.RC:$dst),
                       (ins _.RC:$src1, _.RC:$src2, i32u8imm:$src3),
                       OpcodeStr, "$src3, $src2, $src1", "$src1, $src2, $src3",
@@ -7339,11 +7356,13 @@ multiclass avx512_fp_scalar_imm<bits<8> opc, string OpcodeStr, SDNode OpNode,
                     OpcodeStr, "$src3, $src2, $src1", "$src1, $src2, $src3",
                     []>;
   }
+  }
 }
 
 //handle instruction  reg_vec1 = op(reg_vec2,reg_vec3,imm),{sae}
 multiclass avx512_fp_sae_packed_imm<bits<8> opc, string OpcodeStr,
                                              SDNode OpNode, X86VectorVTInfo _>{
+  let ExeDomain = _.ExeDomain in
   defm rrib : AVX512_maskable<opc, MRMSrcReg, _, (outs _.RC:$dst),
                       (ins _.RC:$src1, _.RC:$src2, i32u8imm:$src3),
                       OpcodeStr, "$src3, {sae}, $src2, $src1",
@@ -7907,8 +7926,8 @@ defm VPSADBW : avx512_psadbw_packed_all<0xf6, X86psadbw, "vpsadbw",
                                        HasBWI>, EVEX_4V;
 
 multiclass avx512_ternlog<bits<8> opc, string OpcodeStr, SDNode OpNode,
-                                                            X86VectorVTInfo _>{
-  let Constraints = "$src1 = $dst" in {
+                          X86VectorVTInfo _>{
+  let Constraints = "$src1 = $dst", ExeDomain = _.ExeDomain in {
   defm rri : AVX512_maskable_3src<opc, MRMSrcReg, _, (outs _.RC:$dst),
                       (ins _.RC:$src2, _.RC:$src3, u8imm:$src4),
                       OpcodeStr, "$src4, $src3, $src2", "$src2, $src3, $src4",
@@ -7953,8 +7972,8 @@ defm VPTERNLOGQ : avx512_common_ternlog<"vpternlogq", avx512vl_i64_info>, VEX_W;
 //===----------------------------------------------------------------------===//
 
 multiclass avx512_fixupimm_packed<bits<8> opc, string OpcodeStr, SDNode OpNode,
-                                                            X86VectorVTInfo _>{
-  let Constraints = "$src1 = $dst" in {
+                                  X86VectorVTInfo _>{
+  let Constraints = "$src1 = $dst", ExeDomain = _.ExeDomain in {
     defm rri : AVX512_maskable_3src<opc, MRMSrcReg, _, (outs _.RC:$dst),
                         (ins _.RC:$src2, _.RC:$src3, i32u8imm:$src4),
                          OpcodeStr##_.Suffix, "$src4, $src3, $src2", "$src2, $src3, $src4",
@@ -7984,8 +8003,8 @@ multiclass avx512_fixupimm_packed<bits<8> opc, string OpcodeStr, SDNode OpNode,
 }
 
 multiclass avx512_fixupimm_packed_sae<bits<8> opc, string OpcodeStr,
-                                             SDNode OpNode, X86VectorVTInfo _>{
-let Constraints = "$src1 = $dst" in {
+                                      SDNode OpNode, X86VectorVTInfo _>{
+let Constraints = "$src1 = $dst", ExeDomain = _.ExeDomain in {
   defm rrib : AVX512_maskable_3src<opc, MRMSrcReg, _, (outs _.RC:$dst),
                       (ins _.RC:$src2, _.RC:$src3, i32u8imm:$src4),
                       OpcodeStr##_.Suffix, "$src4, {sae}, $src3, $src2",
@@ -8000,7 +8019,8 @@ let Constraints = "$src1 = $dst" in {
 
 multiclass avx512_fixupimm_scalar<bits<8> opc, string OpcodeStr, SDNode OpNode,
                                   X86VectorVTInfo _, X86VectorVTInfo _src3VT> {
-  let Constraints = "$src1 = $dst" , Predicates = [HasAVX512] in {
+  let Constraints = "$src1 = $dst" , Predicates = [HasAVX512],
+      ExeDomain = _.ExeDomain in {
     defm rri : AVX512_maskable_3src_scalar<opc, MRMSrcReg, _, (outs _.RC:$dst),
                       (ins _.RC:$src2, _.RC:$src3, i32u8imm:$src4),
                       OpcodeStr##_.Suffix, "$src4, $src3, $src2", "$src2, $src3, $src4",
diff --git a/llvm/lib/Target/X86/X86InstrFormats.td b/llvm/lib/Target/X86/X86InstrFormats.td
index 5183adc..8cbadd3 100644
--- a/llvm/lib/Target/X86/X86InstrFormats.td
+++ b/llvm/lib/Target/X86/X86InstrFormats.td
@@ -785,7 +785,6 @@ class AVX512AIi8<bits<8> o, Format F, dag outs, dag ins, string asm,
       : Ii8<o, F, outs, ins, asm, pattern, itin, SSEPackedInt>, TAPD,
         Requires<[HasAVX512]>;
 class AVX512AIi8Base : TAPD {
-  Domain ExeDomain = SSEPackedInt;
   ImmType ImmT = Imm8;
 }
 class AVX512Ii8<bits<8> o, Format F, dag outs, dag ins, string asm,
diff --git a/llvm/test/CodeGen/X86/avx512-intrinsics.ll b/llvm/test/CodeGen/X86/avx512-intrinsics.ll
index d0dfb62..63f50ef 100644
--- a/llvm/test/CodeGen/X86/avx512-intrinsics.ll
+++ b/llvm/test/CodeGen/X86/avx512-intrinsics.ll
@@ -4956,7 +4956,7 @@ define <2 x double>@test_int_x86_avx512_mask_getmant_sd(<2 x double> %x0, <2 x d
 ; CHECK:       ## BB#0:
 ; CHECK-NEXT:    andl $1, %edi
 ; CHECK-NEXT:    kmovw %edi, %k1
-; CHECK-NEXT:    vmovdqa64 %zmm2, %zmm3
+; CHECK-NEXT:    vmovapd %zmm2, %zmm3
 ; CHECK-NEXT:    vgetmantsd $11, %xmm1, %xmm0, %xmm3 {%k1}
 ; CHECK-NEXT:    vgetmantsd $11, %xmm1, %xmm0, %xmm4 {%k1} {z}
 ; CHECK-NEXT:    vgetmantsd $11, %xmm1, %xmm0, %xmm5
@@ -5995,7 +5995,7 @@ define <8 x double>@test_int_x86_avx512_mask_fixupimm_pd_512(<8 x double> %x0, <
 ; CHECK-LABEL: test_int_x86_avx512_mask_fixupimm_pd_512:
 ; CHECK:       ## BB#0:
 ; CHECK-NEXT:    kmovw %edi, %k1
-; CHECK-NEXT:    vmovdqa64 %zmm0, %zmm3
+; CHECK-NEXT:    vmovapd %zmm0, %zmm3
 ; CHECK-NEXT:    vfixupimmpd $4, %zmm2, %zmm1, %zmm3 {%k1}
 ; CHECK-NEXT:    vpxord %zmm4, %zmm4, %zmm4
 ; CHECK-NEXT:    vfixupimmpd $5, %zmm2, %zmm1, %zmm4 {%k1} {z}
@@ -6017,10 +6017,10 @@ define <8 x double>@test_int_x86_avx512_maskz_fixupimm_pd_512(<8 x double> %x0,
 ; CHECK-LABEL: test_int_x86_avx512_maskz_fixupimm_pd_512:
 ; CHECK:       ## BB#0:
 ; CHECK-NEXT:    kmovw %edi, %k1
-; CHECK-NEXT:    vmovdqa64 %zmm0, %zmm3
+; CHECK-NEXT:    vmovapd %zmm0, %zmm3
 ; CHECK-NEXT:    vfixupimmpd $3, %zmm2, %zmm1, %zmm3 {%k1} {z}
 ; CHECK-NEXT:    vpxord %zmm4, %zmm4, %zmm4
-; CHECK-NEXT:    vmovdqa64 %zmm0, %zmm5
+; CHECK-NEXT:    vmovapd %zmm0, %zmm5
 ; CHECK-NEXT:    vfixupimmpd $5, %zmm4, %zmm1, %zmm5 {%k1} {z}
 ; CHECK-NEXT:    vfixupimmpd $2, {sae}, %zmm2, %zmm1, %zmm0
 ; CHECK-NEXT:    vaddpd %zmm5, %zmm3, %zmm1
@@ -6041,10 +6041,10 @@ define <4 x float>@test_int_x86_avx512_mask_fixupimm_ss(<4 x float> %x0, <4 x fl
 ; CHECK:       ## BB#0:
 ; CHECK-NEXT:    andl $1, %edi
 ; CHECK-NEXT:    kmovw %edi, %k1
-; CHECK-NEXT:    vmovdqa64 %zmm0, %zmm3
+; CHECK-NEXT:    vmovaps %zmm0, %zmm3
 ; CHECK-NEXT:    vfixupimmss $5, %xmm2, %xmm1, %xmm3 {%k1}
-; CHECK-NEXT:    vpxor %xmm4, %xmm4, %xmm4
-; CHECK-NEXT:    vmovdqa64 %zmm0, %zmm5
+; CHECK-NEXT:    vxorps %xmm4, %xmm4, %xmm4
+; CHECK-NEXT:    vmovaps %zmm0, %zmm5
 ; CHECK-NEXT:    vfixupimmss $5, %xmm4, %xmm1, %xmm5 {%k1}
 ; CHECK-NEXT:    vfixupimmss $5, {sae}, %xmm2, %xmm1, %xmm0
 ; CHECK-NEXT:    vaddps %xmm5, %xmm3, %xmm1
@@ -6065,11 +6065,11 @@ define <4 x float>@test_int_x86_avx512_maskz_fixupimm_ss(<4 x float> %x0, <4 x f
 ; CHECK:       ## BB#0:
 ; CHECK-NEXT:    andl $1, %edi
 ; CHECK-NEXT:    kmovw %edi, %k1
-; CHECK-NEXT:    vmovdqa64 %zmm0, %zmm3
+; CHECK-NEXT:    vmovaps %zmm0, %zmm3
 ; CHECK-NEXT:    vfixupimmss $5, %xmm2, %xmm1, %xmm3 {%k1} {z}
-; CHECK-NEXT:    vmovdqa64 %zmm0, %zmm4
+; CHECK-NEXT:    vmovaps %zmm0, %zmm4
 ; CHECK-NEXT:    vfixupimmss $5, %xmm2, %xmm1, %xmm4
-; CHECK-NEXT:    vpxor %xmm2, %xmm2, %xmm2
+; CHECK-NEXT:    vxorps %xmm2, %xmm2, %xmm2
 ; CHECK-NEXT:    vfixupimmss $5, {sae}, %xmm2, %xmm1, %xmm0 {%k1} {z}
 ; CHECK-NEXT:    vaddps %xmm0, %xmm3, %xmm0
 ; CHECK-NEXT:    vaddps %xmm4, %xmm0, %xmm0
@@ -6088,10 +6088,10 @@ define <16 x float>@test_int_x86_avx512_mask_fixupimm_ps_512(<16 x float> %x0, <
 ; CHECK-LABEL: test_int_x86_avx512_mask_fixupimm_ps_512:
 ; CHECK:       ## BB#0:
 ; CHECK-NEXT:    kmovw %edi, %k1
-; CHECK-NEXT:    vmovdqa64 %zmm0, %zmm3
+; CHECK-NEXT:    vmovaps %zmm0, %zmm3
 ; CHECK-NEXT:    vfixupimmps $5, %zmm2, %zmm1, %zmm3 {%k1}
 ; CHECK-NEXT:    vpxord %zmm4, %zmm4, %zmm4
-; CHECK-NEXT:    vmovdqa64 %zmm0, %zmm5
+; CHECK-NEXT:    vmovaps %zmm0, %zmm5
 ; CHECK-NEXT:    vfixupimmps $5, %zmm4, %zmm1, %zmm5 {%k1}
 ; CHECK-NEXT:    vfixupimmps $5, {sae}, %zmm2, %zmm1, %zmm0
 ; CHECK-NEXT:    vaddps %zmm5, %zmm3, %zmm1
@@ -6111,9 +6111,9 @@ define <16 x float>@test_int_x86_avx512_maskz_fixupimm_ps_512(<16 x float> %x0,
 ; CHECK-LABEL: test_int_x86_avx512_maskz_fixupimm_ps_512:
 ; CHECK:       ## BB#0:
 ; CHECK-NEXT:    kmovw %edi, %k1
-; CHECK-NEXT:    vmovdqa64 %zmm0, %zmm3
+; CHECK-NEXT:    vmovaps %zmm0, %zmm3
 ; CHECK-NEXT:    vfixupimmps $5, %zmm2, %zmm1, %zmm3 {%k1} {z}
-; CHECK-NEXT:    vmovdqa64 %zmm0, %zmm4
+; CHECK-NEXT:    vmovaps %zmm0, %zmm4
 ; CHECK-NEXT:    vfixupimmps $5, %zmm2, %zmm1, %zmm4
 ; CHECK-NEXT:    vpxord %zmm2, %zmm2, %zmm2
 ; CHECK-NEXT:    vfixupimmps $5, {sae}, %zmm2, %zmm1, %zmm0 {%k1} {z}
@@ -6135,11 +6135,11 @@ define <2 x double>@test_int_x86_avx512_mask_fixupimm_sd(<2 x double> %x0, <2 x
 ; CHECK:       ## BB#0:
 ; CHECK-NEXT:    andl $1, %edi
 ; CHECK-NEXT:    kmovw %edi, %k1
-; CHECK-NEXT:    vmovdqa64 %zmm0, %zmm3
+; CHECK-NEXT:    vmovapd %zmm0, %zmm3
 ; CHECK-NEXT:    vfixupimmsd $5, %xmm2, %xmm1, %xmm3 {%k1}
-; CHECK-NEXT:    vmovdqa64 %zmm0, %zmm4
+; CHECK-NEXT:    vmovapd %zmm0, %zmm4
 ; CHECK-NEXT:    vfixupimmsd $5, %xmm2, %xmm1, %xmm4
-; CHECK-NEXT:    vpxor %xmm2, %xmm2, %xmm2
+; CHECK-NEXT:    vxorpd %xmm2, %xmm2, %xmm2
 ; CHECK-NEXT:    vfixupimmsd $5, {sae}, %xmm2, %xmm1, %xmm0 {%k1}
 ; CHECK-NEXT:    vaddpd %xmm0, %xmm3, %xmm0
 ; CHECK-NEXT:    vaddpd %xmm4, %xmm0, %xmm0
@@ -6159,10 +6159,10 @@ define <2 x double>@test_int_x86_avx512_maskz_fixupimm_sd(<2 x double> %x0, <2 x
 ; CHECK:       ## BB#0:
 ; CHECK-NEXT:    andl $1, %edi
 ; CHECK-NEXT:    kmovw %edi, %k1
-; CHECK-NEXT:    vmovdqa64 %zmm0, %zmm3
+; CHECK-NEXT:    vmovapd %zmm0, %zmm3
 ; CHECK-NEXT:    vfixupimmsd $5, %xmm2, %xmm1, %xmm3 {%k1} {z}
-; CHECK-NEXT:    vpxor %xmm4, %xmm4, %xmm4
-; CHECK-NEXT:    vmovdqa64 %zmm0, %zmm5
+; CHECK-NEXT:    vxorpd %xmm4, %xmm4, %xmm4
+; CHECK-NEXT:    vmovapd %zmm0, %zmm5
 ; CHECK-NEXT:    vfixupimmsd $5, {sae}, %xmm4, %xmm1, %xmm5 {%k1} {z}
 ; CHECK-NEXT:    vfixupimmsd $5, {sae}, %xmm2, %xmm1, %xmm0 {%k1} {z}
 ; CHECK-NEXT:    vaddpd %xmm5, %xmm3, %xmm1
diff --git a/llvm/test/CodeGen/X86/avx512vl-intrinsics.ll b/llvm/test/CodeGen/X86/avx512vl-intrinsics.ll
index 3079f0e..f0c7a3e 100644
--- a/llvm/test/CodeGen/X86/avx512vl-intrinsics.ll
+++ b/llvm/test/CodeGen/X86/avx512vl-intrinsics.ll
@@ -6889,7 +6889,7 @@ define <2 x double>@test_int_x86_avx512_mask_fixupimm_pd_128(<2 x double> %x0, <
 ; CHECK-LABEL: test_int_x86_avx512_mask_fixupimm_pd_128:
 ; CHECK:       ## BB#0:
 ; CHECK-NEXT:    kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
-; CHECK-NEXT:    vmovdqa64 %xmm0, %xmm3 ## encoding: [0x62,0xf1,0xfd,0x08,0x6f,0xd8]
+; CHECK-NEXT:    vmovapd %xmm0, %xmm3 ## encoding: [0x62,0xf1,0xfd,0x08,0x28,0xd8]
 ; CHECK-NEXT:    vfixupimmpd $5, %xmm2, %xmm1, %xmm3 {%k1} ## encoding: [0x62,0xf3,0xf5,0x09,0x54,0xda,0x05]
 ; CHECK-NEXT:    vpxord %xmm4, %xmm4, %xmm4 ## encoding: [0x62,0xf1,0x5d,0x08,0xef,0xe4]
 ; CHECK-NEXT:    vfixupimmpd $4, %xmm2, %xmm1, %xmm4 {%k1} {z} ## encoding: [0x62,0xf3,0xf5,0x89,0x54,0xe2,0x04]
@@ -6911,7 +6911,7 @@ define <2 x double>@test_int_x86_avx512_maskz_fixupimm_pd_128(<2 x double> %x0,
 ; CHECK-LABEL: test_int_x86_avx512_maskz_fixupimm_pd_128:
 ; CHECK:       ## BB#0:
 ; CHECK-NEXT:    kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
-; CHECK-NEXT:    vmovdqa64 %xmm0, %xmm3 ## encoding: [0x62,0xf1,0xfd,0x08,0x6f,0xd8]
+; CHECK-NEXT:    vmovapd %xmm0, %xmm3 ## encoding: [0x62,0xf1,0xfd,0x08,0x28,0xd8]
 ; CHECK-NEXT:    vfixupimmpd $5, %xmm2, %xmm1, %xmm3 {%k1} {z} ## encoding: [0x62,0xf3,0xf5,0x89,0x54,0xda,0x05]
 ; CHECK-NEXT:    vpxord %xmm2, %xmm2, %xmm2 ## encoding: [0x62,0xf1,0x6d,0x08,0xef,0xd2]
 ; CHECK-NEXT:    vfixupimmpd $3, %xmm2, %xmm1, %xmm0 {%k1} {z} ## encoding: [0x62,0xf3,0xf5,0x89,0x54,0xc2,0x03]
@@ -6931,7 +6931,7 @@ define <4 x double>@test_int_x86_avx512_mask_fixupimm_pd_256(<4 x double> %x0, <
 ; CHECK-LABEL: test_int_x86_avx512_mask_fixupimm_pd_256:
 ; CHECK:       ## BB#0:
 ; CHECK-NEXT:    kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
-; CHECK-NEXT:    vmovdqa64 %ymm0, %ymm3 ## encoding: [0x62,0xf1,0xfd,0x28,0x6f,0xd8]
+; CHECK-NEXT:    vmovapd %ymm0, %ymm3 ## encoding: [0x62,0xf1,0xfd,0x28,0x28,0xd8]
 ; CHECK-NEXT:    vfixupimmpd $4, %ymm2, %ymm1, %ymm3 {%k1} ## encoding: [0x62,0xf3,0xf5,0x29,0x54,0xda,0x04]
 ; CHECK-NEXT:    vpxord %ymm4, %ymm4, %ymm4 ## encoding: [0x62,0xf1,0x5d,0x28,0xef,0xe4]
 ; CHECK-NEXT:    vfixupimmpd $5, %ymm2, %ymm1, %ymm4 {%k1} {z} ## encoding: [0x62,0xf3,0xf5,0xa9,0x54,0xe2,0x05]
@@ -6953,10 +6953,10 @@ define <4 x double>@test_int_x86_avx512_maskz_fixupimm_pd_256(<4 x double> %x0,
 ; CHECK-LABEL: test_int_x86_avx512_maskz_fixupimm_pd_256:
 ; CHECK:       ## BB#0:
 ; CHECK-NEXT:    kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
-; CHECK-NEXT:    vmovdqa64 %ymm0, %ymm3 ## encoding: [0x62,0xf1,0xfd,0x28,0x6f,0xd8]
+; CHECK-NEXT:    vmovapd %ymm0, %ymm3 ## encoding: [0x62,0xf1,0xfd,0x28,0x28,0xd8]
 ; CHECK-NEXT:    vfixupimmpd $5, %ymm2, %ymm1, %ymm3 {%k1} {z} ## encoding: [0x62,0xf3,0xf5,0xa9,0x54,0xda,0x05]
 ; CHECK-NEXT:    vpxord %ymm4, %ymm4, %ymm4 ## encoding: [0x62,0xf1,0x5d,0x28,0xef,0xe4]
-; CHECK-NEXT:    vmovdqa64 %ymm0, %ymm5 ## encoding: [0x62,0xf1,0xfd,0x28,0x6f,0xe8]
+; CHECK-NEXT:    vmovapd %ymm0, %ymm5 ## encoding: [0x62,0xf1,0xfd,0x28,0x28,0xe8]
 ; CHECK-NEXT:    vfixupimmpd $4, %ymm4, %ymm1, %ymm5 {%k1} {z} ## encoding: [0x62,0xf3,0xf5,0xa9,0x54,0xec,0x04]
 ; CHECK-NEXT:    vfixupimmpd $3, %ymm2, %ymm1, %ymm0 ## encoding: [0x62,0xf3,0xf5,0x28,0x54,0xc2,0x03]
 ; CHECK-NEXT:    vaddpd %ymm5, %ymm3, %ymm1 ## encoding: [0x62,0xf1,0xe5,0x28,0x58,0xcd]
@@ -6976,9 +6976,9 @@ define <4 x float>@test_int_x86_avx512_mask_fixupimm_ps_128(<4 x float> %x0, <4
 ; CHECK-LABEL: test_int_x86_avx512_mask_fixupimm_ps_128:
 ; CHECK:       ## BB#0:
 ; CHECK-NEXT:    kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
-; CHECK-NEXT:    vmovdqa64 %xmm0, %xmm3 ## encoding: [0x62,0xf1,0xfd,0x08,0x6f,0xd8]
+; CHECK-NEXT:    vmovaps %xmm0, %xmm3 ## encoding: [0x62,0xf1,0x7c,0x08,0x28,0xd8]
 ; CHECK-NEXT:    vfixupimmps $5, %xmm2, %xmm1, %xmm3 {%k1} ## encoding: [0x62,0xf3,0x75,0x09,0x54,0xda,0x05]
-; CHECK-NEXT:    vmovdqa64 %xmm0, %xmm4 ## encoding: [0x62,0xf1,0xfd,0x08,0x6f,0xe0]
+; CHECK-NEXT:    vmovaps %xmm0, %xmm4 ## encoding: [0x62,0xf1,0x7c,0x08,0x28,0xe0]
 ; CHECK-NEXT:    vfixupimmps $5, %xmm2, %xmm1, %xmm4 ## encoding: [0x62,0xf3,0x75,0x08,0x54,0xe2,0x05]
 ; CHECK-NEXT:    vpxord %xmm2, %xmm2, %xmm2 ## encoding: [0x62,0xf1,0x6d,0x08,0xef,0xd2]
 ; CHECK-NEXT:    vfixupimmps $5, %xmm2, %xmm1, %xmm0 {%k1} ## encoding: [0x62,0xf3,0x75,0x09,0x54,0xc2,0x05]
@@ -6999,9 +6999,9 @@ define <4 x float>@test_int_x86_avx512_maskz_fixupimm_ps_128(<4 x float> %x0, <4
 ; CHECK-LABEL: test_int_x86_avx512_maskz_fixupimm_ps_128:
 ; CHECK:       ## BB#0:
 ; CHECK-NEXT:    kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
-; CHECK-NEXT:    vmovdqa64 %xmm0, %xmm3 ## encoding: [0x62,0xf1,0xfd,0x08,0x6f,0xd8]
+; CHECK-NEXT:    vmovaps %xmm0, %xmm3 ## encoding: [0x62,0xf1,0x7c,0x08,0x28,0xd8]
 ; CHECK-NEXT:    vfixupimmps $5, %xmm2, %xmm1, %xmm3 {%k1} {z} ## encoding: [0x62,0xf3,0x75,0x89,0x54,0xda,0x05]
-; CHECK-NEXT:    vmovdqa64 %xmm0, %xmm4 ## encoding: [0x62,0xf1,0xfd,0x08,0x6f,0xe0]
+; CHECK-NEXT:    vmovaps %xmm0, %xmm4 ## encoding: [0x62,0xf1,0x7c,0x08,0x28,0xe0]
 ; CHECK-NEXT:    vfixupimmps $5, %xmm2, %xmm1, %xmm4 ## encoding: [0x62,0xf3,0x75,0x08,0x54,0xe2,0x05]
 ; CHECK-NEXT:    vpxord %xmm2, %xmm2, %xmm2 ## encoding: [0x62,0xf1,0x6d,0x08,0xef,0xd2]
 ; CHECK-NEXT:    vfixupimmps $5, %xmm2, %xmm1, %xmm0 {%k1} {z} ## encoding: [0x62,0xf3,0x75,0x89,0x54,0xc2,0x05]
@@ -7022,9 +7022,9 @@ define <8 x float>@test_int_x86_avx512_mask_fixupimm_ps_256(<8 x float> %x0, <8
 ; CHECK-LABEL: test_int_x86_avx512_mask_fixupimm_ps_256:
 ; CHECK:       ## BB#0:
 ; CHECK-NEXT:    kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
-; CHECK-NEXT:    vmovdqa64 %ymm0, %ymm3 ## encoding: [0x62,0xf1,0xfd,0x28,0x6f,0xd8]
+; CHECK-NEXT:    vmovaps %ymm0, %ymm3 ## encoding: [0x62,0xf1,0x7c,0x28,0x28,0xd8]
 ; CHECK-NEXT:    vfixupimmps $5, %ymm2, %ymm1, %ymm3 {%k1} ## encoding: [0x62,0xf3,0x75,0x29,0x54,0xda,0x05]
-; CHECK-NEXT:    vmovdqa64 %ymm0, %ymm4 ## encoding: [0x62,0xf1,0xfd,0x28,0x6f,0xe0]
+; CHECK-NEXT:    vmovaps %ymm0, %ymm4 ## encoding: [0x62,0xf1,0x7c,0x28,0x28,0xe0]
 ; CHECK-NEXT:    vfixupimmps $5, %ymm2, %ymm1, %ymm4 ## encoding: [0x62,0xf3,0x75,0x28,0x54,0xe2,0x05]
 ; CHECK-NEXT:    vpxord %ymm2, %ymm2, %ymm2 ## encoding: [0x62,0xf1,0x6d,0x28,0xef,0xd2]
 ; CHECK-NEXT:    vfixupimmps $5, %ymm2, %ymm1, %ymm0 {%k1} ## encoding: [0x62,0xf3,0x75,0x29,0x54,0xc2,0x05]
@@ -7045,9 +7045,9 @@ define <8 x float>@test_int_x86_avx512_maskz_fixupimm_ps_256(<8 x float> %x0, <8
 ; CHECK-LABEL: test_int_x86_avx512_maskz_fixupimm_ps_256:
 ; CHECK:       ## BB#0:
 ; CHECK-NEXT:    kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
-; CHECK-NEXT:    vmovdqa64 %ymm0, %ymm3 ## encoding: [0x62,0xf1,0xfd,0x28,0x6f,0xd8]
+; CHECK-NEXT:    vmovaps %ymm0, %ymm3 ## encoding: [0x62,0xf1,0x7c,0x28,0x28,0xd8]
 ; CHECK-NEXT:    vfixupimmps $5, %ymm2, %ymm1, %ymm3 {%k1} {z} ## encoding: [0x62,0xf3,0x75,0xa9,0x54,0xda,0x05]
-; CHECK-NEXT:    vmovdqa64 %ymm0, %ymm4 ## encoding: [0x62,0xf1,0xfd,0x28,0x6f,0xe0]
+; CHECK-NEXT:    vmovaps %ymm0, %ymm4 ## encoding: [0x62,0xf1,0x7c,0x28,0x28,0xe0]
 ; CHECK-NEXT:    vfixupimmps $5, %ymm2, %ymm1, %ymm4 ## encoding: [0x62,0xf3,0x75,0x28,0x54,0xe2,0x05]
 ; CHECK-NEXT:    vpxord %ymm2, %ymm2, %ymm2 ## encoding: [0x62,0xf1,0x6d,0x28,0xef,0xd2]
 ; CHECK-NEXT:    vfixupimmps $5, %ymm2, %ymm1, %ymm0 {%k1} {z} ## encoding: [0x62,0xf3,0x75,0xa9,0x54,0xc2,0x05]
diff --git a/llvm/test/CodeGen/X86/vector-half-conversions.ll b/llvm/test/CodeGen/X86/vector-half-conversions.ll
index aeb93a2..6fc6395 100644
--- a/llvm/test/CodeGen/X86/vector-half-conversions.ll
+++ b/llvm/test/CodeGen/X86/vector-half-conversions.ll
@@ -2997,12 +2997,12 @@ define <4 x i16> @cvt_4f32_to_4i16(<4 x float> %a0) nounwind {
 ; AVX512VL-NEXT:    callq __gnu_f2h_ieee
 ; AVX512VL-NEXT:    movzwl %ax, %r14d
 ; AVX512VL-NEXT:    orl %ebx, %r14d
-; AVX512VL-NEXT:    vmovdqa64 (%rsp), %xmm0 # 16-byte Reload
+; AVX512VL-NEXT:    vmovaps (%rsp), %xmm0 # 16-byte Reload
 ; AVX512VL-NEXT:    vpermilps {{.*#+}} xmm0 = xmm0[3,1,2,3]
 ; AVX512VL-NEXT:    callq __gnu_f2h_ieee
 ; AVX512VL-NEXT:    movw %ax, %bx
 ; AVX512VL-NEXT:    shll $16, %ebx
-; AVX512VL-NEXT:    vmovdqa64 (%rsp), %xmm0 # 16-byte Reload
+; AVX512VL-NEXT:    vmovapd (%rsp), %xmm0 # 16-byte Reload
 ; AVX512VL-NEXT:    vpermilpd {{.*#+}} xmm0 = xmm0[1,0]
 ; AVX512VL-NEXT:    callq __gnu_f2h_ieee
 ; AVX512VL-NEXT:    movzwl %ax, %eax
@@ -3109,12 +3109,12 @@ define <8 x i16> @cvt_4f32_to_8i16_undef(<4 x float> %a0) nounwind {
 ; AVX512VL-NEXT:    callq __gnu_f2h_ieee
 ; AVX512VL-NEXT:    movzwl %ax, %r14d
 ; AVX512VL-NEXT:    orl %ebx, %r14d
-; AVX512VL-NEXT:    vmovdqa64 (%rsp), %xmm0 # 16-byte Reload
+; AVX512VL-NEXT:    vmovaps (%rsp), %xmm0 # 16-byte Reload
 ; AVX512VL-NEXT:    vpermilps {{.*#+}} xmm0 = xmm0[3,1,2,3]
 ; AVX512VL-NEXT:    callq __gnu_f2h_ieee
 ; AVX512VL-NEXT:    movw %ax, %bx
 ; AVX512VL-NEXT:    shll $16, %ebx
-; AVX512VL-NEXT:    vmovdqa64 (%rsp), %xmm0 # 16-byte Reload
+; AVX512VL-NEXT:    vmovapd (%rsp), %xmm0 # 16-byte Reload
 ; AVX512VL-NEXT:    vpermilpd {{.*#+}} xmm0 = xmm0[1,0]
 ; AVX512VL-NEXT:    callq __gnu_f2h_ieee
 ; AVX512VL-NEXT:    movzwl %ax, %eax
@@ -3225,12 +3225,12 @@ define <8 x i16> @cvt_4f32_to_8i16_zero(<4 x float> %a0) nounwind {
 ; AVX512VL-NEXT:    callq __gnu_f2h_ieee
 ; AVX512VL-NEXT:    movzwl %ax, %r14d
 ; AVX512VL-NEXT:    orl %ebx, %r14d
-; AVX512VL-NEXT:    vmovdqa64 (%rsp), %xmm0 # 16-byte Reload
+; AVX512VL-NEXT:    vmovaps (%rsp), %xmm0 # 16-byte Reload
 ; AVX512VL-NEXT:    vpermilps {{.*#+}} xmm0 = xmm0[3,1,2,3]
 ; AVX512VL-NEXT:    callq __gnu_f2h_ieee
 ; AVX512VL-NEXT:    movw %ax, %bx
 ; AVX512VL-NEXT:    shll $16, %ebx
-; AVX512VL-NEXT:    vmovdqa64 (%rsp), %xmm0 # 16-byte Reload
+; AVX512VL-NEXT:    vmovapd (%rsp), %xmm0 # 16-byte Reload
 ; AVX512VL-NEXT:    vpermilpd {{.*#+}} xmm0 = xmm0[1,0]
 ; AVX512VL-NEXT:    callq __gnu_f2h_ieee
 ; AVX512VL-NEXT:    movzwl %ax, %eax
@@ -3410,12 +3410,12 @@ define <8 x i16> @cvt_8f32_to_8i16(<8 x float> %a0) nounwind {
 ; AVX512VL-NEXT:    callq __gnu_f2h_ieee
 ; AVX512VL-NEXT:    movzwl %ax, %r15d
 ; AVX512VL-NEXT:    orl %ebx, %r15d
-; AVX512VL-NEXT:    vmovdqu64 (%rsp), %ymm0 # 32-byte Reload
+; AVX512VL-NEXT:    vmovups (%rsp), %ymm0 # 32-byte Reload
 ; AVX512VL-NEXT:    vpermilps {{.*#+}} xmm0 = xmm0[3,1,2,3]
 ; AVX512VL-NEXT:    callq __gnu_f2h_ieee
 ; AVX512VL-NEXT:    movw %ax, %bx
 ; AVX512VL-NEXT:    shll $16, %ebx
-; AVX512VL-NEXT:    vmovdqu64 (%rsp), %ymm0 # 32-byte Reload
+; AVX512VL-NEXT:    vmovupd (%rsp), %ymm0 # 32-byte Reload
 ; AVX512VL-NEXT:    vpermilpd {{.*#+}} xmm0 = xmm0[1,0]
 ; AVX512VL-NEXT:    callq __gnu_f2h_ieee
 ; AVX512VL-NEXT:    movzwl %ax, %r14d
@@ -3433,12 +3433,12 @@ define <8 x i16> @cvt_8f32_to_8i16(<8 x float> %a0) nounwind {
 ; AVX512VL-NEXT:    callq __gnu_f2h_ieee
 ; AVX512VL-NEXT:    movzwl %ax, %r15d
 ; AVX512VL-NEXT:    orl %ebx, %r15d
-; AVX512VL-NEXT:    vmovdqa64 (%rsp), %xmm0 # 16-byte Reload
+; AVX512VL-NEXT:    vmovaps (%rsp), %xmm0 # 16-byte Reload
 ; AVX512VL-NEXT:    vpermilps {{.*#+}} xmm0 = xmm0[3,1,2,3]
 ; AVX512VL-NEXT:    callq __gnu_f2h_ieee
 ; AVX512VL-NEXT:    movw %ax, %bx
 ; AVX512VL-NEXT:    shll $16, %ebx
-; AVX512VL-NEXT:    vmovdqa64 (%rsp), %xmm0 # 16-byte Reload
+; AVX512VL-NEXT:    vmovapd (%rsp), %xmm0 # 16-byte Reload
 ; AVX512VL-NEXT:    vpermilpd {{.*#+}} xmm0 = xmm0[1,0]
 ; AVX512VL-NEXT:    callq __gnu_f2h_ieee
 ; AVX512VL-NEXT:    movzwl %ax, %eax
@@ -3677,7 +3677,7 @@ define <16 x i16> @cvt_16f32_to_16i16(<16 x float> %a0) nounwind {
 ; AVX512VL-NEXT:    callq __gnu_f2h_ieee
 ; AVX512VL-NEXT:    # kill: %AX<def> %AX<kill> %EAX<def>
 ; AVX512VL-NEXT:    movl %eax, {{[0-9]+}}(%rsp) # 4-byte Spill
-; AVX512VL-NEXT:    vmovdqa64 (%rsp), %xmm0 # 16-byte Reload
+; AVX512VL-NEXT:    vmovapd (%rsp), %xmm0 # 16-byte Reload
 ; AVX512VL-NEXT:    vpermilpd {{.*#+}} xmm0 = xmm0[1,0]
 ; AVX512VL-NEXT:    callq __gnu_f2h_ieee
 ; AVX512VL-NEXT:    # kill: %AX<def> %AX<kill> %EAX<def>
@@ -3691,12 +3691,12 @@ define <16 x i16> @cvt_16f32_to_16i16(<16 x float> %a0) nounwind {
 ; AVX512VL-NEXT:    callq __gnu_f2h_ieee
 ; AVX512VL-NEXT:    # kill: %AX<def> %AX<kill> %EAX<def>
 ; AVX512VL-NEXT:    movl %eax, {{[0-9]+}}(%rsp) # 4-byte Spill
-; AVX512VL-NEXT:    vmovdqu64 {{[0-9]+}}(%rsp), %zmm0 # 64-byte Reload
+; AVX512VL-NEXT:    vmovups {{[0-9]+}}(%rsp), %zmm0 # 64-byte Reload
 ; AVX512VL-NEXT:    vpermilps {{.*#+}} xmm0 = xmm0[3,1,2,3]
 ; AVX512VL-NEXT:    callq __gnu_f2h_ieee
 ; AVX512VL-NEXT:    # kill: %AX<def> %AX<kill> %EAX<def>
 ; AVX512VL-NEXT:    movl %eax, {{[0-9]+}}(%rsp) # 4-byte Spill
-; AVX512VL-NEXT:    vmovdqu64 {{[0-9]+}}(%rsp), %zmm0 # 64-byte Reload
+; AVX512VL-NEXT:    vmovupd {{[0-9]+}}(%rsp), %zmm0 # 64-byte Reload
 ; AVX512VL-NEXT:    vpermilpd {{.*#+}} xmm0 = xmm0[1,0]
 ; AVX512VL-NEXT:    callq __gnu_f2h_ieee
 ; AVX512VL-NEXT:    # kill: %AX<def> %AX<kill> %EAX<def>
@@ -3720,7 +3720,7 @@ define <16 x i16> @cvt_16f32_to_16i16(<16 x float> %a0) nounwind {
 ; AVX512VL-NEXT:    callq __gnu_f2h_ieee
 ; AVX512VL-NEXT:    # kill: %AX<def> %AX<kill> %EAX<def>
 ; AVX512VL-NEXT:    movl %eax, {{[0-9]+}}(%rsp) # 4-byte Spill
-; AVX512VL-NEXT:    vmovdqa64 (%rsp), %xmm0 # 16-byte Reload
+; AVX512VL-NEXT:    vmovapd (%rsp), %xmm0 # 16-byte Reload
 ; AVX512VL-NEXT:    vpermilpd {{.*#+}} xmm0 = xmm0[1,0]
 ; AVX512VL-NEXT:    callq __gnu_f2h_ieee
 ; AVX512VL-NEXT:    movw %ax, %r13w
@@ -3731,11 +3731,11 @@ define <16 x i16> @cvt_16f32_to_16i16(<16 x float> %a0) nounwind {
 ; AVX512VL-NEXT:    vmovaps (%rsp), %xmm0 # 16-byte Reload
 ; AVX512VL-NEXT:    callq __gnu_f2h_ieee
 ; AVX512VL-NEXT:    movw %ax, %bp
-; AVX512VL-NEXT:    vmovdqu64 {{[0-9]+}}(%rsp), %ymm0 # 32-byte Reload
+; AVX512VL-NEXT:    vmovups {{[0-9]+}}(%rsp), %ymm0 # 32-byte Reload
 ; AVX512VL-NEXT:    vpermilps {{.*#+}} xmm0 = xmm0[3,1,2,3]
 ; AVX512VL-NEXT:    callq __gnu_f2h_ieee
 ; AVX512VL-NEXT:    movw %ax, %r14w
-; AVX512VL-NEXT:    vmovdqu64 {{[0-9]+}}(%rsp), %ymm0 # 32-byte Reload
+; AVX512VL-NEXT:    vmovupd {{[0-9]+}}(%rsp), %ymm0 # 32-byte Reload
 ; AVX512VL-NEXT:    vpermilpd {{.*#+}} xmm0 = xmm0[1,0]
 ; AVX512VL-NEXT:    callq __gnu_f2h_ieee
 ; AVX512VL-NEXT:    movw %ax, %r15w
@@ -3888,11 +3888,11 @@ define void @store_cvt_4f32_to_4i16(<4 x float> %a0, <4 x i16>* %a1) nounwind {
 ; AVX512VL-NEXT:    vmovshdup {{.*#+}} xmm0 = xmm0[1,1,3,3]
 ; AVX512VL-NEXT:    callq __gnu_f2h_ieee
 ; AVX512VL-NEXT:    movl %eax, %r14d
-; AVX512VL-NEXT:    vmovdqa64 (%rsp), %xmm0 # 16-byte Reload
+; AVX512VL-NEXT:    vmovapd (%rsp), %xmm0 # 16-byte Reload
 ; AVX512VL-NEXT:    vpermilpd {{.*#+}} xmm0 = xmm0[1,0]
 ; AVX512VL-NEXT:    callq __gnu_f2h_ieee
 ; AVX512VL-NEXT:    movl %eax, %r15d
-; AVX512VL-NEXT:    vmovdqa64 (%rsp), %xmm0 # 16-byte Reload
+; AVX512VL-NEXT:    vmovaps (%rsp), %xmm0 # 16-byte Reload
 ; AVX512VL-NEXT:    vpermilps {{.*#+}} xmm0 = xmm0[3,1,2,3]
 ; AVX512VL-NEXT:    callq __gnu_f2h_ieee
 ; AVX512VL-NEXT:    movl %eax, %ebp
@@ -4009,12 +4009,12 @@ define void @store_cvt_4f32_to_8i16_undef(<4 x float> %a0, <8 x i16>* %a1) nounw
 ; AVX512VL-NEXT:    callq __gnu_f2h_ieee
 ; AVX512VL-NEXT:    movzwl %ax, %ebx
 ; AVX512VL-NEXT:    orl %ebp, %ebx
-; AVX512VL-NEXT:    vmovdqa64 (%rsp), %xmm0 # 16-byte Reload
+; AVX512VL-NEXT:    vmovaps (%rsp), %xmm0 # 16-byte Reload
 ; AVX512VL-NEXT:    vpermilps {{.*#+}} xmm0 = xmm0[3,1,2,3]
 ; AVX512VL-NEXT:    callq __gnu_f2h_ieee
 ; AVX512VL-NEXT:    movw %ax, %bp
 ; AVX512VL-NEXT:    shll $16, %ebp
-; AVX512VL-NEXT:    vmovdqa64 (%rsp), %xmm0 # 16-byte Reload
+; AVX512VL-NEXT:    vmovapd (%rsp), %xmm0 # 16-byte Reload
 ; AVX512VL-NEXT:    vpermilpd {{.*#+}} xmm0 = xmm0[1,0]
 ; AVX512VL-NEXT:    callq __gnu_f2h_ieee
 ; AVX512VL-NEXT:    movzwl %ax, %eax
@@ -4133,12 +4133,12 @@ define void @store_cvt_4f32_to_8i16_zero(<4 x float> %a0, <8 x i16>* %a1) nounwi
 ; AVX512VL-NEXT:    callq __gnu_f2h_ieee
 ; AVX512VL-NEXT:    movzwl %ax, %ebx
 ; AVX512VL-NEXT:    orl %ebp, %ebx
-; AVX512VL-NEXT:    vmovdqa64 (%rsp), %xmm0 # 16-byte Reload
+; AVX512VL-NEXT:    vmovaps (%rsp), %xmm0 # 16-byte Reload
 ; AVX512VL-NEXT:    vpermilps {{.*#+}} xmm0 = xmm0[3,1,2,3]
 ; AVX512VL-NEXT:    callq __gnu_f2h_ieee
 ; AVX512VL-NEXT:    movw %ax, %bp
 ; AVX512VL-NEXT:    shll $16, %ebp
-; AVX512VL-NEXT:    vmovdqa64 (%rsp), %xmm0 # 16-byte Reload
+; AVX512VL-NEXT:    vmovapd (%rsp), %xmm0 # 16-byte Reload
 ; AVX512VL-NEXT:    vpermilpd {{.*#+}} xmm0 = xmm0[1,0]
 ; AVX512VL-NEXT:    callq __gnu_f2h_ieee
 ; AVX512VL-NEXT:    movzwl %ax, %eax
@@ -4286,11 +4286,11 @@ define void @store_cvt_8f32_to_8i16(<8 x float> %a0, <8 x i16>* %a1) nounwind {
 ; AVX512VL-NEXT:    vmovshdup {{.*#+}} xmm0 = xmm0[1,1,3,3]
 ; AVX512VL-NEXT:    callq __gnu_f2h_ieee
 ; AVX512VL-NEXT:    movw %ax, {{[0-9]+}}(%rsp) # 2-byte Spill
-; AVX512VL-NEXT:    vmovdqu64 {{[0-9]+}}(%rsp), %ymm0 # 32-byte Reload
+; AVX512VL-NEXT:    vmovupd {{[0-9]+}}(%rsp), %ymm0 # 32-byte Reload
 ; AVX512VL-NEXT:    vpermilpd {{.*#+}} xmm0 = xmm0[1,0]
 ; AVX512VL-NEXT:    callq __gnu_f2h_ieee
 ; AVX512VL-NEXT:    movw %ax, {{[0-9]+}}(%rsp) # 2-byte Spill
-; AVX512VL-NEXT:    vmovdqu64 {{[0-9]+}}(%rsp), %ymm0 # 32-byte Reload
+; AVX512VL-NEXT:    vmovups {{[0-9]+}}(%rsp), %ymm0 # 32-byte Reload
 ; AVX512VL-NEXT:    vpermilps {{.*#+}} xmm0 = xmm0[3,1,2,3]
 ; AVX512VL-NEXT:    callq __gnu_f2h_ieee
 ; AVX512VL-NEXT:    movl %eax, %r12d
@@ -4300,11 +4300,11 @@ define void @store_cvt_8f32_to_8i16(<8 x float> %a0, <8 x i16>* %a1) nounwind {
 ; AVX512VL-NEXT:    vmovshdup {{.*#+}} xmm0 = xmm0[1,1,3,3]
 ; AVX512VL-NEXT:    callq __gnu_f2h_ieee
 ; AVX512VL-NEXT:    movl %eax, %r13d
-; AVX512VL-NEXT:    vmovdqa64 {{[0-9]+}}(%rsp), %xmm0 # 16-byte Reload
+; AVX512VL-NEXT:    vmovapd {{[0-9]+}}(%rsp), %xmm0 # 16-byte Reload
 ; AVX512VL-NEXT:    vpermilpd {{.*#+}} xmm0 = xmm0[1,0]
 ; AVX512VL-NEXT:    callq __gnu_f2h_ieee
 ; AVX512VL-NEXT:    movl %eax, %ebp
-; AVX512VL-NEXT:    vmovdqa64 {{[0-9]+}}(%rsp), %xmm0 # 16-byte Reload
+; AVX512VL-NEXT:    vmovaps {{[0-9]+}}(%rsp), %xmm0 # 16-byte Reload
 ; AVX512VL-NEXT:    vpermilps {{.*#+}} xmm0 = xmm0[3,1,2,3]
 ; AVX512VL-NEXT:    callq __gnu_f2h_ieee
 ; AVX512VL-NEXT:    movl %eax, %r14d
@@ -4554,11 +4554,11 @@ define void @store_cvt_16f32_to_16i16(<16 x float> %a0, <16 x i16>* %a1) nounwin
 ; AVX512VL-NEXT:    vmovshdup {{.*#+}} xmm0 = xmm0[1,1,3,3]
 ; AVX512VL-NEXT:    callq __gnu_f2h_ieee
 ; AVX512VL-NEXT:    movw %ax, {{[0-9]+}}(%rsp) # 2-byte Spill
-; AVX512VL-NEXT:    vmovdqu64 {{[0-9]+}}(%rsp), %zmm0 # 64-byte Reload
+; AVX512VL-NEXT:    vmovupd {{[0-9]+}}(%rsp), %zmm0 # 64-byte Reload
 ; AVX512VL-NEXT:    vpermilpd {{.*#+}} xmm0 = xmm0[1,0]
 ; AVX512VL-NEXT:    callq __gnu_f2h_ieee
 ; AVX512VL-NEXT:    movw %ax, {{[0-9]+}}(%rsp) # 2-byte Spill
-; AVX512VL-NEXT:    vmovdqu64 {{[0-9]+}}(%rsp), %zmm0 # 64-byte Reload
+; AVX512VL-NEXT:    vmovups {{[0-9]+}}(%rsp), %zmm0 # 64-byte Reload
 ; AVX512VL-NEXT:    vpermilps {{.*#+}} xmm0 = xmm0[3,1,2,3]
 ; AVX512VL-NEXT:    callq __gnu_f2h_ieee
 ; AVX512VL-NEXT:    movw %ax, {{[0-9]+}}(%rsp) # 2-byte Spill
@@ -4568,11 +4568,11 @@ define void @store_cvt_16f32_to_16i16(<16 x float> %a0, <16 x i16>* %a1) nounwin
 ; AVX512VL-NEXT:    vmovshdup {{.*#+}} xmm0 = xmm0[1,1,3,3]
 ; AVX512VL-NEXT:    callq __gnu_f2h_ieee
 ; AVX512VL-NEXT:    movw %ax, {{[0-9]+}}(%rsp) # 2-byte Spill
-; AVX512VL-NEXT:    vmovdqa64 {{[0-9]+}}(%rsp), %xmm0 # 16-byte Reload
+; AVX512VL-NEXT:    vmovapd {{[0-9]+}}(%rsp), %xmm0 # 16-byte Reload
 ; AVX512VL-NEXT:    vpermilpd {{.*#+}} xmm0 = xmm0[1,0]
 ; AVX512VL-NEXT:    callq __gnu_f2h_ieee
 ; AVX512VL-NEXT:    movw %ax, {{[0-9]+}}(%rsp) # 2-byte Spill
-; AVX512VL-NEXT:    vmovdqa64 {{[0-9]+}}(%rsp), %xmm0 # 16-byte Reload
+; AVX512VL-NEXT:    vmovaps {{[0-9]+}}(%rsp), %xmm0 # 16-byte Reload
 ; AVX512VL-NEXT:    vpermilps {{.*#+}} xmm0 = xmm0[3,1,2,3]
 ; AVX512VL-NEXT:    callq __gnu_f2h_ieee
 ; AVX512VL-NEXT:    movw %ax, {{[0-9]+}}(%rsp) # 2-byte Spill
@@ -4582,11 +4582,11 @@ define void @store_cvt_16f32_to_16i16(<16 x float> %a0, <16 x i16>* %a1) nounwin
 ; AVX512VL-NEXT:    vmovshdup {{.*#+}} xmm0 = xmm0[1,1,3,3]
 ; AVX512VL-NEXT:    callq __gnu_f2h_ieee
 ; AVX512VL-NEXT:    movw %ax, {{[0-9]+}}(%rsp) # 2-byte Spill
-; AVX512VL-NEXT:    vmovdqu64 {{[0-9]+}}(%rsp), %ymm0 # 32-byte Reload
+; AVX512VL-NEXT:    vmovupd {{[0-9]+}}(%rsp), %ymm0 # 32-byte Reload
 ; AVX512VL-NEXT:    vpermilpd {{.*#+}} xmm0 = xmm0[1,0]
 ; AVX512VL-NEXT:    callq __gnu_f2h_ieee
 ; AVX512VL-NEXT:    movw %ax, {{[0-9]+}}(%rsp) # 2-byte Spill
-; AVX512VL-NEXT:    vmovdqu64 {{[0-9]+}}(%rsp), %ymm0 # 32-byte Reload
+; AVX512VL-NEXT:    vmovups {{[0-9]+}}(%rsp), %ymm0 # 32-byte Reload
 ; AVX512VL-NEXT:    vpermilps {{.*#+}} xmm0 = xmm0[3,1,2,3]
 ; AVX512VL-NEXT:    callq __gnu_f2h_ieee
 ; AVX512VL-NEXT:    movw %ax, {{[0-9]+}}(%rsp) # 2-byte Spill
@@ -4596,11 +4596,11 @@ define void @store_cvt_16f32_to_16i16(<16 x float> %a0, <16 x i16>* %a1) nounwin
 ; AVX512VL-NEXT:    vmovshdup {{.*#+}} xmm0 = xmm0[1,1,3,3]
 ; AVX512VL-NEXT:    callq __gnu_f2h_ieee
 ; AVX512VL-NEXT:    movw %ax, {{[0-9]+}}(%rsp) # 2-byte Spill
-; AVX512VL-NEXT:    vmovdqa64 {{[0-9]+}}(%rsp), %xmm0 # 16-byte Reload
+; AVX512VL-NEXT:    vmovapd {{[0-9]+}}(%rsp), %xmm0 # 16-byte Reload
 ; AVX512VL-NEXT:    vpermilpd {{.*#+}} xmm0 = xmm0[1,0]
 ; AVX512VL-NEXT:    callq __gnu_f2h_ieee
 ; AVX512VL-NEXT:    movl %eax, %r14d
-; AVX512VL-NEXT:    vmovdqa64 {{[0-9]+}}(%rsp), %xmm0 # 16-byte Reload
+; AVX512VL-NEXT:    vmovaps {{[0-9]+}}(%rsp), %xmm0 # 16-byte Reload
 ; AVX512VL-NEXT:    vpermilps {{.*#+}} xmm0 = xmm0[3,1,2,3]
 ; AVX512VL-NEXT:    callq __gnu_f2h_ieee
 ; AVX512VL-NEXT:    movl %eax, %r15d
@@ -4671,77 +4671,23 @@ define i16 @cvt_f64_to_i16(double %a0) nounwind {
 }
 
 define <2 x i16> @cvt_2f64_to_2i16(<2 x double> %a0) nounwind {
-; AVX1-LABEL: cvt_2f64_to_2i16:
-; AVX1:       # BB#0:
-; AVX1-NEXT:    pushq %rbx
-; AVX1-NEXT:    subq $16, %rsp
-; AVX1-NEXT:    vmovapd %xmm0, (%rsp) # 16-byte Spill
-; AVX1-NEXT:    vpermilpd {{.*#+}} xmm0 = xmm0[1,0]
-; AVX1-NEXT:    callq __truncdfhf2
-; AVX1-NEXT:    movw %ax, %bx
-; AVX1-NEXT:    shll $16, %ebx
-; AVX1-NEXT:    vmovaps (%rsp), %xmm0 # 16-byte Reload
-; AVX1-NEXT:    callq __truncdfhf2
-; AVX1-NEXT:    movzwl %ax, %eax
-; AVX1-NEXT:    orl %ebx, %eax
-; AVX1-NEXT:    vmovd %eax, %xmm0
-; AVX1-NEXT:    addq $16, %rsp
-; AVX1-NEXT:    popq %rbx
-; AVX1-NEXT:    retq
-;
-; AVX2-LABEL: cvt_2f64_to_2i16:
-; AVX2:       # BB#0:
-; AVX2-NEXT:    pushq %rbx
-; AVX2-NEXT:    subq $16, %rsp
-; AVX2-NEXT:    vmovapd %xmm0, (%rsp) # 16-byte Spill
-; AVX2-NEXT:    vpermilpd {{.*#+}} xmm0 = xmm0[1,0]
-; AVX2-NEXT:    callq __truncdfhf2
-; AVX2-NEXT:    movw %ax, %bx
-; AVX2-NEXT:    shll $16, %ebx
-; AVX2-NEXT:    vmovaps (%rsp), %xmm0 # 16-byte Reload
-; AVX2-NEXT:    callq __truncdfhf2
-; AVX2-NEXT:    movzwl %ax, %eax
-; AVX2-NEXT:    orl %ebx, %eax
-; AVX2-NEXT:    vmovd %eax, %xmm0
-; AVX2-NEXT:    addq $16, %rsp
-; AVX2-NEXT:    popq %rbx
-; AVX2-NEXT:    retq
-;
-; AVX512F-LABEL: cvt_2f64_to_2i16:
-; AVX512F:       # BB#0:
-; AVX512F-NEXT:    pushq %rbx
-; AVX512F-NEXT:    subq $16, %rsp
-; AVX512F-NEXT:    vmovapd %xmm0, (%rsp) # 16-byte Spill
-; AVX512F-NEXT:    vpermilpd {{.*#+}} xmm0 = xmm0[1,0]
-; AVX512F-NEXT:    callq __truncdfhf2
-; AVX512F-NEXT:    movw %ax, %bx
-; AVX512F-NEXT:    shll $16, %ebx
-; AVX512F-NEXT:    vmovaps (%rsp), %xmm0 # 16-byte Reload
-; AVX512F-NEXT:    callq __truncdfhf2
-; AVX512F-NEXT:    movzwl %ax, %eax
-; AVX512F-NEXT:    orl %ebx, %eax
-; AVX512F-NEXT:    vmovd %eax, %xmm0
-; AVX512F-NEXT:    addq $16, %rsp
-; AVX512F-NEXT:    popq %rbx
-; AVX512F-NEXT:    retq
-;
-; AVX512VL-LABEL: cvt_2f64_to_2i16:
-; AVX512VL:       # BB#0:
-; AVX512VL-NEXT:    pushq %rbx
-; AVX512VL-NEXT:    subq $16, %rsp
-; AVX512VL-NEXT:    vmovdqa64 %xmm0, (%rsp) # 16-byte Spill
-; AVX512VL-NEXT:    vpermilpd {{.*#+}} xmm0 = xmm0[1,0]
-; AVX512VL-NEXT:    callq __truncdfhf2
-; AVX512VL-NEXT:    movw %ax, %bx
-; AVX512VL-NEXT:    shll $16, %ebx
-; AVX512VL-NEXT:    vmovaps (%rsp), %xmm0 # 16-byte Reload
-; AVX512VL-NEXT:    callq __truncdfhf2
-; AVX512VL-NEXT:    movzwl %ax, %eax
-; AVX512VL-NEXT:    orl %ebx, %eax
-; AVX512VL-NEXT:    vmovd %eax, %xmm0
-; AVX512VL-NEXT:    addq $16, %rsp
-; AVX512VL-NEXT:    popq %rbx
-; AVX512VL-NEXT:    retq
+; ALL-LABEL: cvt_2f64_to_2i16:
+; ALL:       # BB#0:
+; ALL-NEXT:    pushq %rbx
+; ALL-NEXT:    subq $16, %rsp
+; ALL-NEXT:    vmovapd %xmm0, (%rsp) # 16-byte Spill
+; ALL-NEXT:    vpermilpd {{.*#+}} xmm0 = xmm0[1,0]
+; ALL-NEXT:    callq __truncdfhf2
+; ALL-NEXT:    movw %ax, %bx
+; ALL-NEXT:    shll $16, %ebx
+; ALL-NEXT:    vmovaps (%rsp), %xmm0 # 16-byte Reload
+; ALL-NEXT:    callq __truncdfhf2
+; ALL-NEXT:    movzwl %ax, %eax
+; ALL-NEXT:    orl %ebx, %eax
+; ALL-NEXT:    vmovd %eax, %xmm0
+; ALL-NEXT:    addq $16, %rsp
+; ALL-NEXT:    popq %rbx
+; ALL-NEXT:    retq
   %1 = fptrunc <2 x double> %a0 to <2 x half>
   %2 = bitcast <2 x half> %1 to <2 x i16>
   ret <2 x i16> %2
@@ -4861,7 +4807,7 @@ define <4 x i16> @cvt_4f64_to_4i16(<4 x double> %a0) nounwind {
 ; AVX512VL-NEXT:    pushq %r14
 ; AVX512VL-NEXT:    pushq %rbx
 ; AVX512VL-NEXT:    subq $40, %rsp
-; AVX512VL-NEXT:    vmovdqu64 %ymm0, (%rsp) # 32-byte Spill
+; AVX512VL-NEXT:    vmovupd %ymm0, (%rsp) # 32-byte Spill
 ; AVX512VL-NEXT:    vpermilpd {{.*#+}} xmm0 = xmm0[1,0]
 ; AVX512VL-NEXT:    callq __truncdfhf2
 ; AVX512VL-NEXT:    movw %ax, %bx
@@ -5011,7 +4957,7 @@ define <8 x i16> @cvt_4f64_to_8i16_undef(<4 x double> %a0) nounwind {
 ; AVX512VL-NEXT:    pushq %r14
 ; AVX512VL-NEXT:    pushq %rbx
 ; AVX512VL-NEXT:    subq $40, %rsp
-; AVX512VL-NEXT:    vmovdqu64 %ymm0, (%rsp) # 32-byte Spill
+; AVX512VL-NEXT:    vmovupd %ymm0, (%rsp) # 32-byte Spill
 ; AVX512VL-NEXT:    vpermilpd {{.*#+}} xmm0 = xmm0[1,0]
 ; AVX512VL-NEXT:    callq __truncdfhf2
 ; AVX512VL-NEXT:    movw %ax, %bx
@@ -5165,7 +5111,7 @@ define <8 x i16> @cvt_4f64_to_8i16_zero(<4 x double> %a0) nounwind {
 ; AVX512VL-NEXT:    pushq %r14
 ; AVX512VL-NEXT:    pushq %rbx
 ; AVX512VL-NEXT:    subq $40, %rsp
-; AVX512VL-NEXT:    vmovdqu64 %ymm0, (%rsp) # 32-byte Spill
+; AVX512VL-NEXT:    vmovupd %ymm0, (%rsp) # 32-byte Spill
 ; AVX512VL-NEXT:    vpermilpd {{.*#+}} xmm0 = xmm0[1,0]
 ; AVX512VL-NEXT:    callq __truncdfhf2
 ; AVX512VL-NEXT:    movw %ax, %bx
@@ -5410,7 +5356,7 @@ define <8 x i16> @cvt_8f64_to_8i16(<8 x double> %a0) nounwind {
 ; AVX512VL-NEXT:    pushq %r14
 ; AVX512VL-NEXT:    pushq %rbx
 ; AVX512VL-NEXT:    subq $96, %rsp
-; AVX512VL-NEXT:    vmovdqu64 %zmm0, (%rsp) # 64-byte Spill
+; AVX512VL-NEXT:    vmovupd %zmm0, (%rsp) # 64-byte Spill
 ; AVX512VL-NEXT:    vpermilpd {{.*#+}} xmm0 = xmm0[1,0]
 ; AVX512VL-NEXT:    callq __truncdfhf2
 ; AVX512VL-NEXT:    movw %ax, %bx
@@ -5491,81 +5437,24 @@ define void @store_cvt_f64_to_i16(double %a0, i16* %a1) nounwind {
 }
 
 define void @store_cvt_2f64_to_2i16(<2 x double> %a0, <2 x i16>* %a1) nounwind {
-; AVX1-LABEL: store_cvt_2f64_to_2i16:
-; AVX1:       # BB#0:
-; AVX1-NEXT:    pushq %rbp
-; AVX1-NEXT:    pushq %rbx
-; AVX1-NEXT:    subq $24, %rsp
-; AVX1-NEXT:    movq %rdi, %rbx
-; AVX1-NEXT:    vmovapd %xmm0, (%rsp) # 16-byte Spill
-; AVX1-NEXT:    vpermilpd {{.*#+}} xmm0 = xmm0[1,0]
-; AVX1-NEXT:    callq __truncdfhf2
-; AVX1-NEXT:    movl %eax, %ebp
-; AVX1-NEXT:    vmovaps (%rsp), %xmm0 # 16-byte Reload
-; AVX1-NEXT:    callq __truncdfhf2
-; AVX1-NEXT:    movw %ax, (%rbx)
-; AVX1-NEXT:    movw %bp, 2(%rbx)
-; AVX1-NEXT:    addq $24, %rsp
-; AVX1-NEXT:    popq %rbx
-; AVX1-NEXT:    popq %rbp
-; AVX1-NEXT:    retq
-;
-; AVX2-LABEL: store_cvt_2f64_to_2i16:
-; AVX2:       # BB#0:
-; AVX2-NEXT:    pushq %rbp
-; AVX2-NEXT:    pushq %rbx
-; AVX2-NEXT:    subq $24, %rsp
-; AVX2-NEXT:    movq %rdi, %rbx
-; AVX2-NEXT:    vmovapd %xmm0, (%rsp) # 16-byte Spill
-; AVX2-NEXT:    vpermilpd {{.*#+}} xmm0 = xmm0[1,0]
-; AVX2-NEXT:    callq __truncdfhf2
-; AVX2-NEXT:    movl %eax, %ebp
-; AVX2-NEXT:    vmovaps (%rsp), %xmm0 # 16-byte Reload
-; AVX2-NEXT:    callq __truncdfhf2
-; AVX2-NEXT:    movw %ax, (%rbx)
-; AVX2-NEXT:    movw %bp, 2(%rbx)
-; AVX2-NEXT:    addq $24, %rsp
-; AVX2-NEXT:    popq %rbx
-; AVX2-NEXT:    popq %rbp
-; AVX2-NEXT:    retq
-;
-; AVX512F-LABEL: store_cvt_2f64_to_2i16:
-; AVX512F:       # BB#0:
-; AVX512F-NEXT:    pushq %rbp
-; AVX512F-NEXT:    pushq %rbx
-; AVX512F-NEXT:    subq $24, %rsp
-; AVX512F-NEXT:    movq %rdi, %rbx
-; AVX512F-NEXT:    vmovapd %xmm0, (%rsp) # 16-byte Spill
-; AVX512F-NEXT:    vpermilpd {{.*#+}} xmm0 = xmm0[1,0]
-; AVX512F-NEXT:    callq __truncdfhf2
-; AVX512F-NEXT:    movl %eax, %ebp
-; AVX512F-NEXT:    vmovaps (%rsp), %xmm0 # 16-byte Reload
-; AVX512F-NEXT:    callq __truncdfhf2
-; AVX512F-NEXT:    movw %ax, (%rbx)
-; AVX512F-NEXT:    movw %bp, 2(%rbx)
-; AVX512F-NEXT:    addq $24, %rsp
-; AVX512F-NEXT:    popq %rbx
-; AVX512F-NEXT:    popq %rbp
-; AVX512F-NEXT:    retq
-;
-; AVX512VL-LABEL: store_cvt_2f64_to_2i16:
-; AVX512VL:       # BB#0:
-; AVX512VL-NEXT:    pushq %rbp
-; AVX512VL-NEXT:    pushq %rbx
-; AVX512VL-NEXT:    subq $24, %rsp
-; AVX512VL-NEXT:    movq %rdi, %rbx
-; AVX512VL-NEXT:    vmovdqa64 %xmm0, (%rsp) # 16-byte Spill
-; AVX512VL-NEXT:    vpermilpd {{.*#+}} xmm0 = xmm0[1,0]
-; AVX512VL-NEXT:    callq __truncdfhf2
-; AVX512VL-NEXT:    movl %eax, %ebp
-; AVX512VL-NEXT:    vmovaps (%rsp), %xmm0 # 16-byte Reload
-; AVX512VL-NEXT:    callq __truncdfhf2
-; AVX512VL-NEXT:    movw %ax, (%rbx)
-; AVX512VL-NEXT:    movw %bp, 2(%rbx)
-; AVX512VL-NEXT:    addq $24, %rsp
-; AVX512VL-NEXT:    popq %rbx
-; AVX512VL-NEXT:    popq %rbp
-; AVX512VL-NEXT:    retq
+; ALL-LABEL: store_cvt_2f64_to_2i16:
+; ALL:       # BB#0:
+; ALL-NEXT:    pushq %rbp
+; ALL-NEXT:    pushq %rbx
+; ALL-NEXT:    subq $24, %rsp
+; ALL-NEXT:    movq %rdi, %rbx
+; ALL-NEXT:    vmovapd %xmm0, (%rsp) # 16-byte Spill
+; ALL-NEXT:    vpermilpd {{.*#+}} xmm0 = xmm0[1,0]
+; ALL-NEXT:    callq __truncdfhf2
+; ALL-NEXT:    movl %eax, %ebp
+; ALL-NEXT:    vmovaps (%rsp), %xmm0 # 16-byte Reload
+; ALL-NEXT:    callq __truncdfhf2
+; ALL-NEXT:    movw %ax, (%rbx)
+; ALL-NEXT:    movw %bp, 2(%rbx)
+; ALL-NEXT:    addq $24, %rsp
+; ALL-NEXT:    popq %rbx
+; ALL-NEXT:    popq %rbp
+; ALL-NEXT:    retq
   %1 = fptrunc <2 x double> %a0 to <2 x half>
   %2 = bitcast <2 x half> %1 to <2 x i16>
   store <2 x i16> %2, <2 x i16>* %a1
@@ -5692,7 +5581,7 @@ define void @store_cvt_4f64_to_4i16(<4 x double> %a0, <4 x i16>* %a1) nounwind {
 ; AVX512VL-NEXT:    pushq %rbx
 ; AVX512VL-NEXT:    subq $88, %rsp
 ; AVX512VL-NEXT:    movq %rdi, %rbx
-; AVX512VL-NEXT:    vmovdqu64 %ymm0, {{[0-9]+}}(%rsp) # 32-byte Spill
+; AVX512VL-NEXT:    vmovupd %ymm0, {{[0-9]+}}(%rsp) # 32-byte Spill
 ; AVX512VL-NEXT:    vpermilpd {{.*#+}} xmm0 = xmm0[1,0]
 ; AVX512VL-NEXT:    callq __truncdfhf2
 ; AVX512VL-NEXT:    movl %eax, %r14d
@@ -5855,7 +5744,7 @@ define void @store_cvt_4f64_to_8i16_undef(<4 x double> %a0, <8 x i16>* %a1) noun
 ; AVX512VL-NEXT:    pushq %rbx
 ; AVX512VL-NEXT:    subq $32, %rsp
 ; AVX512VL-NEXT:    movq %rdi, %r14
-; AVX512VL-NEXT:    vmovdqu64 %ymm0, (%rsp) # 32-byte Spill
+; AVX512VL-NEXT:    vmovupd %ymm0, (%rsp) # 32-byte Spill
 ; AVX512VL-NEXT:    vpermilpd {{.*#+}} xmm0 = xmm0[1,0]
 ; AVX512VL-NEXT:    callq __truncdfhf2
 ; AVX512VL-NEXT:    movw %ax, %bp
@@ -6026,7 +5915,7 @@ define void @store_cvt_4f64_to_8i16_zero(<4 x double> %a0, <8 x i16>* %a1) nounw
 ; AVX512VL-NEXT:    pushq %rbx
 ; AVX512VL-NEXT:    subq $32, %rsp
 ; AVX512VL-NEXT:    movq %rdi, %r14
-; AVX512VL-NEXT:    vmovdqu64 %ymm0, (%rsp) # 32-byte Spill
+; AVX512VL-NEXT:    vmovupd %ymm0, (%rsp) # 32-byte Spill
 ; AVX512VL-NEXT:    vpermilpd {{.*#+}} xmm0 = xmm0[1,0]
 ; AVX512VL-NEXT:    callq __truncdfhf2
 ; AVX512VL-NEXT:    movw %ax, %bp
@@ -6281,7 +6170,7 @@ define void @store_cvt_8f64_to_8i16(<8 x double> %a0, <8 x i16>* %a1) nounwind {
 ; AVX512VL-NEXT:    pushq %rbx
 ; AVX512VL-NEXT:    subq $200, %rsp
 ; AVX512VL-NEXT:    movq %rdi, %rbx
-; AVX512VL-NEXT:    vmovdqu64 %zmm0, {{[0-9]+}}(%rsp) # 64-byte Spill
+; AVX512VL-NEXT:    vmovupd %zmm0, {{[0-9]+}}(%rsp) # 64-byte Spill
 ; AVX512VL-NEXT:    vpermilpd {{.*#+}} xmm0 = xmm0[1,0]
 ; AVX512VL-NEXT:    callq __truncdfhf2
 ; AVX512VL-NEXT:    movw %ax, {{[0-9]+}}(%rsp) # 2-byte Spill
diff --git a/llvm/test/CodeGen/X86/vector-shuffle-256-v4.ll b/llvm/test/CodeGen/X86/vector-shuffle-256-v4.ll
index d853655..22b6e28 100644
--- a/llvm/test/CodeGen/X86/vector-shuffle-256-v4.ll
+++ b/llvm/test/CodeGen/X86/vector-shuffle-256-v4.ll
@@ -389,45 +389,21 @@ define <4 x double> @shuffle_v4f64_1054(<4 x double> %a, <4 x double> %b) {
 }
 
 define <4 x double> @shuffle_v4f64_3254(<4 x double> %a, <4 x double> %b) {
-; AVX1-LABEL: shuffle_v4f64_3254:
-; AVX1:       # BB#0:
-; AVX1-NEXT:    vperm2f128 {{.*#+}} ymm0 = ymm0[2,3],ymm1[0,1]
-; AVX1-NEXT:    vpermilpd {{.*#+}} ymm0 = ymm0[1,0,3,2]
-; AVX1-NEXT:    retq
-;
-; AVX2-LABEL: shuffle_v4f64_3254:
-; AVX2:       # BB#0:
-; AVX2-NEXT:    vperm2f128 {{.*#+}} ymm0 = ymm0[2,3],ymm1[0,1]
-; AVX2-NEXT:    vpermilpd {{.*#+}} ymm0 = ymm0[1,0,3,2]
-; AVX2-NEXT:    retq
-;
-; AVX512VL-LABEL: shuffle_v4f64_3254:
-; AVX512VL:       # BB#0:
-; AVX512VL-NEXT:    vperm2i128 {{.*#+}} ymm0 = ymm0[2,3],ymm1[0,1]
-; AVX512VL-NEXT:    vpermilpd {{.*#+}} ymm0 = ymm0[1,0,3,2]
-; AVX512VL-NEXT:    retq
+; ALL-LABEL: shuffle_v4f64_3254:
+; ALL:       # BB#0:
+; ALL-NEXT:    vperm2f128 {{.*#+}} ymm0 = ymm0[2,3],ymm1[0,1]
+; ALL-NEXT:    vpermilpd {{.*#+}} ymm0 = ymm0[1,0,3,2]
+; ALL-NEXT:    retq
   %shuffle = shufflevector <4 x double> %a, <4 x double> %b, <4 x i32> <i32 3, i32 2, i32 5, i32 4>
   ret <4 x double> %shuffle
 }
 
 define <4 x double> @shuffle_v4f64_3276(<4 x double> %a, <4 x double> %b) {
-; AVX1-LABEL: shuffle_v4f64_3276:
-; AVX1:       # BB#0:
-; AVX1-NEXT:    vperm2f128 {{.*#+}} ymm0 = ymm0[2,3],ymm1[2,3]
-; AVX1-NEXT:    vpermilpd {{.*#+}} ymm0 = ymm0[1,0,3,2]
-; AVX1-NEXT:    retq
-;
-; AVX2-LABEL: shuffle_v4f64_3276:
-; AVX2:       # BB#0:
-; AVX2-NEXT:    vperm2f128 {{.*#+}} ymm0 = ymm0[2,3],ymm1[2,3]
-; AVX2-NEXT:    vpermilpd {{.*#+}} ymm0 = ymm0[1,0,3,2]
-; AVX2-NEXT:    retq
-;
-; AVX512VL-LABEL: shuffle_v4f64_3276:
-; AVX512VL:       # BB#0:
-; AVX512VL-NEXT:    vperm2i128 {{.*#+}} ymm0 = ymm0[2,3],ymm1[2,3]
-; AVX512VL-NEXT:    vpermilpd {{.*#+}} ymm0 = ymm0[1,0,3,2]
-; AVX512VL-NEXT:    retq
+; ALL-LABEL: shuffle_v4f64_3276:
+; ALL:       # BB#0:
+; ALL-NEXT:    vperm2f128 {{.*#+}} ymm0 = ymm0[2,3],ymm1[2,3]
+; ALL-NEXT:    vpermilpd {{.*#+}} ymm0 = ymm0[1,0,3,2]
+; ALL-NEXT:    retq
   %shuffle = shufflevector <4 x double> %a, <4 x double> %b, <4 x i32> <i32 3, i32 2, i32 7, i32 6>
   ret <4 x double> %shuffle
 }
diff --git a/llvm/test/CodeGen/X86/vector-shuffle-512-v16.ll b/llvm/test/CodeGen/X86/vector-shuffle-512-v16.ll
index 4d441aa..6cd0366 100644
--- a/llvm/test/CodeGen/X86/vector-shuffle-512-v16.ll
+++ b/llvm/test/CodeGen/X86/vector-shuffle-512-v16.ll
@@ -190,7 +190,7 @@ define <16 x i32> @shuffle_v16i32_01_02_03_16_05_06_07_20_09_10_11_24_13_14_15_2
 define <16 x float> @shuffle_v16f32_02_05_u_u_07_u_0a_01_00_05_u_04_07_u_0a_01(<16 x float> %a)  {
 ; ALL-LABEL: shuffle_v16f32_02_05_u_u_07_u_0a_01_00_05_u_04_07_u_0a_01:
 ; ALL:       # BB#0:
-; ALL-NEXT:    vmovdqa32 {{.*#+}} zmm1 = <2,5,u,u,7,u,10,1,0,5,u,4,7,u,10,1>
+; ALL-NEXT:    vmovaps {{.*#+}} zmm1 = <2,5,u,u,7,u,10,1,0,5,u,4,7,u,10,1>
 ; ALL-NEXT:    vpermps %zmm0, %zmm1, %zmm0
 ; ALL-NEXT:    retq
   %c = shufflevector <16 x float> %a, <16 x float> undef, <16 x i32> <i32 2, i32 5, i32 undef, i32 undef, i32 7, i32 undef, i32 10, i32 1,  i32 0, i32 5, i32 undef, i32 4, i32 7, i32 undef, i32 10, i32 1>
diff --git a/llvm/test/CodeGen/X86/vector-shuffle-512-v8.ll b/llvm/test/CodeGen/X86/vector-shuffle-512-v8.ll
index da8ea83..6e2df62 100644
--- a/llvm/test/CodeGen/X86/vector-shuffle-512-v8.ll
+++ b/llvm/test/CodeGen/X86/vector-shuffle-512-v8.ll
@@ -51,13 +51,13 @@ define <8 x double> @shuffle_v8f64_44444444(<8 x double> %a, <8 x double> %b) {
 define <8 x double> @shuffle_v8f64_00000010(<8 x double> %a, <8 x double> %b) {
 ; AVX512F-LABEL: shuffle_v8f64_00000010:
 ; AVX512F:       # BB#0:
-; AVX512F-NEXT:    vmovdqa64 {{.*#+}} zmm1 = [0,0,0,0,0,0,1,0]
+; AVX512F-NEXT:    vmovapd {{.*#+}} zmm1 = [0,0,0,0,0,0,1,0]
 ; AVX512F-NEXT:    vpermpd %zmm0, %zmm1, %zmm0
 ; AVX512F-NEXT:    retq
 ;
 ; AVX512F-32-LABEL: shuffle_v8f64_00000010:
 ; AVX512F-32:       # BB#0:
-; AVX512F-32-NEXT:    vmovdqa64 {{.*#+}} zmm1 = [0,0,0,0,0,0,0,0,0,0,0,0,1,0,0,0]
+; AVX512F-32-NEXT:    vmovapd {{.*#+}} zmm1 = [0,0,0,0,0,0,0,0,0,0,0,0,1,0,0,0]
 ; AVX512F-32-NEXT:    vpermpd %zmm0, %zmm1, %zmm0
 ; AVX512F-32-NEXT:    retl
   %shuffle = shufflevector <8 x double> %a, <8 x double> %b, <8 x i32> <i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 1, i32 0>
@@ -67,13 +67,13 @@ define <8 x double> @shuffle_v8f64_00000010(<8 x double> %a, <8 x double> %b) {
 define <8 x double> @shuffle_v8f64_00000200(<8 x double> %a, <8 x double> %b) {
 ; AVX512F-LABEL: shuffle_v8f64_00000200:
 ; AVX512F:       # BB#0:
-; AVX512F-NEXT:    vmovdqa64 {{.*#+}} zmm1 = [0,0,0,0,0,2,0,0]
+; AVX512F-NEXT:    vmovapd {{.*#+}} zmm1 = [0,0,0,0,0,2,0,0]
 ; AVX512F-NEXT:    vpermpd %zmm0, %zmm1, %zmm0
 ; AVX512F-NEXT:    retq
 ;
 ; AVX512F-32-LABEL: shuffle_v8f64_00000200:
 ; AVX512F-32:       # BB#0:
-; AVX512F-32-NEXT:    vmovdqa64 {{.*#+}} zmm1 = [0,0,0,0,0,0,0,0,0,0,2,0,0,0,0,0]
+; AVX512F-32-NEXT:    vmovapd {{.*#+}} zmm1 = [0,0,0,0,0,0,0,0,0,0,2,0,0,0,0,0]
 ; AVX512F-32-NEXT:    vpermpd %zmm0, %zmm1, %zmm0
 ; AVX512F-32-NEXT:    retl
   %shuffle = shufflevector <8 x double> %a, <8 x double> %b, <8 x i32> <i32 0, i32 0, i32 0, i32 0, i32 0, i32 2, i32 0, i32 0>
@@ -83,13 +83,13 @@ define <8 x double> @shuffle_v8f64_00000200(<8 x double> %a, <8 x double> %b) {
 define <8 x double> @shuffle_v8f64_00003000(<8 x double> %a, <8 x double> %b) {
 ; AVX512F-LABEL: shuffle_v8f64_00003000:
 ; AVX512F:       # BB#0:
-; AVX512F-NEXT:    vmovdqa64 {{.*#+}} zmm1 = [0,0,0,0,3,0,0,0]
+; AVX512F-NEXT:    vmovapd {{.*#+}} zmm1 = [0,0,0,0,3,0,0,0]
 ; AVX512F-NEXT:    vpermpd %zmm0, %zmm1, %zmm0
 ; AVX512F-NEXT:    retq
 ;
 ; AVX512F-32-LABEL: shuffle_v8f64_00003000:
 ; AVX512F-32:       # BB#0:
-; AVX512F-32-NEXT:    vmovdqa64 {{.*#+}} zmm1 = [0,0,0,0,0,0,0,0,3,0,0,0,0,0,0,0]
+; AVX512F-32-NEXT:    vmovapd {{.*#+}} zmm1 = [0,0,0,0,0,0,0,0,3,0,0,0,0,0,0,0]
 ; AVX512F-32-NEXT:    vpermpd %zmm0, %zmm1, %zmm0
 ; AVX512F-32-NEXT:    retl
   %shuffle = shufflevector <8 x double> %a, <8 x double> %b, <8 x i32> <i32 0, i32 0, i32 0, i32 0, i32 3, i32 0, i32 0, i32 0>
@@ -99,13 +99,13 @@ define <8 x double> @shuffle_v8f64_00003000(<8 x double> %a, <8 x double> %b) {
 define <8 x double> @shuffle_v8f64_00040000(<8 x double> %a, <8 x double> %b) {
 ; AVX512F-LABEL: shuffle_v8f64_00040000:
 ; AVX512F:       # BB#0:
-; AVX512F-NEXT:    vmovdqa64 {{.*#+}} zmm1 = [0,0,0,4,0,0,0,0]
+; AVX512F-NEXT:    vmovapd {{.*#+}} zmm1 = [0,0,0,4,0,0,0,0]
 ; AVX512F-NEXT:    vpermpd %zmm0, %zmm1, %zmm0
 ; AVX512F-NEXT:    retq
 ;
 ; AVX512F-32-LABEL: shuffle_v8f64_00040000:
 ; AVX512F-32:       # BB#0:
-; AVX512F-32-NEXT:    vmovdqa64 {{.*#+}} zmm1 = [0,0,0,0,0,0,4,0,0,0,0,0,0,0,0,0]
+; AVX512F-32-NEXT:    vmovapd {{.*#+}} zmm1 = [0,0,0,0,0,0,4,0,0,0,0,0,0,0,0,0]
 ; AVX512F-32-NEXT:    vpermpd %zmm0, %zmm1, %zmm0
 ; AVX512F-32-NEXT:    retl
   %shuffle = shufflevector <8 x double> %a, <8 x double> %b, <8 x i32> <i32 0, i32 0, i32 0, i32 4, i32 0, i32 0, i32 0, i32 0>
@@ -115,13 +115,13 @@ define <8 x double> @shuffle_v8f64_00040000(<8 x double> %a, <8 x double> %b) {
 define <8 x double> @shuffle_v8f64_00500000(<8 x double> %a, <8 x double> %b) {
 ; AVX512F-LABEL: shuffle_v8f64_00500000:
 ; AVX512F:       # BB#0:
-; AVX512F-NEXT:    vmovdqa64 {{.*#+}} zmm1 = [0,0,5,0,0,0,0,0]
+; AVX512F-NEXT:    vmovapd {{.*#+}} zmm1 = [0,0,5,0,0,0,0,0]
 ; AVX512F-NEXT:    vpermpd %zmm0, %zmm1, %zmm0
 ; AVX512F-NEXT:    retq
 ;
 ; AVX512F-32-LABEL: shuffle_v8f64_00500000:
 ; AVX512F-32:       # BB#0:
-; AVX512F-32-NEXT:    vmovdqa64 {{.*#+}} zmm1 = [0,0,0,0,5,0,0,0,0,0,0,0,0,0,0,0]
+; AVX512F-32-NEXT:    vmovapd {{.*#+}} zmm1 = [0,0,0,0,5,0,0,0,0,0,0,0,0,0,0,0]
 ; AVX512F-32-NEXT:    vpermpd %zmm0, %zmm1, %zmm0
 ; AVX512F-32-NEXT:    retl
   %shuffle = shufflevector <8 x double> %a, <8 x double> %b, <8 x i32> <i32 0, i32 0, i32 5, i32 0, i32 0, i32 0, i32 0, i32 0>
@@ -131,13 +131,13 @@ define <8 x double> @shuffle_v8f64_00500000(<8 x double> %a, <8 x double> %b) {
 define <8 x double> @shuffle_v8f64_06000000(<8 x double> %a, <8 x double> %b) {
 ; AVX512F-LABEL: shuffle_v8f64_06000000:
 ; AVX512F:       # BB#0:
-; AVX512F-NEXT:    vmovdqa64 {{.*#+}} zmm1 = [0,6,0,0,0,0,0,0]
+; AVX512F-NEXT:    vmovapd {{.*#+}} zmm1 = [0,6,0,0,0,0,0,0]
 ; AVX512F-NEXT:    vpermpd %zmm0, %zmm1, %zmm0
 ; AVX512F-NEXT:    retq
 ;
 ; AVX512F-32-LABEL: shuffle_v8f64_06000000:
 ; AVX512F-32:       # BB#0:
-; AVX512F-32-NEXT:    vmovdqa64 {{.*#+}} zmm1 = [0,0,6,0,0,0,0,0,0,0,0,0,0,0,0,0]
+; AVX512F-32-NEXT:    vmovapd {{.*#+}} zmm1 = [0,0,6,0,0,0,0,0,0,0,0,0,0,0,0,0]
 ; AVX512F-32-NEXT:    vpermpd %zmm0, %zmm1, %zmm0
 ; AVX512F-32-NEXT:    retl
   %shuffle = shufflevector <8 x double> %a, <8 x double> %b, <8 x i32> <i32 0, i32 6, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0>
@@ -184,13 +184,13 @@ define <8 x double> @shuffle_v8f64_01014545(<8 x double> %a, <8 x double> %b) {
 define <8 x double> @shuffle_v8f64_00112233(<8 x double> %a, <8 x double> %b) {
 ; AVX512F-LABEL: shuffle_v8f64_00112233:
 ; AVX512F:       # BB#0:
-; AVX512F-NEXT:    vmovdqa64 {{.*#+}} zmm1 = [0,0,1,1,2,2,3,3]
+; AVX512F-NEXT:    vmovapd {{.*#+}} zmm1 = [0,0,1,1,2,2,3,3]
 ; AVX512F-NEXT:    vpermpd %zmm0, %zmm1, %zmm0
 ; AVX512F-NEXT:    retq
 ;
 ; AVX512F-32-LABEL: shuffle_v8f64_00112233:
 ; AVX512F-32:       # BB#0:
-; AVX512F-32-NEXT:    vmovdqa64 {{.*#+}} zmm1 = [0,0,0,0,1,0,1,0,2,0,2,0,3,0,3,0]
+; AVX512F-32-NEXT:    vmovapd {{.*#+}} zmm1 = [0,0,0,0,1,0,1,0,2,0,2,0,3,0,3,0]
 ; AVX512F-32-NEXT:    vpermpd %zmm0, %zmm1, %zmm0
 ; AVX512F-32-NEXT:    retl
   %shuffle = shufflevector <8 x double> %a, <8 x double> %b, <8 x i32> <i32 0, i32 0, i32 1, i32 1, i32 2, i32 2, i32 3, i32 3>
@@ -200,13 +200,13 @@ define <8 x double> @shuffle_v8f64_00112233(<8 x double> %a, <8 x double> %b) {
 define <8 x double> @shuffle_v8f64_00001111(<8 x double> %a, <8 x double> %b) {
 ; AVX512F-LABEL: shuffle_v8f64_00001111:
 ; AVX512F:       # BB#0:
-; AVX512F-NEXT:    vmovdqa64 {{.*#+}} zmm1 = [0,0,0,0,1,1,1,1]
+; AVX512F-NEXT:    vmovapd {{.*#+}} zmm1 = [0,0,0,0,1,1,1,1]
 ; AVX512F-NEXT:    vpermpd %zmm0, %zmm1, %zmm0
 ; AVX512F-NEXT:    retq
 ;
 ; AVX512F-32-LABEL: shuffle_v8f64_00001111:
 ; AVX512F-32:       # BB#0:
-; AVX512F-32-NEXT:    vmovdqa64 {{.*#+}} zmm1 = [0,0,0,0,0,0,0,0,1,0,1,0,1,0,1,0]
+; AVX512F-32-NEXT:    vmovapd {{.*#+}} zmm1 = [0,0,0,0,0,0,0,0,1,0,1,0,1,0,1,0]
 ; AVX512F-32-NEXT:    vpermpd %zmm0, %zmm1, %zmm0
 ; AVX512F-32-NEXT:    retl
   %shuffle = shufflevector <8 x double> %a, <8 x double> %b, <8 x i32> <i32 0, i32 0, i32 0, i32 0, i32 1, i32 1, i32 1, i32 1>
@@ -624,13 +624,13 @@ define <8 x double> @shuffle_v8f64_00015444(<8 x double> %a, <8 x double> %b) {
 ;
 ; AVX512F-LABEL: shuffle_v8f64_00015444:
 ; AVX512F:       # BB#0:
-; AVX512F-NEXT:    vmovdqa64 {{.*#+}} zmm1 = [0,0,0,1,5,4,4,4]
+; AVX512F-NEXT:    vmovapd {{.*#+}} zmm1 = [0,0,0,1,5,4,4,4]
 ; AVX512F-NEXT:    vpermpd %zmm0, %zmm1, %zmm0
 ; AVX512F-NEXT:    retq
 ;
 ; AVX512F-32-LABEL: shuffle_v8f64_00015444:
 ; AVX512F-32:       # BB#0:
-; AVX512F-32-NEXT:    vmovdqa64 {{.*#+}} zmm1 = [0,0,0,0,0,0,1,0,5,0,4,0,4,0,4,0]
+; AVX512F-32-NEXT:    vmovapd {{.*#+}} zmm1 = [0,0,0,0,0,0,1,0,5,0,4,0,4,0,4,0]
 ; AVX512F-32-NEXT:    vpermpd %zmm0, %zmm1, %zmm0
 ; AVX512F-32-NEXT:    retl
   %shuffle = shufflevector <8 x double> %a, <8 x double> %b, <8 x i32> <i32 0, i32 0, i32 0, i32 1, i32 5, i32 4, i32 4, i32 4>
@@ -641,13 +641,13 @@ define <8 x double> @shuffle_v8f64_00204644(<8 x double> %a, <8 x double> %b) {
 ;
 ; AVX512F-LABEL: shuffle_v8f64_00204644:
 ; AVX512F:       # BB#0:
-; AVX512F-NEXT:    vmovdqa64 {{.*#+}} zmm1 = [0,0,2,0,4,6,4,4]
+; AVX512F-NEXT:    vmovapd {{.*#+}} zmm1 = [0,0,2,0,4,6,4,4]
 ; AVX512F-NEXT:    vpermpd %zmm0, %zmm1, %zmm0
 ; AVX512F-NEXT:    retq
 ;
 ; AVX512F-32-LABEL: shuffle_v8f64_00204644:
 ; AVX512F-32:       # BB#0:
-; AVX512F-32-NEXT:    vmovdqa64 {{.*#+}} zmm1 = [0,0,0,0,2,0,0,0,4,0,6,0,4,0,4,0]
+; AVX512F-32-NEXT:    vmovapd {{.*#+}} zmm1 = [0,0,0,0,2,0,0,0,4,0,6,0,4,0,4,0]
 ; AVX512F-32-NEXT:    vpermpd %zmm0, %zmm1, %zmm0
 ; AVX512F-32-NEXT:    retl
   %shuffle = shufflevector <8 x double> %a, <8 x double> %b, <8 x i32> <i32 0, i32 0, i32 2, i32 0, i32 4, i32 6, i32 4, i32 4>
@@ -658,13 +658,13 @@ define <8 x double> @shuffle_v8f64_03004474(<8 x double> %a, <8 x double> %b) {
 ;
 ; AVX512F-LABEL: shuffle_v8f64_03004474:
 ; AVX512F:       # BB#0:
-; AVX512F-NEXT:    vmovdqa64 {{.*#+}} zmm1 = [0,3,0,0,4,4,7,4]
+; AVX512F-NEXT:    vmovapd {{.*#+}} zmm1 = [0,3,0,0,4,4,7,4]
 ; AVX512F-NEXT:    vpermpd %zmm0, %zmm1, %zmm0
 ; AVX512F-NEXT:    retq
 ;
 ; AVX512F-32-LABEL: shuffle_v8f64_03004474:
 ; AVX512F-32:       # BB#0:
-; AVX512F-32-NEXT:    vmovdqa64 {{.*#+}} zmm1 = [0,0,3,0,0,0,0,0,4,0,4,0,7,0,4,0]
+; AVX512F-32-NEXT:    vmovapd {{.*#+}} zmm1 = [0,0,3,0,0,0,0,0,4,0,4,0,7,0,4,0]
 ; AVX512F-32-NEXT:    vpermpd %zmm0, %zmm1, %zmm0
 ; AVX512F-32-NEXT:    retl
   %shuffle = shufflevector <8 x double> %a, <8 x double> %b, <8 x i32> <i32 0, i32 3, i32 0, i32 0, i32 4, i32 4, i32 7, i32 4>
@@ -675,13 +675,13 @@ define <8 x double> @shuffle_v8f64_10004444(<8 x double> %a, <8 x double> %b) {
 ;
 ; AVX512F-LABEL: shuffle_v8f64_10004444:
 ; AVX512F:       # BB#0:
-; AVX512F-NEXT:    vmovdqa64 {{.*#+}} zmm1 = [1,0,0,0,4,4,4,4]
+; AVX512F-NEXT:    vmovapd {{.*#+}} zmm1 = [1,0,0,0,4,4,4,4]
 ; AVX512F-NEXT:    vpermpd %zmm0, %zmm1, %zmm0
 ; AVX512F-NEXT:    retq
 ;
 ; AVX512F-32-LABEL: shuffle_v8f64_10004444:
 ; AVX512F-32:       # BB#0:
-; AVX512F-32-NEXT:    vmovdqa64 {{.*#+}} zmm1 = [1,0,0,0,0,0,0,0,4,0,4,0,4,0,4,0]
+; AVX512F-32-NEXT:    vmovapd {{.*#+}} zmm1 = [1,0,0,0,0,0,0,0,4,0,4,0,4,0,4,0]
 ; AVX512F-32-NEXT:    vpermpd %zmm0, %zmm1, %zmm0
 ; AVX512F-32-NEXT:    retl
   %shuffle = shufflevector <8 x double> %a, <8 x double> %b, <8 x i32> <i32 1, i32 0, i32 0, i32 0, i32 4, i32 4, i32 4, i32 4>
@@ -692,13 +692,13 @@ define <8 x double> @shuffle_v8f64_22006446(<8 x double> %a, <8 x double> %b) {
 ;
 ; AVX512F-LABEL: shuffle_v8f64_22006446:
 ; AVX512F:       # BB#0:
-; AVX512F-NEXT:    vmovdqa64 {{.*#+}} zmm1 = [2,2,0,0,6,4,4,6]
+; AVX512F-NEXT:    vmovapd {{.*#+}} zmm1 = [2,2,0,0,6,4,4,6]
 ; AVX512F-NEXT:    vpermpd %zmm0, %zmm1, %zmm0
 ; AVX512F-NEXT:    retq
 ;
 ; AVX512F-32-LABEL: shuffle_v8f64_22006446:
 ; AVX512F-32:       # BB#0:
-; AVX512F-32-NEXT:    vmovdqa64 {{.*#+}} zmm1 = [2,0,2,0,0,0,0,0,6,0,4,0,4,0,6,0]
+; AVX512F-32-NEXT:    vmovapd {{.*#+}} zmm1 = [2,0,2,0,0,0,0,0,6,0,4,0,4,0,6,0]
 ; AVX512F-32-NEXT:    vpermpd %zmm0, %zmm1, %zmm0
 ; AVX512F-32-NEXT:    retl
   %shuffle = shufflevector <8 x double> %a, <8 x double> %b, <8 x i32> <i32 2, i32 2, i32 0, i32 0, i32 6, i32 4, i32 4, i32 6>
@@ -709,13 +709,13 @@ define <8 x double> @shuffle_v8f64_33307474(<8 x double> %a, <8 x double> %b) {
 ;
 ; AVX512F-LABEL: shuffle_v8f64_33307474:
 ; AVX512F:       # BB#0:
-; AVX512F-NEXT:    vmovdqa64 {{.*#+}} zmm1 = [3,3,3,0,7,4,7,4]
+; AVX512F-NEXT:    vmovapd {{.*#+}} zmm1 = [3,3,3,0,7,4,7,4]
 ; AVX512F-NEXT:    vpermpd %zmm0, %zmm1, %zmm0
 ; AVX512F-NEXT:    retq
 ;
 ; AVX512F-32-LABEL: shuffle_v8f64_33307474:
 ; AVX512F-32:       # BB#0:
-; AVX512F-32-NEXT:    vmovdqa64 {{.*#+}} zmm1 = [3,0,3,0,3,0,0,0,7,0,4,0,7,0,4,0]
+; AVX512F-32-NEXT:    vmovapd {{.*#+}} zmm1 = [3,0,3,0,3,0,0,0,7,0,4,0,7,0,4,0]
 ; AVX512F-32-NEXT:    vpermpd %zmm0, %zmm1, %zmm0
 ; AVX512F-32-NEXT:    retl
   %shuffle = shufflevector <8 x double> %a, <8 x double> %b, <8 x i32> <i32 3, i32 3, i32 3, i32 0, i32 7, i32 4, i32 7, i32 4>
@@ -726,13 +726,13 @@ define <8 x double> @shuffle_v8f64_32104567(<8 x double> %a, <8 x double> %b) {
 ;
 ; AVX512F-LABEL: shuffle_v8f64_32104567:
 ; AVX512F:       # BB#0:
-; AVX512F-NEXT:    vmovdqa64 {{.*#+}} zmm1 = [3,2,1,0,4,5,6,7]
+; AVX512F-NEXT:    vmovapd {{.*#+}} zmm1 = [3,2,1,0,4,5,6,7]
 ; AVX512F-NEXT:    vpermpd %zmm0, %zmm1, %zmm0
 ; AVX512F-NEXT:    retq
 ;
 ; AVX512F-32-LABEL: shuffle_v8f64_32104567:
 ; AVX512F-32:       # BB#0:
-; AVX512F-32-NEXT:    vmovdqa64 {{.*#+}} zmm1 = [3,0,2,0,1,0,0,0,4,0,5,0,6,0,7,0]
+; AVX512F-32-NEXT:    vmovapd {{.*#+}} zmm1 = [3,0,2,0,1,0,0,0,4,0,5,0,6,0,7,0]
 ; AVX512F-32-NEXT:    vpermpd %zmm0, %zmm1, %zmm0
 ; AVX512F-32-NEXT:    retl
   %shuffle = shufflevector <8 x double> %a, <8 x double> %b, <8 x i32> <i32 3, i32 2, i32 1, i32 0, i32 4, i32 5, i32 6, i32 7>
@@ -743,13 +743,13 @@ define <8 x double> @shuffle_v8f64_00236744(<8 x double> %a, <8 x double> %b) {
 ;
 ; AVX512F-LABEL: shuffle_v8f64_00236744:
 ; AVX512F:       # BB#0:
-; AVX512F-NEXT:    vmovdqa64 {{.*#+}} zmm1 = [0,0,2,3,6,7,4,4]
+; AVX512F-NEXT:    vmovapd {{.*#+}} zmm1 = [0,0,2,3,6,7,4,4]
 ; AVX512F-NEXT:    vpermpd %zmm0, %zmm1, %zmm0
 ; AVX512F-NEXT:    retq
 ;
 ; AVX512F-32-LABEL: shuffle_v8f64_00236744:
 ; AVX512F-32:       # BB#0:
-; AVX512F-32-NEXT:    vmovdqa64 {{.*#+}} zmm1 = [0,0,0,0,2,0,3,0,6,0,7,0,4,0,4,0]
+; AVX512F-32-NEXT:    vmovapd {{.*#+}} zmm1 = [0,0,0,0,2,0,3,0,6,0,7,0,4,0,4,0]
 ; AVX512F-32-NEXT:    vpermpd %zmm0, %zmm1, %zmm0
 ; AVX512F-32-NEXT:    retl
   %shuffle = shufflevector <8 x double> %a, <8 x double> %b, <8 x i32> <i32 0, i32 0, i32 2, i32 3, i32 6, i32 7, i32 4, i32 4>
@@ -760,13 +760,13 @@ define <8 x double> @shuffle_v8f64_00226644(<8 x double> %a, <8 x double> %b) {
 ;
 ; AVX512F-LABEL: shuffle_v8f64_00226644:
 ; AVX512F:       # BB#0:
-; AVX512F-NEXT:    vmovdqa64 {{.*#+}} zmm1 = [0,0,2,2,6,6,4,4]
+; AVX512F-NEXT:    vmovapd {{.*#+}} zmm1 = [0,0,2,2,6,6,4,4]
 ; AVX512F-NEXT:    vpermpd %zmm0, %zmm1, %zmm0
 ; AVX512F-NEXT:    retq
 ;
 ; AVX512F-32-LABEL: shuffle_v8f64_00226644:
 ; AVX512F-32:       # BB#0:
-; AVX512F-32-NEXT:    vmovdqa64 {{.*#+}} zmm1 = [0,0,0,0,2,0,2,0,6,0,6,0,4,0,4,0]
+; AVX512F-32-NEXT:    vmovapd {{.*#+}} zmm1 = [0,0,0,0,2,0,2,0,6,0,6,0,4,0,4,0]
 ; AVX512F-32-NEXT:    vpermpd %zmm0, %zmm1, %zmm0
 ; AVX512F-32-NEXT:    retl
   %shuffle = shufflevector <8 x double> %a, <8 x double> %b, <8 x i32> <i32 0, i32 0, i32 2, i32 2, i32 6, i32 6, i32 4, i32 4>
@@ -837,13 +837,13 @@ define <8 x double> @shuffle_v8f64_002u6u44(<8 x double> %a, <8 x double> %b) {
 ;
 ; AVX512F-LABEL: shuffle_v8f64_002u6u44:
 ; AVX512F:       # BB#0:
-; AVX512F-NEXT:    vmovdqa64 {{.*#+}} zmm1 = <0,0,2,u,6,u,4,4>
+; AVX512F-NEXT:    vmovapd {{.*#+}} zmm1 = <0,0,2,u,6,u,4,4>
 ; AVX512F-NEXT:    vpermpd %zmm0, %zmm1, %zmm0
 ; AVX512F-NEXT:    retq
 ;
 ; AVX512F-32-LABEL: shuffle_v8f64_002u6u44:
 ; AVX512F-32:       # BB#0:
-; AVX512F-32-NEXT:    vmovdqa64 {{.*#+}} zmm1 = <0,0,0,0,2,0,u,u,6,0,u,u,4,0,4,0>
+; AVX512F-32-NEXT:    vmovapd {{.*#+}} zmm1 = <0,0,0,0,2,0,u,u,6,0,u,u,4,0,4,0>
 ; AVX512F-32-NEXT:    vpermpd %zmm0, %zmm1, %zmm0
 ; AVX512F-32-NEXT:    retl
   %shuffle = shufflevector <8 x double> %a, <8 x double> %b, <8 x i32> <i32 0, i32 0, i32 2, i32 undef, i32 6, i32 undef, i32 4, i32 4>
@@ -854,13 +854,13 @@ define <8 x double> @shuffle_v8f64_00uu66uu(<8 x double> %a, <8 x double> %b) {
 ;
 ; AVX512F-LABEL: shuffle_v8f64_00uu66uu:
 ; AVX512F:       # BB#0:
-; AVX512F-NEXT:    vmovdqa64 {{.*#+}} zmm1 = <0,0,u,u,6,6,u,u>
+; AVX512F-NEXT:    vmovapd {{.*#+}} zmm1 = <0,0,u,u,6,6,u,u>
 ; AVX512F-NEXT:    vpermpd %zmm0, %zmm1, %zmm0
 ; AVX512F-NEXT:    retq
 ;
 ; AVX512F-32-LABEL: shuffle_v8f64_00uu66uu:
 ; AVX512F-32:       # BB#0:
-; AVX512F-32-NEXT:    vmovdqa64 {{.*#+}} zmm1 = <0,0,0,0,u,u,u,u,6,0,6,0,u,u,u,u>
+; AVX512F-32-NEXT:    vmovapd {{.*#+}} zmm1 = <0,0,0,0,u,u,u,u,6,0,6,0,u,u,u,u>
 ; AVX512F-32-NEXT:    vpermpd %zmm0, %zmm1, %zmm0
 ; AVX512F-32-NEXT:    retl
   %shuffle = shufflevector <8 x double> %a, <8 x double> %b, <8 x i32> <i32 0, i32 0, i32 undef, i32 undef, i32 6, i32 6, i32 undef, i32 undef>
diff --git a/llvm/test/CodeGen/X86/vector-shuffle-combining-avx512bw.ll b/llvm/test/CodeGen/X86/vector-shuffle-combining-avx512bw.ll
index 3e0c3e4..d57e1fe 100644
--- a/llvm/test/CodeGen/X86/vector-shuffle-combining-avx512bw.ll
+++ b/llvm/test/CodeGen/X86/vector-shuffle-combining-avx512bw.ll
@@ -29,11 +29,11 @@ define <8 x double> @combine_permvar_8f64_identity_mask(<8 x double> %x0, <8 x d
 ; CHECK-LABEL: combine_permvar_8f64_identity_mask:
 ; CHECK:       # BB#0:
 ; CHECK-NEXT:    kmovw %edi, %k1
-; CHECK-NEXT:    vmovdqa64 {{.*#+}} zmm2 = [7,6,5,4,3,2,1,0]
+; CHECK-NEXT:    vmovapd {{.*#+}} zmm2 = [7,6,5,4,3,2,1,0]
 ; CHECK-NEXT:    vpermpd %zmm0, %zmm2, %zmm1 {%k1}
-; CHECK-NEXT:    vmovdqa64 {{.*#+}} zmm0 = [7,14,5,12,3,10,1,8]
+; CHECK-NEXT:    vmovapd {{.*#+}} zmm0 = [7,14,5,12,3,10,1,8]
 ; CHECK-NEXT:    vpermpd %zmm1, %zmm0, %zmm1 {%k1}
-; CHECK-NEXT:    vmovdqa64 %zmm1, %zmm0
+; CHECK-NEXT:    vmovapd %zmm1, %zmm0
 ; CHECK-NEXT:    retq
   %res0 = call <8 x double> @llvm.x86.avx512.mask.permvar.df.512(<8 x double> %x0, <8 x i64> <i64 7, i64 6, i64 5, i64 4, i64 3, i64 2, i64 1, i64 0>, <8 x double> %x1, i8 %m)
   %res1 = call <8 x double> @llvm.x86.avx512.mask.permvar.df.512(<8 x double> %res0, <8 x i64> <i64 7, i64 14, i64 5, i64 12, i64 3, i64 10, i64 1, i64 8>, <8 x double> %res0, i8 %m)
@@ -433,7 +433,7 @@ define <8 x double> @combine_permvar_8f64_as_permpd_mask(<8 x double> %x0, <8 x
 ; CHECK:       # BB#0:
 ; CHECK-NEXT:    kmovw %edi, %k1
 ; CHECK-NEXT:    vpermpd {{.*#+}} zmm1 {%k1} = zmm0[3,2,1,0,7,6,5,4]
-; CHECK-NEXT:    vmovdqa64 %zmm1, %zmm0
+; CHECK-NEXT:    vmovapd %zmm1, %zmm0
 ; CHECK-NEXT:    retq
   %1 = call <8 x double> @llvm.x86.avx512.mask.permvar.df.512(<8 x double> %x0, <8 x i64> <i64 3, i64 2, i64 1, i64 undef, i64 undef, i64 6, i64 5, i64 4>, <8 x double> %x1, i8 %m)
   ret <8 x double> %1