[X86][AVX512VL] Added AVX512VL to 128/256 bit vector shift tests

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@291488 91177308-0d34-0410-b5e6-96231b3b80d8

diff --git a/test/CodeGen/X86/vector-shift-ashr-128.ll b/test/CodeGen/X86/vector-shift-ashr-128.ll
index 440faa689fb8aaaeb42111762f467ab746374955..9f0d4a7d7264a97e2fa73d19970215d91dc30328 100644 (file)
--- a/test/CodeGen/X86/vector-shift-ashr-128.ll
+++ b/test/CodeGen/X86/vector-shift-ashr-128.ll
@@ -7,7 +7,8 @@
 ; RUN: llc < %s -mtriple=x86_64-unknown-unknown -mattr=+xop,+avx2 | FileCheck %s --check-prefix=ALL --check-prefix=XOP --check-prefix=XOPAVX2
 ; RUN: llc < %s -mtriple=x86_64-unknown-unknown -mcpu=knl -mattr=+avx512dq | FileCheck %s --check-prefix=ALL --check-prefix=AVX512 --check-prefix=AVX512DQ
 ; RUN: llc < %s -mtriple=x86_64-unknown-unknown -mcpu=knl -mattr=+avx512bw | FileCheck %s --check-prefix=ALL --check-prefix=AVX512 --check-prefix=AVX512BW
-
+; RUN: llc < %s -mtriple=x86_64-unknown-unknown -mcpu=knl -mattr=+avx512dq,+avx512vl | FileCheck %s --check-prefix=ALL --check-prefix=AVX512VL --check-prefix=AVX512DQVL
+; RUN: llc < %s -mtriple=x86_64-unknown-unknown -mcpu=knl -mattr=+avx512bw,+avx512vl | FileCheck %s --check-prefix=ALL --check-prefix=AVX512VL --check-prefix=AVX512BWVL
 ;
 ; Just one 32-bit run to make sure we do reasonable things for i64 shifts.
 ; RUN: llc < %s -mtriple=i686-unknown-unknown -mattr=+sse2 | FileCheck %s --check-prefix=ALL --check-prefix=X32-SSE --check-prefix=X32-SSE2
@@ -89,6 +90,11 @@ define <2 x i64> @var_shift_v2i64(<2 x i64> %a, <2 x i64> %b) nounwind {
 ; AVX512-NEXT:    vpsubq %xmm3, %xmm0, %xmm0
 ; AVX512-NEXT:    retq
 ;
+; AVX512VL-LABEL: var_shift_v2i64:
+; AVX512VL:       # BB#0:
+; AVX512VL-NEXT:    vpsravq %xmm1, %xmm0, %xmm0
+; AVX512VL-NEXT:    retq
+;
 ; X32-SSE-LABEL: var_shift_v2i64:
 ; X32-SSE:       # BB#0:
 ; X32-SSE-NEXT:    pshufd {{.*#+}} xmm3 = xmm1[2,3,0,1]
@@ -193,6 +199,11 @@ define <4 x i32> @var_shift_v4i32(<4 x i32> %a, <4 x i32> %b) nounwind {
 ; AVX512-NEXT:    vpsravd %xmm1, %xmm0, %xmm0
 ; AVX512-NEXT:    retq
 ;
+; AVX512VL-LABEL: var_shift_v4i32:
+; AVX512VL:       # BB#0:
+; AVX512VL-NEXT:    vpsravd %xmm1, %xmm0, %xmm0
+; AVX512VL-NEXT:    retq
+;
 ; X32-SSE-LABEL: var_shift_v4i32:
 ; X32-SSE:       # BB#0:
 ; X32-SSE-NEXT:    movdqa %xmm1, %xmm2
@@ -339,6 +350,19 @@ define <8 x i16> @var_shift_v8i16(<8 x i16> %a, <8 x i16> %b) nounwind {
 ; AVX512BW-NEXT:    # kill: %XMM0<def> %XMM0<kill> %ZMM0<kill>
 ; AVX512BW-NEXT:    retq
 ;
+; AVX512DQVL-LABEL: var_shift_v8i16:
+; AVX512DQVL:       # BB#0:
+; AVX512DQVL-NEXT:    vpmovzxwd {{.*#+}} ymm1 = xmm1[0],zero,xmm1[1],zero,xmm1[2],zero,xmm1[3],zero,xmm1[4],zero,xmm1[5],zero,xmm1[6],zero,xmm1[7],zero
+; AVX512DQVL-NEXT:    vpmovsxwd %xmm0, %ymm0
+; AVX512DQVL-NEXT:    vpsravd %ymm1, %ymm0, %ymm0
+; AVX512DQVL-NEXT:    vpmovdw %ymm0, %xmm0
+; AVX512DQVL-NEXT:    retq
+;
+; AVX512BWVL-LABEL: var_shift_v8i16:
+; AVX512BWVL:       # BB#0:
+; AVX512BWVL-NEXT:    vpsravw %xmm1, %xmm0, %xmm0
+; AVX512BWVL-NEXT:    retq
+;
 ; X32-SSE-LABEL: var_shift_v8i16:
 ; X32-SSE:       # BB#0:
 ; X32-SSE-NEXT:    psllw $12, %xmm1
@@ -515,6 +539,14 @@ define <16 x i8> @var_shift_v16i8(<16 x i8> %a, <16 x i8> %b) nounwind {
 ; AVX512-NEXT:    vpmovdb %zmm0, %xmm0
 ; AVX512-NEXT:    retq
 ;
+; AVX512VL-LABEL: var_shift_v16i8:
+; AVX512VL:       # BB#0:
+; AVX512VL-NEXT:    vpmovzxbd {{.*#+}} zmm1 = xmm1[0],zero,zero,zero,xmm1[1],zero,zero,zero,xmm1[2],zero,zero,zero,xmm1[3],zero,zero,zero,xmm1[4],zero,zero,zero,xmm1[5],zero,zero,zero,xmm1[6],zero,zero,zero,xmm1[7],zero,zero,zero,xmm1[8],zero,zero,zero,xmm1[9],zero,zero,zero,xmm1[10],zero,zero,zero,xmm1[11],zero,zero,zero,xmm1[12],zero,zero,zero,xmm1[13],zero,zero,zero,xmm1[14],zero,zero,zero,xmm1[15],zero,zero,zero
+; AVX512VL-NEXT:    vpmovsxbd %xmm0, %zmm0
+; AVX512VL-NEXT:    vpsravd %zmm1, %zmm0, %zmm0
+; AVX512VL-NEXT:    vpmovdb %zmm0, %xmm0
+; AVX512VL-NEXT:    retq
+;
 ; X32-SSE-LABEL: var_shift_v16i8:
 ; X32-SSE:       # BB#0:
 ; X32-SSE-NEXT:    punpckhbw {{.*#+}} xmm2 = xmm2[8],xmm0[8],xmm2[9],xmm0[9],xmm2[10],xmm0[10],xmm2[11],xmm0[11],xmm2[12],xmm0[12],xmm2[13],xmm0[13],xmm2[14],xmm0[14],xmm2[15],xmm0[15]
@@ -624,6 +656,11 @@ define <2 x i64> @splatvar_shift_v2i64(<2 x i64> %a, <2 x i64> %b) nounwind {
 ; AVX512-NEXT:    vpsubq %xmm2, %xmm0, %xmm0
 ; AVX512-NEXT:    retq
 ;
+; AVX512VL-LABEL: splatvar_shift_v2i64:
+; AVX512VL:       # BB#0:
+; AVX512VL-NEXT:    vpsraq %xmm1, %xmm0, %xmm0
+; AVX512VL-NEXT:    retq
+;
 ; X32-SSE-LABEL: splatvar_shift_v2i64:
 ; X32-SSE:       # BB#0:
 ; X32-SSE-NEXT:    movdqa {{.*#+}} xmm2 = [0,2147483648,0,2147483648]
@@ -669,6 +706,12 @@ define <4 x i32> @splatvar_shift_v4i32(<4 x i32> %a, <4 x i32> %b) nounwind {
 ; AVX512-NEXT:    vpsrad %xmm1, %xmm0, %xmm0
 ; AVX512-NEXT:    retq
 ;
+; AVX512VL-LABEL: splatvar_shift_v4i32:
+; AVX512VL:       # BB#0:
+; AVX512VL-NEXT:    vpmovzxdq {{.*#+}} xmm1 = xmm1[0],zero,xmm1[1],zero
+; AVX512VL-NEXT:    vpsrad %xmm1, %xmm0, %xmm0
+; AVX512VL-NEXT:    retq
+;
 ; X32-SSE-LABEL: splatvar_shift_v4i32:
 ; X32-SSE:       # BB#0:
 ; X32-SSE-NEXT:    xorps %xmm2, %xmm2
@@ -712,6 +755,12 @@ define <8 x i16> @splatvar_shift_v8i16(<8 x i16> %a, <8 x i16> %b) nounwind {
 ; AVX512-NEXT:    vpsraw %xmm1, %xmm0, %xmm0
 ; AVX512-NEXT:    retq
 ;
+; AVX512VL-LABEL: splatvar_shift_v8i16:
+; AVX512VL:       # BB#0:
+; AVX512VL-NEXT:    vpmovzxwq {{.*#+}} xmm1 = xmm1[0],zero,zero,zero,xmm1[1],zero,zero,zero
+; AVX512VL-NEXT:    vpsraw %xmm1, %xmm0, %xmm0
+; AVX512VL-NEXT:    retq
+;
 ; X32-SSE-LABEL: splatvar_shift_v8i16:
 ; X32-SSE:       # BB#0:
 ; X32-SSE-NEXT:    pextrw $0, %xmm1, %eax
@@ -907,6 +956,15 @@ define <16 x i8> @splatvar_shift_v16i8(<16 x i8> %a, <16 x i8> %b) nounwind {
 ; AVX512-NEXT:    vpmovdb %zmm0, %xmm0
 ; AVX512-NEXT:    retq
 ;
+; AVX512VL-LABEL: splatvar_shift_v16i8:
+; AVX512VL:       # BB#0:
+; AVX512VL-NEXT:    vpbroadcastb %xmm1, %xmm1
+; AVX512VL-NEXT:    vpmovsxbd %xmm0, %zmm0
+; AVX512VL-NEXT:    vpmovzxbd {{.*#+}} zmm1 = xmm1[0],zero,zero,zero,xmm1[1],zero,zero,zero,xmm1[2],zero,zero,zero,xmm1[3],zero,zero,zero,xmm1[4],zero,zero,zero,xmm1[5],zero,zero,zero,xmm1[6],zero,zero,zero,xmm1[7],zero,zero,zero,xmm1[8],zero,zero,zero,xmm1[9],zero,zero,zero,xmm1[10],zero,zero,zero,xmm1[11],zero,zero,zero,xmm1[12],zero,zero,zero,xmm1[13],zero,zero,zero,xmm1[14],zero,zero,zero,xmm1[15],zero,zero,zero
+; AVX512VL-NEXT:    vpsravd %zmm1, %zmm0, %zmm0
+; AVX512VL-NEXT:    vpmovdb %zmm0, %xmm0
+; AVX512VL-NEXT:    retq
+;
 ; X32-SSE-LABEL: splatvar_shift_v16i8:
 ; X32-SSE:       # BB#0:
 ; X32-SSE-NEXT:    punpcklbw {{.*#+}} xmm1 = xmm1[0,0,1,1,2,2,3,3,4,4,5,5,6,6,7,7]
@@ -1033,6 +1091,11 @@ define <2 x i64> @constant_shift_v2i64(<2 x i64> %a) nounwind {
 ; AVX512-NEXT:    vpsubq %xmm1, %xmm0, %xmm0
 ; AVX512-NEXT:    retq
 ;
+; AVX512VL-LABEL: constant_shift_v2i64:
+; AVX512VL:       # BB#0:
+; AVX512VL-NEXT:    vpsravq {{.*}}(%rip), %xmm0, %xmm0
+; AVX512VL-NEXT:    retq
+;
 ; X32-SSE-LABEL: constant_shift_v2i64:
 ; X32-SSE:       # BB#0:
 ; X32-SSE-NEXT:    movdqa {{.*#+}} xmm1 = [0,2147483648,0,2147483648]
@@ -1114,6 +1177,11 @@ define <4 x i32> @constant_shift_v4i32(<4 x i32> %a) nounwind {
 ; AVX512-NEXT:    vpsravd {{.*}}(%rip), %xmm0, %xmm0
 ; AVX512-NEXT:    retq
 ;
+; AVX512VL-LABEL: constant_shift_v4i32:
+; AVX512VL:       # BB#0:
+; AVX512VL-NEXT:    vpsravd {{.*}}(%rip), %xmm0, %xmm0
+; AVX512VL-NEXT:    retq
+;
 ; X32-SSE-LABEL: constant_shift_v4i32:
 ; X32-SSE:       # BB#0:
 ; X32-SSE-NEXT:    movdqa %xmm0, %xmm1
@@ -1207,6 +1275,18 @@ define <8 x i16> @constant_shift_v8i16(<8 x i16> %a) nounwind {
 ; AVX512BW-NEXT:    # kill: %XMM0<def> %XMM0<kill> %ZMM0<kill>
 ; AVX512BW-NEXT:    retq
 ;
+; AVX512DQVL-LABEL: constant_shift_v8i16:
+; AVX512DQVL:       # BB#0:
+; AVX512DQVL-NEXT:    vpmovsxwd %xmm0, %ymm0
+; AVX512DQVL-NEXT:    vpsravd {{.*}}(%rip), %ymm0, %ymm0
+; AVX512DQVL-NEXT:    vpmovdw %ymm0, %xmm0
+; AVX512DQVL-NEXT:    retq
+;
+; AVX512BWVL-LABEL: constant_shift_v8i16:
+; AVX512BWVL:       # BB#0:
+; AVX512BWVL-NEXT:    vpsravw {{.*}}(%rip), %xmm0, %xmm0
+; AVX512BWVL-NEXT:    retq
+;
 ; X32-SSE-LABEL: constant_shift_v8i16:
 ; X32-SSE:       # BB#0:
 ; X32-SSE-NEXT:    movdqa %xmm0, %xmm1
@@ -1367,6 +1447,13 @@ define <16 x i8> @constant_shift_v16i8(<16 x i8> %a) nounwind {
 ; AVX512-NEXT:    vpmovdb %zmm0, %xmm0
 ; AVX512-NEXT:    retq
 ;
+; AVX512VL-LABEL: constant_shift_v16i8:
+; AVX512VL:       # BB#0:
+; AVX512VL-NEXT:    vpmovsxbd %xmm0, %zmm0
+; AVX512VL-NEXT:    vpsravd {{.*}}(%rip), %zmm0, %zmm0
+; AVX512VL-NEXT:    vpmovdb %zmm0, %xmm0
+; AVX512VL-NEXT:    retq
+;
 ; X32-SSE-LABEL: constant_shift_v16i8:
 ; X32-SSE:       # BB#0:
 ; X32-SSE-NEXT:    punpckhbw {{.*#+}} xmm1 = xmm1[8],xmm0[8],xmm1[9],xmm0[9],xmm1[10],xmm0[10],xmm1[11],xmm0[11],xmm1[12],xmm0[12],xmm1[13],xmm0[13],xmm1[14],xmm0[14],xmm1[15],xmm0[15]
@@ -1480,6 +1567,11 @@ define <2 x i64> @splatconstant_shift_v2i64(<2 x i64> %a) nounwind {
 ; AVX512-NEXT:    vpblendd {{.*#+}} xmm0 = xmm0[0],xmm1[1],xmm0[2],xmm1[3]
 ; AVX512-NEXT:    retq
 ;
+; AVX512VL-LABEL: splatconstant_shift_v2i64:
+; AVX512VL:       # BB#0:
+; AVX512VL-NEXT:    vpsraq $7, %xmm0, %xmm0
+; AVX512VL-NEXT:    retq
+;
 ; X32-SSE-LABEL: splatconstant_shift_v2i64:
 ; X32-SSE:       # BB#0:
 ; X32-SSE-NEXT:    movdqa %xmm0, %xmm1
@@ -1514,6 +1606,11 @@ define <4 x i32> @splatconstant_shift_v4i32(<4 x i32> %a) nounwind {
 ; AVX512-NEXT:    vpsrad $5, %xmm0, %xmm0
 ; AVX512-NEXT:    retq
 ;
+; AVX512VL-LABEL: splatconstant_shift_v4i32:
+; AVX512VL:       # BB#0:
+; AVX512VL-NEXT:    vpsrad $5, %xmm0, %xmm0
+; AVX512VL-NEXT:    retq
+;
 ; X32-SSE-LABEL: splatconstant_shift_v4i32:
 ; X32-SSE:       # BB#0:
 ; X32-SSE-NEXT:    psrad $5, %xmm0
@@ -1543,6 +1640,11 @@ define <8 x i16> @splatconstant_shift_v8i16(<8 x i16> %a) nounwind {
 ; AVX512-NEXT:    vpsraw $3, %xmm0, %xmm0
 ; AVX512-NEXT:    retq
 ;
+; AVX512VL-LABEL: splatconstant_shift_v8i16:
+; AVX512VL:       # BB#0:
+; AVX512VL-NEXT:    vpsraw $3, %xmm0, %xmm0
+; AVX512VL-NEXT:    retq
+;
 ; X32-SSE-LABEL: splatconstant_shift_v8i16:
 ; X32-SSE:       # BB#0:
 ; X32-SSE-NEXT:    psraw $3, %xmm0
@@ -1586,6 +1688,15 @@ define <16 x i8> @splatconstant_shift_v16i8(<16 x i8> %a) nounwind {
 ; AVX512-NEXT:    vpsubb %xmm1, %xmm0, %xmm0
 ; AVX512-NEXT:    retq
 ;
+; AVX512VL-LABEL: splatconstant_shift_v16i8:
+; AVX512VL:       # BB#0:
+; AVX512VL-NEXT:    vpsrlw $3, %xmm0, %xmm0
+; AVX512VL-NEXT:    vpand {{.*}}(%rip), %xmm0, %xmm0
+; AVX512VL-NEXT:    vmovdqa {{.*#+}} xmm1 = [16,16,16,16,16,16,16,16,16,16,16,16,16,16,16,16]
+; AVX512VL-NEXT:    vpxor %xmm1, %xmm0, %xmm0
+; AVX512VL-NEXT:    vpsubb %xmm1, %xmm0, %xmm0
+; AVX512VL-NEXT:    retq
+;
 ; X32-SSE-LABEL: splatconstant_shift_v16i8:
 ; X32-SSE:       # BB#0:
 ; X32-SSE-NEXT:    psrlw $3, %xmm0

diff --git a/test/CodeGen/X86/vector-shift-ashr-256.ll b/test/CodeGen/X86/vector-shift-ashr-256.ll
index 79902acfec24c6ae9a72ca98961d7b7486d2a6da..aee2857157b6ba6739c6bc15d40efee975127960 100644 (file)
--- a/test/CodeGen/X86/vector-shift-ashr-256.ll
+++ b/test/CodeGen/X86/vector-shift-ashr-256.ll
@@ -5,6 +5,9 @@
 ; RUN: llc < %s -mtriple=x86_64-unknown-unknown -mattr=+xop,+avx2 | FileCheck %s --check-prefix=ALL --check-prefix=XOP --check-prefix=XOPAVX2
 ; RUN: llc < %s -mtriple=x86_64-unknown-unknown -mcpu=knl -mattr=+avx512dq | FileCheck %s --check-prefix=ALL --check-prefix=AVX512 --check-prefix=AVX512DQ
 ; RUN: llc < %s -mtriple=x86_64-unknown-unknown -mcpu=knl -mattr=+avx512bw | FileCheck %s --check-prefix=ALL --check-prefix=AVX512 --check-prefix=AVX512BW
+; RUN: llc < %s -mtriple=x86_64-unknown-unknown -mcpu=knl -mattr=+avx512dq,+avx512vl | FileCheck %s --check-prefix=ALL --check-prefix=AVX512VL --check-prefix=AVX512DQVL
+; RUN: llc < %s -mtriple=x86_64-unknown-unknown -mcpu=knl -mattr=+avx512bw,+avx512vl | FileCheck %s --check-prefix=ALL --check-prefix=AVX512VL --check-prefix=AVX512BWVL
+
 ;
 ; Variable Shifts
 ;
@@ -74,6 +77,11 @@ define <4 x i64> @var_shift_v4i64(<4 x i64> %a, <4 x i64> %b) nounwind {
 ; AVX512-NEXT:    vpsrlvq %ymm1, %ymm0, %ymm0
 ; AVX512-NEXT:    vpsubq %ymm3, %ymm0, %ymm0
 ; AVX512-NEXT:    retq
+;
+; AVX512VL-LABEL: var_shift_v4i64:
+; AVX512VL:       # BB#0:
+; AVX512VL-NEXT:    vpsravq %ymm1, %ymm0, %ymm0
+; AVX512VL-NEXT:    retq
   %shift = ashr <4 x i64> %a, %b
   ret <4 x i64> %shift
 }
@@ -135,6 +143,11 @@ define <8 x i32> @var_shift_v8i32(<8 x i32> %a, <8 x i32> %b) nounwind {
 ; AVX512:       # BB#0:
 ; AVX512-NEXT:    vpsravd %ymm1, %ymm0, %ymm0
 ; AVX512-NEXT:    retq
+;
+; AVX512VL-LABEL: var_shift_v8i32:
+; AVX512VL:       # BB#0:
+; AVX512VL-NEXT:    vpsravd %ymm1, %ymm0, %ymm0
+; AVX512VL-NEXT:    retq
   %shift = ashr <8 x i32> %a, %b
   ret <8 x i32> %shift
 }
@@ -228,6 +241,19 @@ define <16 x i16> @var_shift_v16i16(<16 x i16> %a, <16 x i16> %b) nounwind {
 ; AVX512BW-NEXT:    vpsravw %zmm1, %zmm0, %zmm0
 ; AVX512BW-NEXT:    # kill: %YMM0<def> %YMM0<kill> %ZMM0<kill>
 ; AVX512BW-NEXT:    retq
+;
+; AVX512DQVL-LABEL: var_shift_v16i16:
+; AVX512DQVL:       # BB#0:
+; AVX512DQVL-NEXT:    vpmovzxwd {{.*#+}} zmm1 = ymm1[0],zero,ymm1[1],zero,ymm1[2],zero,ymm1[3],zero,ymm1[4],zero,ymm1[5],zero,ymm1[6],zero,ymm1[7],zero,ymm1[8],zero,ymm1[9],zero,ymm1[10],zero,ymm1[11],zero,ymm1[12],zero,ymm1[13],zero,ymm1[14],zero,ymm1[15],zero
+; AVX512DQVL-NEXT:    vpmovsxwd %ymm0, %zmm0
+; AVX512DQVL-NEXT:    vpsravd %zmm1, %zmm0, %zmm0
+; AVX512DQVL-NEXT:    vpmovdw %zmm0, %ymm0
+; AVX512DQVL-NEXT:    retq
+;
+; AVX512BWVL-LABEL: var_shift_v16i16:
+; AVX512BWVL:       # BB#0:
+; AVX512BWVL-NEXT:    vpsravw %ymm1, %ymm0, %ymm0
+; AVX512BWVL-NEXT:    retq
   %shift = ashr <16 x i16> %a, %b
   ret <16 x i16> %shift
 }
@@ -375,6 +401,42 @@ define <32 x i8> @var_shift_v32i8(<32 x i8> %a, <32 x i8> %b) nounwind {
 ; AVX512BW-NEXT:    vpsravw %zmm1, %zmm0, %zmm0
 ; AVX512BW-NEXT:    vpmovwb %zmm0, %ymm0
 ; AVX512BW-NEXT:    retq
+;
+; AVX512DQVL-LABEL: var_shift_v32i8:
+; AVX512DQVL:       # BB#0:
+; AVX512DQVL-NEXT:    vpsllw $5, %ymm1, %ymm1
+; AVX512DQVL-NEXT:    vpunpckhbw {{.*#+}} ymm2 = ymm0[8],ymm1[8],ymm0[9],ymm1[9],ymm0[10],ymm1[10],ymm0[11],ymm1[11],ymm0[12],ymm1[12],ymm0[13],ymm1[13],ymm0[14],ymm1[14],ymm0[15],ymm1[15],ymm0[24],ymm1[24],ymm0[25],ymm1[25],ymm0[26],ymm1[26],ymm0[27],ymm1[27],ymm0[28],ymm1[28],ymm0[29],ymm1[29],ymm0[30],ymm1[30],ymm0[31],ymm1[31]
+; AVX512DQVL-NEXT:    vpunpckhbw {{.*#+}} ymm3 = ymm0[8,8,9,9,10,10,11,11,12,12,13,13,14,14,15,15,24,24,25,25,26,26,27,27,28,28,29,29,30,30,31,31]
+; AVX512DQVL-NEXT:    vpsraw $4, %ymm3, %ymm4
+; AVX512DQVL-NEXT:    vpblendvb %ymm2, %ymm4, %ymm3, %ymm3
+; AVX512DQVL-NEXT:    vpsraw $2, %ymm3, %ymm4
+; AVX512DQVL-NEXT:    vpaddw %ymm2, %ymm2, %ymm2
+; AVX512DQVL-NEXT:    vpblendvb %ymm2, %ymm4, %ymm3, %ymm3
+; AVX512DQVL-NEXT:    vpsraw $1, %ymm3, %ymm4
+; AVX512DQVL-NEXT:    vpaddw %ymm2, %ymm2, %ymm2
+; AVX512DQVL-NEXT:    vpblendvb %ymm2, %ymm4, %ymm3, %ymm2
+; AVX512DQVL-NEXT:    vpsrlw $8, %ymm2, %ymm2
+; AVX512DQVL-NEXT:    vpunpcklbw {{.*#+}} ymm1 = ymm0[0],ymm1[0],ymm0[1],ymm1[1],ymm0[2],ymm1[2],ymm0[3],ymm1[3],ymm0[4],ymm1[4],ymm0[5],ymm1[5],ymm0[6],ymm1[6],ymm0[7],ymm1[7],ymm0[16],ymm1[16],ymm0[17],ymm1[17],ymm0[18],ymm1[18],ymm0[19],ymm1[19],ymm0[20],ymm1[20],ymm0[21],ymm1[21],ymm0[22],ymm1[22],ymm0[23],ymm1[23]
+; AVX512DQVL-NEXT:    vpunpcklbw {{.*#+}} ymm0 = ymm0[0,0,1,1,2,2,3,3,4,4,5,5,6,6,7,7,16,16,17,17,18,18,19,19,20,20,21,21,22,22,23,23]
+; AVX512DQVL-NEXT:    vpsraw $4, %ymm0, %ymm3
+; AVX512DQVL-NEXT:    vpblendvb %ymm1, %ymm3, %ymm0, %ymm0
+; AVX512DQVL-NEXT:    vpsraw $2, %ymm0, %ymm3
+; AVX512DQVL-NEXT:    vpaddw %ymm1, %ymm1, %ymm1
+; AVX512DQVL-NEXT:    vpblendvb %ymm1, %ymm3, %ymm0, %ymm0
+; AVX512DQVL-NEXT:    vpsraw $1, %ymm0, %ymm3
+; AVX512DQVL-NEXT:    vpaddw %ymm1, %ymm1, %ymm1
+; AVX512DQVL-NEXT:    vpblendvb %ymm1, %ymm3, %ymm0, %ymm0
+; AVX512DQVL-NEXT:    vpsrlw $8, %ymm0, %ymm0
+; AVX512DQVL-NEXT:    vpackuswb %ymm2, %ymm0, %ymm0
+; AVX512DQVL-NEXT:    retq
+;
+; AVX512BWVL-LABEL: var_shift_v32i8:
+; AVX512BWVL:       # BB#0:
+; AVX512BWVL-NEXT:    vpmovzxbw {{.*#+}} zmm1 = ymm1[0],zero,ymm1[1],zero,ymm1[2],zero,ymm1[3],zero,ymm1[4],zero,ymm1[5],zero,ymm1[6],zero,ymm1[7],zero,ymm1[8],zero,ymm1[9],zero,ymm1[10],zero,ymm1[11],zero,ymm1[12],zero,ymm1[13],zero,ymm1[14],zero,ymm1[15],zero,ymm1[16],zero,ymm1[17],zero,ymm1[18],zero,ymm1[19],zero,ymm1[20],zero,ymm1[21],zero,ymm1[22],zero,ymm1[23],zero,ymm1[24],zero,ymm1[25],zero,ymm1[26],zero,ymm1[27],zero,ymm1[28],zero,ymm1[29],zero,ymm1[30],zero,ymm1[31],zero
+; AVX512BWVL-NEXT:    vpmovsxbw %ymm0, %zmm0
+; AVX512BWVL-NEXT:    vpsravw %zmm1, %zmm0, %zmm0
+; AVX512BWVL-NEXT:    vpmovwb %zmm0, %ymm0
+; AVX512BWVL-NEXT:    retq
   %shift = ashr <32 x i8> %a, %b
   ret <32 x i8> %shift
 }
@@ -435,6 +497,11 @@ define <4 x i64> @splatvar_shift_v4i64(<4 x i64> %a, <4 x i64> %b) nounwind {
 ; AVX512-NEXT:    vpxor %ymm2, %ymm0, %ymm0
 ; AVX512-NEXT:    vpsubq %ymm2, %ymm0, %ymm0
 ; AVX512-NEXT:    retq
+;
+; AVX512VL-LABEL: splatvar_shift_v4i64:
+; AVX512VL:       # BB#0:
+; AVX512VL-NEXT:    vpsraq %xmm1, %ymm0, %ymm0
+; AVX512VL-NEXT:    retq
   %splat = shufflevector <4 x i64> %b, <4 x i64> undef, <4 x i32> zeroinitializer
   %shift = ashr <4 x i64> %a, %splat
   ret <4 x i64> %shift
@@ -476,6 +543,12 @@ define <8 x i32> @splatvar_shift_v8i32(<8 x i32> %a, <8 x i32> %b) nounwind {
 ; AVX512-NEXT:    vpmovzxdq {{.*#+}} xmm1 = xmm1[0],zero,xmm1[1],zero
 ; AVX512-NEXT:    vpsrad %xmm1, %ymm0, %ymm0
 ; AVX512-NEXT:    retq
+;
+; AVX512VL-LABEL: splatvar_shift_v8i32:
+; AVX512VL:       # BB#0:
+; AVX512VL-NEXT:    vpmovzxdq {{.*#+}} xmm1 = xmm1[0],zero,xmm1[1],zero
+; AVX512VL-NEXT:    vpsrad %xmm1, %ymm0, %ymm0
+; AVX512VL-NEXT:    retq
   %splat = shufflevector <8 x i32> %b, <8 x i32> undef, <8 x i32> zeroinitializer
   %shift = ashr <8 x i32> %a, %splat
   ret <8 x i32> %shift
@@ -517,6 +590,12 @@ define <16 x i16> @splatvar_shift_v16i16(<16 x i16> %a, <16 x i16> %b) nounwind
 ; AVX512-NEXT:    vpmovzxwq {{.*#+}} xmm1 = xmm1[0],zero,zero,zero,xmm1[1],zero,zero,zero
 ; AVX512-NEXT:    vpsraw %xmm1, %ymm0, %ymm0
 ; AVX512-NEXT:    retq
+;
+; AVX512VL-LABEL: splatvar_shift_v16i16:
+; AVX512VL:       # BB#0:
+; AVX512VL-NEXT:    vpmovzxwq {{.*#+}} xmm1 = xmm1[0],zero,zero,zero,xmm1[1],zero,zero,zero
+; AVX512VL-NEXT:    vpsraw %xmm1, %ymm0, %ymm0
+; AVX512VL-NEXT:    retq
   %splat = shufflevector <16 x i16> %b, <16 x i16> undef, <16 x i32> zeroinitializer
   %shift = ashr <16 x i16> %a, %splat
   ret <16 x i16> %shift
@@ -662,6 +741,44 @@ define <32 x i8> @splatvar_shift_v32i8(<32 x i8> %a, <32 x i8> %b) nounwind {
 ; AVX512BW-NEXT:    vpsravw %zmm1, %zmm0, %zmm0
 ; AVX512BW-NEXT:    vpmovwb %zmm0, %ymm0
 ; AVX512BW-NEXT:    retq
+;
+; AVX512DQVL-LABEL: splatvar_shift_v32i8:
+; AVX512DQVL:       # BB#0:
+; AVX512DQVL-NEXT:    vpbroadcastb %xmm1, %ymm1
+; AVX512DQVL-NEXT:    vpsllw $5, %ymm1, %ymm1
+; AVX512DQVL-NEXT:    vpunpckhbw {{.*#+}} ymm2 = ymm0[8],ymm1[8],ymm0[9],ymm1[9],ymm0[10],ymm1[10],ymm0[11],ymm1[11],ymm0[12],ymm1[12],ymm0[13],ymm1[13],ymm0[14],ymm1[14],ymm0[15],ymm1[15],ymm0[24],ymm1[24],ymm0[25],ymm1[25],ymm0[26],ymm1[26],ymm0[27],ymm1[27],ymm0[28],ymm1[28],ymm0[29],ymm1[29],ymm0[30],ymm1[30],ymm0[31],ymm1[31]
+; AVX512DQVL-NEXT:    vpunpckhbw {{.*#+}} ymm3 = ymm0[8,8,9,9,10,10,11,11,12,12,13,13,14,14,15,15,24,24,25,25,26,26,27,27,28,28,29,29,30,30,31,31]
+; AVX512DQVL-NEXT:    vpsraw $4, %ymm3, %ymm4
+; AVX512DQVL-NEXT:    vpblendvb %ymm2, %ymm4, %ymm3, %ymm3
+; AVX512DQVL-NEXT:    vpsraw $2, %ymm3, %ymm4
+; AVX512DQVL-NEXT:    vpaddw %ymm2, %ymm2, %ymm2
+; AVX512DQVL-NEXT:    vpblendvb %ymm2, %ymm4, %ymm3, %ymm3
+; AVX512DQVL-NEXT:    vpsraw $1, %ymm3, %ymm4
+; AVX512DQVL-NEXT:    vpaddw %ymm2, %ymm2, %ymm2
+; AVX512DQVL-NEXT:    vpblendvb %ymm2, %ymm4, %ymm3, %ymm2
+; AVX512DQVL-NEXT:    vpsrlw $8, %ymm2, %ymm2
+; AVX512DQVL-NEXT:    vpunpcklbw {{.*#+}} ymm1 = ymm0[0],ymm1[0],ymm0[1],ymm1[1],ymm0[2],ymm1[2],ymm0[3],ymm1[3],ymm0[4],ymm1[4],ymm0[5],ymm1[5],ymm0[6],ymm1[6],ymm0[7],ymm1[7],ymm0[16],ymm1[16],ymm0[17],ymm1[17],ymm0[18],ymm1[18],ymm0[19],ymm1[19],ymm0[20],ymm1[20],ymm0[21],ymm1[21],ymm0[22],ymm1[22],ymm0[23],ymm1[23]
+; AVX512DQVL-NEXT:    vpunpcklbw {{.*#+}} ymm0 = ymm0[0,0,1,1,2,2,3,3,4,4,5,5,6,6,7,7,16,16,17,17,18,18,19,19,20,20,21,21,22,22,23,23]
+; AVX512DQVL-NEXT:    vpsraw $4, %ymm0, %ymm3
+; AVX512DQVL-NEXT:    vpblendvb %ymm1, %ymm3, %ymm0, %ymm0
+; AVX512DQVL-NEXT:    vpsraw $2, %ymm0, %ymm3
+; AVX512DQVL-NEXT:    vpaddw %ymm1, %ymm1, %ymm1
+; AVX512DQVL-NEXT:    vpblendvb %ymm1, %ymm3, %ymm0, %ymm0
+; AVX512DQVL-NEXT:    vpsraw $1, %ymm0, %ymm3
+; AVX512DQVL-NEXT:    vpaddw %ymm1, %ymm1, %ymm1
+; AVX512DQVL-NEXT:    vpblendvb %ymm1, %ymm3, %ymm0, %ymm0
+; AVX512DQVL-NEXT:    vpsrlw $8, %ymm0, %ymm0
+; AVX512DQVL-NEXT:    vpackuswb %ymm2, %ymm0, %ymm0
+; AVX512DQVL-NEXT:    retq
+;
+; AVX512BWVL-LABEL: splatvar_shift_v32i8:
+; AVX512BWVL:       # BB#0:
+; AVX512BWVL-NEXT:    vpbroadcastb %xmm1, %ymm1
+; AVX512BWVL-NEXT:    vpmovsxbw %ymm0, %zmm0
+; AVX512BWVL-NEXT:    vpmovzxbw {{.*#+}} zmm1 = ymm1[0],zero,ymm1[1],zero,ymm1[2],zero,ymm1[3],zero,ymm1[4],zero,ymm1[5],zero,ymm1[6],zero,ymm1[7],zero,ymm1[8],zero,ymm1[9],zero,ymm1[10],zero,ymm1[11],zero,ymm1[12],zero,ymm1[13],zero,ymm1[14],zero,ymm1[15],zero,ymm1[16],zero,ymm1[17],zero,ymm1[18],zero,ymm1[19],zero,ymm1[20],zero,ymm1[21],zero,ymm1[22],zero,ymm1[23],zero,ymm1[24],zero,ymm1[25],zero,ymm1[26],zero,ymm1[27],zero,ymm1[28],zero,ymm1[29],zero,ymm1[30],zero,ymm1[31],zero
+; AVX512BWVL-NEXT:    vpsravw %zmm1, %zmm0, %zmm0
+; AVX512BWVL-NEXT:    vpmovwb %zmm0, %ymm0
+; AVX512BWVL-NEXT:    retq
   %splat = shufflevector <32 x i8> %b, <32 x i8> undef, <32 x i32> zeroinitializer
   %shift = ashr <32 x i8> %a, %splat
   ret <32 x i8> %shift
@@ -724,6 +841,11 @@ define <4 x i64> @constant_shift_v4i64(<4 x i64> %a) nounwind {
 ; AVX512-NEXT:    vpxor %ymm1, %ymm0, %ymm0
 ; AVX512-NEXT:    vpsubq %ymm1, %ymm0, %ymm0
 ; AVX512-NEXT:    retq
+;
+; AVX512VL-LABEL: constant_shift_v4i64:
+; AVX512VL:       # BB#0:
+; AVX512VL-NEXT:    vpsravq {{.*}}(%rip), %ymm0, %ymm0
+; AVX512VL-NEXT:    retq
   %shift = ashr <4 x i64> %a, <i64 1, i64 7, i64 31, i64 62>
   ret <4 x i64> %shift
 }
@@ -769,6 +891,11 @@ define <8 x i32> @constant_shift_v8i32(<8 x i32> %a) nounwind {
 ; AVX512:       # BB#0:
 ; AVX512-NEXT:    vpsravd {{.*}}(%rip), %ymm0, %ymm0
 ; AVX512-NEXT:    retq
+;
+; AVX512VL-LABEL: constant_shift_v8i32:
+; AVX512VL:       # BB#0:
+; AVX512VL-NEXT:    vpsravd {{.*}}(%rip), %ymm0, %ymm0
+; AVX512VL-NEXT:    retq
   %shift = ashr <8 x i32> %a, <i32 4, i32 5, i32 6, i32 7, i32 8, i32 9, i32 8, i32 7>
   ret <8 x i32> %shift
 }
@@ -844,6 +971,18 @@ define <16 x i16> @constant_shift_v16i16(<16 x i16> %a) nounwind {
 ; AVX512BW-NEXT:    vpsravw %zmm1, %zmm0, %zmm0
 ; AVX512BW-NEXT:    # kill: %YMM0<def> %YMM0<kill> %ZMM0<kill>
 ; AVX512BW-NEXT:    retq
+;
+; AVX512DQVL-LABEL: constant_shift_v16i16:
+; AVX512DQVL:       # BB#0:
+; AVX512DQVL-NEXT:    vpmovsxwd %ymm0, %zmm0
+; AVX512DQVL-NEXT:    vpsravd {{.*}}(%rip), %zmm0, %zmm0
+; AVX512DQVL-NEXT:    vpmovdw %zmm0, %ymm0
+; AVX512DQVL-NEXT:    retq
+;
+; AVX512BWVL-LABEL: constant_shift_v16i16:
+; AVX512BWVL:       # BB#0:
+; AVX512BWVL-NEXT:    vpsravw {{.*}}(%rip), %ymm0, %ymm0
+; AVX512BWVL-NEXT:    retq
   %shift = ashr <16 x i16> %a, <i16 0, i16 1, i16 2, i16 3, i16 4, i16 5, i16 6, i16 7, i16 8, i16 9, i16 10, i16 11, i16 12, i16 13, i16 14, i16 15>
   ret <16 x i16> %shift
 }
@@ -981,6 +1120,42 @@ define <32 x i8> @constant_shift_v32i8(<32 x i8> %a) nounwind {
 ; AVX512BW-NEXT:    vpsravw {{.*}}(%rip), %zmm0, %zmm0
 ; AVX512BW-NEXT:    vpmovwb %zmm0, %ymm0
 ; AVX512BW-NEXT:    retq
+;
+; AVX512DQVL-LABEL: constant_shift_v32i8:
+; AVX512DQVL:       # BB#0:
+; AVX512DQVL-NEXT:    vmovdqa {{.*#+}} ymm1 = [0,1,2,3,4,5,6,7,7,6,5,4,3,2,1,0,0,1,2,3,4,5,6,7,7,6,5,4,3,2,1,0]
+; AVX512DQVL-NEXT:    vpsllw $5, %ymm1, %ymm1
+; AVX512DQVL-NEXT:    vpunpckhbw {{.*#+}} ymm2 = ymm0[8],ymm1[8],ymm0[9],ymm1[9],ymm0[10],ymm1[10],ymm0[11],ymm1[11],ymm0[12],ymm1[12],ymm0[13],ymm1[13],ymm0[14],ymm1[14],ymm0[15],ymm1[15],ymm0[24],ymm1[24],ymm0[25],ymm1[25],ymm0[26],ymm1[26],ymm0[27],ymm1[27],ymm0[28],ymm1[28],ymm0[29],ymm1[29],ymm0[30],ymm1[30],ymm0[31],ymm1[31]
+; AVX512DQVL-NEXT:    vpunpckhbw {{.*#+}} ymm3 = ymm0[8,8,9,9,10,10,11,11,12,12,13,13,14,14,15,15,24,24,25,25,26,26,27,27,28,28,29,29,30,30,31,31]
+; AVX512DQVL-NEXT:    vpsraw $4, %ymm3, %ymm4
+; AVX512DQVL-NEXT:    vpblendvb %ymm2, %ymm4, %ymm3, %ymm3
+; AVX512DQVL-NEXT:    vpsraw $2, %ymm3, %ymm4
+; AVX512DQVL-NEXT:    vpaddw %ymm2, %ymm2, %ymm2
+; AVX512DQVL-NEXT:    vpblendvb %ymm2, %ymm4, %ymm3, %ymm3
+; AVX512DQVL-NEXT:    vpsraw $1, %ymm3, %ymm4
+; AVX512DQVL-NEXT:    vpaddw %ymm2, %ymm2, %ymm2
+; AVX512DQVL-NEXT:    vpblendvb %ymm2, %ymm4, %ymm3, %ymm2
+; AVX512DQVL-NEXT:    vpsrlw $8, %ymm2, %ymm2
+; AVX512DQVL-NEXT:    vpunpcklbw {{.*#+}} ymm1 = ymm0[0],ymm1[0],ymm0[1],ymm1[1],ymm0[2],ymm1[2],ymm0[3],ymm1[3],ymm0[4],ymm1[4],ymm0[5],ymm1[5],ymm0[6],ymm1[6],ymm0[7],ymm1[7],ymm0[16],ymm1[16],ymm0[17],ymm1[17],ymm0[18],ymm1[18],ymm0[19],ymm1[19],ymm0[20],ymm1[20],ymm0[21],ymm1[21],ymm0[22],ymm1[22],ymm0[23],ymm1[23]
+; AVX512DQVL-NEXT:    vpunpcklbw {{.*#+}} ymm0 = ymm0[0,0,1,1,2,2,3,3,4,4,5,5,6,6,7,7,16,16,17,17,18,18,19,19,20,20,21,21,22,22,23,23]
+; AVX512DQVL-NEXT:    vpsraw $4, %ymm0, %ymm3
+; AVX512DQVL-NEXT:    vpblendvb %ymm1, %ymm3, %ymm0, %ymm0
+; AVX512DQVL-NEXT:    vpsraw $2, %ymm0, %ymm3
+; AVX512DQVL-NEXT:    vpaddw %ymm1, %ymm1, %ymm1
+; AVX512DQVL-NEXT:    vpblendvb %ymm1, %ymm3, %ymm0, %ymm0
+; AVX512DQVL-NEXT:    vpsraw $1, %ymm0, %ymm3
+; AVX512DQVL-NEXT:    vpaddw %ymm1, %ymm1, %ymm1
+; AVX512DQVL-NEXT:    vpblendvb %ymm1, %ymm3, %ymm0, %ymm0
+; AVX512DQVL-NEXT:    vpsrlw $8, %ymm0, %ymm0
+; AVX512DQVL-NEXT:    vpackuswb %ymm2, %ymm0, %ymm0
+; AVX512DQVL-NEXT:    retq
+;
+; AVX512BWVL-LABEL: constant_shift_v32i8:
+; AVX512BWVL:       # BB#0:
+; AVX512BWVL-NEXT:    vpmovsxbw %ymm0, %zmm0
+; AVX512BWVL-NEXT:    vpsravw {{.*}}(%rip), %zmm0, %zmm0
+; AVX512BWVL-NEXT:    vpmovwb %zmm0, %ymm0
+; AVX512BWVL-NEXT:    retq
   %shift = ashr <32 x i8> %a, <i8 0, i8 1, i8 2, i8 3, i8 4, i8 5, i8 6, i8 7, i8 7, i8 6, i8 5, i8 4, i8 3, i8 2, i8 1, i8 0, i8 0, i8 1, i8 2, i8 3, i8 4, i8 5, i8 6, i8 7, i8 7, i8 6, i8 5, i8 4, i8 3, i8 2, i8 1, i8 0>
   ret <32 x i8> %shift
 }
@@ -1033,6 +1208,11 @@ define <4 x i64> @splatconstant_shift_v4i64(<4 x i64> %a) nounwind {
 ; AVX512-NEXT:    vpsrlq $7, %ymm0, %ymm0
 ; AVX512-NEXT:    vpblendd {{.*#+}} ymm0 = ymm0[0],ymm1[1],ymm0[2],ymm1[3],ymm0[4],ymm1[5],ymm0[6],ymm1[7]
 ; AVX512-NEXT:    retq
+;
+; AVX512VL-LABEL: splatconstant_shift_v4i64:
+; AVX512VL:       # BB#0:
+; AVX512VL-NEXT:    vpsraq $7, %ymm0, %ymm0
+; AVX512VL-NEXT:    retq
   %shift = ashr <4 x i64> %a, <i64 7, i64 7, i64 7, i64 7>
   ret <4 x i64> %shift
 }
@@ -1068,6 +1248,11 @@ define <8 x i32> @splatconstant_shift_v8i32(<8 x i32> %a) nounwind {
 ; AVX512:       # BB#0:
 ; AVX512-NEXT:    vpsrad $5, %ymm0, %ymm0
 ; AVX512-NEXT:    retq
+;
+; AVX512VL-LABEL: splatconstant_shift_v8i32:
+; AVX512VL:       # BB#0:
+; AVX512VL-NEXT:    vpsrad $5, %ymm0, %ymm0
+; AVX512VL-NEXT:    retq
   %shift = ashr <8 x i32> %a, <i32 5, i32 5, i32 5, i32 5, i32 5, i32 5, i32 5, i32 5>
   ret <8 x i32> %shift
 }
@@ -1103,6 +1288,11 @@ define <16 x i16> @splatconstant_shift_v16i16(<16 x i16> %a) nounwind {
 ; AVX512:       # BB#0:
 ; AVX512-NEXT:    vpsraw $3, %ymm0, %ymm0
 ; AVX512-NEXT:    retq
+;
+; AVX512VL-LABEL: splatconstant_shift_v16i16:
+; AVX512VL:       # BB#0:
+; AVX512VL-NEXT:    vpsraw $3, %ymm0, %ymm0
+; AVX512VL-NEXT:    retq
   %shift = ashr <16 x i16> %a, <i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3>
   ret <16 x i16> %shift
 }
@@ -1160,6 +1350,15 @@ define <32 x i8> @splatconstant_shift_v32i8(<32 x i8> %a) nounwind {
 ; AVX512-NEXT:    vpxor %ymm1, %ymm0, %ymm0
 ; AVX512-NEXT:    vpsubb %ymm1, %ymm0, %ymm0
 ; AVX512-NEXT:    retq
+;
+; AVX512VL-LABEL: splatconstant_shift_v32i8:
+; AVX512VL:       # BB#0:
+; AVX512VL-NEXT:    vpsrlw $3, %ymm0, %ymm0
+; AVX512VL-NEXT:    vpand {{.*}}(%rip), %ymm0, %ymm0
+; AVX512VL-NEXT:    vmovdqa {{.*#+}} ymm1 = [16,16,16,16,16,16,16,16,16,16,16,16,16,16,16,16,16,16,16,16,16,16,16,16,16,16,16,16,16,16,16,16]
+; AVX512VL-NEXT:    vpxor %ymm1, %ymm0, %ymm0
+; AVX512VL-NEXT:    vpsubb %ymm1, %ymm0, %ymm0
+; AVX512VL-NEXT:    retq
   %shift = ashr <32 x i8> %a, <i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3>
   ret <32 x i8> %shift
 }

diff --git a/test/CodeGen/X86/vector-shift-ashr-512.ll b/test/CodeGen/X86/vector-shift-ashr-512.ll
index 2c9e433cfb2ca29b3d88b9a5a2520015b6cc84b8..1e9f465c3f3f23c78cedee7b7d2311acb92f2e2e 100644 (file)
--- a/test/CodeGen/X86/vector-shift-ashr-512.ll
+++ b/test/CodeGen/X86/vector-shift-ashr-512.ll
@@ -1,6 +1,7 @@
 ; NOTE: Assertions have been autogenerated by utils/update_llc_test_checks.py
 ; RUN: llc < %s -mtriple=x86_64-unknown-unknown -mcpu=knl -mattr=+avx512dq | FileCheck %s --check-prefix=ALL --check-prefix=AVX512 --check-prefix=AVX512DQ
 ; RUN: llc < %s -mtriple=x86_64-unknown-unknown -mcpu=knl -mattr=+avx512bw | FileCheck %s --check-prefix=ALL --check-prefix=AVX512 --check-prefix=AVX512BW
+
 ;
 ; Variable Shifts
 ;

diff --git a/test/CodeGen/X86/vector-shift-lshr-128.ll b/test/CodeGen/X86/vector-shift-lshr-128.ll
index a7e1a531b659894ea61861717dee68c27c435698..9b8c0def4558f29fde61c9de357f353ca67f9238 100644 (file)
--- a/test/CodeGen/X86/vector-shift-lshr-128.ll
+++ b/test/CodeGen/X86/vector-shift-lshr-128.ll
@@ -7,6 +7,8 @@
 ; RUN: llc < %s -mtriple=x86_64-unknown-unknown -mattr=+xop,+avx2 | FileCheck %s --check-prefix=ALL --check-prefix=XOP --check-prefix=XOPAVX2
 ; RUN: llc < %s -mtriple=x86_64-unknown-unknown -mcpu=knl -mattr=+avx512dq | FileCheck %s --check-prefix=ALL --check-prefix=AVX512 --check-prefix=AVX512DQ
 ; RUN: llc < %s -mtriple=x86_64-unknown-unknown -mcpu=knl -mattr=+avx512bw | FileCheck %s --check-prefix=ALL --check-prefix=AVX512 --check-prefix=AVX512BW
+; RUN: llc < %s -mtriple=x86_64-unknown-unknown -mcpu=knl -mattr=+avx512dq,+avx512vl | FileCheck %s --check-prefix=ALL --check-prefix=AVX512VL --check-prefix=AVX512DQVL
+; RUN: llc < %s -mtriple=x86_64-unknown-unknown -mcpu=knl -mattr=+avx512bw,+avx512vl | FileCheck %s --check-prefix=ALL --check-prefix=AVX512VL --check-prefix=AVX512BWVL
 ;
 ; Just one 32-bit run to make sure we do reasonable things for i64 shifts.
 ; RUN: llc < %s -mtriple=i686-unknown-unknown -mattr=+sse2 | FileCheck %s --check-prefix=ALL --check-prefix=X32-SSE --check-prefix=X32-SSE2
@@ -65,6 +67,11 @@ define <2 x i64> @var_shift_v2i64(<2 x i64> %a, <2 x i64> %b) nounwind {
 ; AVX512-NEXT:    vpsrlvq %xmm1, %xmm0, %xmm0
 ; AVX512-NEXT:    retq
 ;
+; AVX512VL-LABEL: var_shift_v2i64:
+; AVX512VL:       # BB#0:
+; AVX512VL-NEXT:    vpsrlvq %xmm1, %xmm0, %xmm0
+; AVX512VL-NEXT:    retq
+;
 ; X32-SSE-LABEL: var_shift_v2i64:
 ; X32-SSE:       # BB#0:
 ; X32-SSE-NEXT:    pshufd {{.*#+}} xmm3 = xmm1[2,3,0,1]
@@ -162,6 +169,11 @@ define <4 x i32> @var_shift_v4i32(<4 x i32> %a, <4 x i32> %b) nounwind {
 ; AVX512-NEXT:    vpsrlvd %xmm1, %xmm0, %xmm0
 ; AVX512-NEXT:    retq
 ;
+; AVX512VL-LABEL: var_shift_v4i32:
+; AVX512VL:       # BB#0:
+; AVX512VL-NEXT:    vpsrlvd %xmm1, %xmm0, %xmm0
+; AVX512VL-NEXT:    retq
+;
 ; X32-SSE-LABEL: var_shift_v4i32:
 ; X32-SSE:       # BB#0:
 ; X32-SSE-NEXT:    movdqa %xmm1, %xmm2
@@ -308,6 +320,19 @@ define <8 x i16> @var_shift_v8i16(<8 x i16> %a, <8 x i16> %b) nounwind {
 ; AVX512BW-NEXT:    # kill: %XMM0<def> %XMM0<kill> %ZMM0<kill>
 ; AVX512BW-NEXT:    retq
 ;
+; AVX512DQVL-LABEL: var_shift_v8i16:
+; AVX512DQVL:       # BB#0:
+; AVX512DQVL-NEXT:    vpmovzxwd {{.*#+}} ymm1 = xmm1[0],zero,xmm1[1],zero,xmm1[2],zero,xmm1[3],zero,xmm1[4],zero,xmm1[5],zero,xmm1[6],zero,xmm1[7],zero
+; AVX512DQVL-NEXT:    vpmovzxwd {{.*#+}} ymm0 = xmm0[0],zero,xmm0[1],zero,xmm0[2],zero,xmm0[3],zero,xmm0[4],zero,xmm0[5],zero,xmm0[6],zero,xmm0[7],zero
+; AVX512DQVL-NEXT:    vpsrlvd %ymm1, %ymm0, %ymm0
+; AVX512DQVL-NEXT:    vpmovdw %ymm0, %xmm0
+; AVX512DQVL-NEXT:    retq
+;
+; AVX512BWVL-LABEL: var_shift_v8i16:
+; AVX512BWVL:       # BB#0:
+; AVX512BWVL-NEXT:    vpsrlvw %xmm1, %xmm0, %xmm0
+; AVX512BWVL-NEXT:    retq
+;
 ; X32-SSE-LABEL: var_shift_v8i16:
 ; X32-SSE:       # BB#0:
 ; X32-SSE-NEXT:    psllw $12, %xmm1
@@ -433,6 +458,14 @@ define <16 x i8> @var_shift_v16i8(<16 x i8> %a, <16 x i8> %b) nounwind {
 ; AVX512-NEXT:    vpmovdb %zmm0, %xmm0
 ; AVX512-NEXT:    retq
 ;
+; AVX512VL-LABEL: var_shift_v16i8:
+; AVX512VL:       # BB#0:
+; AVX512VL-NEXT:    vpmovzxbd {{.*#+}} zmm1 = xmm1[0],zero,zero,zero,xmm1[1],zero,zero,zero,xmm1[2],zero,zero,zero,xmm1[3],zero,zero,zero,xmm1[4],zero,zero,zero,xmm1[5],zero,zero,zero,xmm1[6],zero,zero,zero,xmm1[7],zero,zero,zero,xmm1[8],zero,zero,zero,xmm1[9],zero,zero,zero,xmm1[10],zero,zero,zero,xmm1[11],zero,zero,zero,xmm1[12],zero,zero,zero,xmm1[13],zero,zero,zero,xmm1[14],zero,zero,zero,xmm1[15],zero,zero,zero
+; AVX512VL-NEXT:    vpmovzxbd {{.*#+}} zmm0 = xmm0[0],zero,zero,zero,xmm0[1],zero,zero,zero,xmm0[2],zero,zero,zero,xmm0[3],zero,zero,zero,xmm0[4],zero,zero,zero,xmm0[5],zero,zero,zero,xmm0[6],zero,zero,zero,xmm0[7],zero,zero,zero,xmm0[8],zero,zero,zero,xmm0[9],zero,zero,zero,xmm0[10],zero,zero,zero,xmm0[11],zero,zero,zero,xmm0[12],zero,zero,zero,xmm0[13],zero,zero,zero,xmm0[14],zero,zero,zero,xmm0[15],zero,zero,zero
+; AVX512VL-NEXT:    vpsrlvd %zmm1, %zmm0, %zmm0
+; AVX512VL-NEXT:    vpmovdb %zmm0, %xmm0
+; AVX512VL-NEXT:    retq
+;
 ; X32-SSE-LABEL: var_shift_v16i8:
 ; X32-SSE:       # BB#0:
 ; X32-SSE-NEXT:    psllw $5, %xmm1
@@ -492,6 +525,11 @@ define <2 x i64> @splatvar_shift_v2i64(<2 x i64> %a, <2 x i64> %b) nounwind {
 ; AVX512-NEXT:    vpsrlq %xmm1, %xmm0, %xmm0
 ; AVX512-NEXT:    retq
 ;
+; AVX512VL-LABEL: splatvar_shift_v2i64:
+; AVX512VL:       # BB#0:
+; AVX512VL-NEXT:    vpsrlq %xmm1, %xmm0, %xmm0
+; AVX512VL-NEXT:    retq
+;
 ; X32-SSE-LABEL: splatvar_shift_v2i64:
 ; X32-SSE:       # BB#0:
 ; X32-SSE-NEXT:    psrlq %xmm1, %xmm0
@@ -533,6 +571,12 @@ define <4 x i32> @splatvar_shift_v4i32(<4 x i32> %a, <4 x i32> %b) nounwind {
 ; AVX512-NEXT:    vpsrld %xmm1, %xmm0, %xmm0
 ; AVX512-NEXT:    retq
 ;
+; AVX512VL-LABEL: splatvar_shift_v4i32:
+; AVX512VL:       # BB#0:
+; AVX512VL-NEXT:    vpmovzxdq {{.*#+}} xmm1 = xmm1[0],zero,xmm1[1],zero
+; AVX512VL-NEXT:    vpsrld %xmm1, %xmm0, %xmm0
+; AVX512VL-NEXT:    retq
+;
 ; X32-SSE-LABEL: splatvar_shift_v4i32:
 ; X32-SSE:       # BB#0:
 ; X32-SSE-NEXT:    xorps %xmm2, %xmm2
@@ -576,6 +620,12 @@ define <8 x i16> @splatvar_shift_v8i16(<8 x i16> %a, <8 x i16> %b) nounwind {
 ; AVX512-NEXT:    vpsrlw %xmm1, %xmm0, %xmm0
 ; AVX512-NEXT:    retq
 ;
+; AVX512VL-LABEL: splatvar_shift_v8i16:
+; AVX512VL:       # BB#0:
+; AVX512VL-NEXT:    vpmovzxwq {{.*#+}} xmm1 = xmm1[0],zero,zero,zero,xmm1[1],zero,zero,zero
+; AVX512VL-NEXT:    vpsrlw %xmm1, %xmm0, %xmm0
+; AVX512VL-NEXT:    retq
+;
 ; X32-SSE-LABEL: splatvar_shift_v8i16:
 ; X32-SSE:       # BB#0:
 ; X32-SSE-NEXT:    pextrw $0, %xmm1, %eax
@@ -709,6 +759,15 @@ define <16 x i8> @splatvar_shift_v16i8(<16 x i8> %a, <16 x i8> %b) nounwind {
 ; AVX512-NEXT:    vpmovdb %zmm0, %xmm0
 ; AVX512-NEXT:    retq
 ;
+; AVX512VL-LABEL: splatvar_shift_v16i8:
+; AVX512VL:       # BB#0:
+; AVX512VL-NEXT:    vpbroadcastb %xmm1, %xmm1
+; AVX512VL-NEXT:    vpmovzxbd {{.*#+}} zmm0 = xmm0[0],zero,zero,zero,xmm0[1],zero,zero,zero,xmm0[2],zero,zero,zero,xmm0[3],zero,zero,zero,xmm0[4],zero,zero,zero,xmm0[5],zero,zero,zero,xmm0[6],zero,zero,zero,xmm0[7],zero,zero,zero,xmm0[8],zero,zero,zero,xmm0[9],zero,zero,zero,xmm0[10],zero,zero,zero,xmm0[11],zero,zero,zero,xmm0[12],zero,zero,zero,xmm0[13],zero,zero,zero,xmm0[14],zero,zero,zero,xmm0[15],zero,zero,zero
+; AVX512VL-NEXT:    vpmovzxbd {{.*#+}} zmm1 = xmm1[0],zero,zero,zero,xmm1[1],zero,zero,zero,xmm1[2],zero,zero,zero,xmm1[3],zero,zero,zero,xmm1[4],zero,zero,zero,xmm1[5],zero,zero,zero,xmm1[6],zero,zero,zero,xmm1[7],zero,zero,zero,xmm1[8],zero,zero,zero,xmm1[9],zero,zero,zero,xmm1[10],zero,zero,zero,xmm1[11],zero,zero,zero,xmm1[12],zero,zero,zero,xmm1[13],zero,zero,zero,xmm1[14],zero,zero,zero,xmm1[15],zero,zero,zero
+; AVX512VL-NEXT:    vpsrlvd %zmm1, %zmm0, %zmm0
+; AVX512VL-NEXT:    vpmovdb %zmm0, %xmm0
+; AVX512VL-NEXT:    retq
+;
 ; X32-SSE-LABEL: splatvar_shift_v16i8:
 ; X32-SSE:       # BB#0:
 ; X32-SSE-NEXT:    punpcklbw {{.*#+}} xmm1 = xmm1[0,0,1,1,2,2,3,3,4,4,5,5,6,6,7,7]
@@ -798,6 +857,11 @@ define <2 x i64> @constant_shift_v2i64(<2 x i64> %a) nounwind {
 ; AVX512-NEXT:    vpsrlvq {{.*}}(%rip), %xmm0, %xmm0
 ; AVX512-NEXT:    retq
 ;
+; AVX512VL-LABEL: constant_shift_v2i64:
+; AVX512VL:       # BB#0:
+; AVX512VL-NEXT:    vpsrlvq {{.*}}(%rip), %xmm0, %xmm0
+; AVX512VL-NEXT:    retq
+;
 ; X32-SSE-LABEL: constant_shift_v2i64:
 ; X32-SSE:       # BB#0:
 ; X32-SSE-NEXT:    movdqa %xmm0, %xmm1
@@ -872,6 +936,11 @@ define <4 x i32> @constant_shift_v4i32(<4 x i32> %a) nounwind {
 ; AVX512-NEXT:    vpsrlvd {{.*}}(%rip), %xmm0, %xmm0
 ; AVX512-NEXT:    retq
 ;
+; AVX512VL-LABEL: constant_shift_v4i32:
+; AVX512VL:       # BB#0:
+; AVX512VL-NEXT:    vpsrlvd {{.*}}(%rip), %xmm0, %xmm0
+; AVX512VL-NEXT:    retq
+;
 ; X32-SSE-LABEL: constant_shift_v4i32:
 ; X32-SSE:       # BB#0:
 ; X32-SSE-NEXT:    movdqa %xmm0, %xmm1
@@ -965,6 +1034,18 @@ define <8 x i16> @constant_shift_v8i16(<8 x i16> %a) nounwind {
 ; AVX512BW-NEXT:    # kill: %XMM0<def> %XMM0<kill> %ZMM0<kill>
 ; AVX512BW-NEXT:    retq
 ;
+; AVX512DQVL-LABEL: constant_shift_v8i16:
+; AVX512DQVL:       # BB#0:
+; AVX512DQVL-NEXT:    vpmovzxwd {{.*#+}} ymm0 = xmm0[0],zero,xmm0[1],zero,xmm0[2],zero,xmm0[3],zero,xmm0[4],zero,xmm0[5],zero,xmm0[6],zero,xmm0[7],zero
+; AVX512DQVL-NEXT:    vpsrlvd {{.*}}(%rip), %ymm0, %ymm0
+; AVX512DQVL-NEXT:    vpmovdw %ymm0, %xmm0
+; AVX512DQVL-NEXT:    retq
+;
+; AVX512BWVL-LABEL: constant_shift_v8i16:
+; AVX512BWVL:       # BB#0:
+; AVX512BWVL-NEXT:    vpsrlvw {{.*}}(%rip), %xmm0, %xmm0
+; AVX512BWVL-NEXT:    retq
+;
 ; X32-SSE-LABEL: constant_shift_v8i16:
 ; X32-SSE:       # BB#0:
 ; X32-SSE-NEXT:    movdqa %xmm0, %xmm1
@@ -1071,6 +1152,13 @@ define <16 x i8> @constant_shift_v16i8(<16 x i8> %a) nounwind {
 ; AVX512-NEXT:    vpmovdb %zmm0, %xmm0
 ; AVX512-NEXT:    retq
 ;
+; AVX512VL-LABEL: constant_shift_v16i8:
+; AVX512VL:       # BB#0:
+; AVX512VL-NEXT:    vpmovzxbd {{.*#+}} zmm0 = xmm0[0],zero,zero,zero,xmm0[1],zero,zero,zero,xmm0[2],zero,zero,zero,xmm0[3],zero,zero,zero,xmm0[4],zero,zero,zero,xmm0[5],zero,zero,zero,xmm0[6],zero,zero,zero,xmm0[7],zero,zero,zero,xmm0[8],zero,zero,zero,xmm0[9],zero,zero,zero,xmm0[10],zero,zero,zero,xmm0[11],zero,zero,zero,xmm0[12],zero,zero,zero,xmm0[13],zero,zero,zero,xmm0[14],zero,zero,zero,xmm0[15],zero,zero,zero
+; AVX512VL-NEXT:    vpsrlvd {{.*}}(%rip), %zmm0, %zmm0
+; AVX512VL-NEXT:    vpmovdb %zmm0, %xmm0
+; AVX512VL-NEXT:    retq
+;
 ; X32-SSE-LABEL: constant_shift_v16i8:
 ; X32-SSE:       # BB#0:
 ; X32-SSE-NEXT:    movdqa {{.*#+}} xmm2 = [0,1,2,3,4,5,6,7,7,6,5,4,3,2,1,0]
@@ -1131,6 +1219,11 @@ define <2 x i64> @splatconstant_shift_v2i64(<2 x i64> %a) nounwind {
 ; AVX512-NEXT:    vpsrlq $7, %xmm0, %xmm0
 ; AVX512-NEXT:    retq
 ;
+; AVX512VL-LABEL: splatconstant_shift_v2i64:
+; AVX512VL:       # BB#0:
+; AVX512VL-NEXT:    vpsrlq $7, %xmm0, %xmm0
+; AVX512VL-NEXT:    retq
+;
 ; X32-SSE-LABEL: splatconstant_shift_v2i64:
 ; X32-SSE:       # BB#0:
 ; X32-SSE-NEXT:    psrlq $7, %xmm0
@@ -1160,6 +1253,11 @@ define <4 x i32> @splatconstant_shift_v4i32(<4 x i32> %a) nounwind {
 ; AVX512-NEXT:    vpsrld $5, %xmm0, %xmm0
 ; AVX512-NEXT:    retq
 ;
+; AVX512VL-LABEL: splatconstant_shift_v4i32:
+; AVX512VL:       # BB#0:
+; AVX512VL-NEXT:    vpsrld $5, %xmm0, %xmm0
+; AVX512VL-NEXT:    retq
+;
 ; X32-SSE-LABEL: splatconstant_shift_v4i32:
 ; X32-SSE:       # BB#0:
 ; X32-SSE-NEXT:    psrld $5, %xmm0
@@ -1189,6 +1287,11 @@ define <8 x i16> @splatconstant_shift_v8i16(<8 x i16> %a) nounwind {
 ; AVX512-NEXT:    vpsrlw $3, %xmm0, %xmm0
 ; AVX512-NEXT:    retq
 ;
+; AVX512VL-LABEL: splatconstant_shift_v8i16:
+; AVX512VL:       # BB#0:
+; AVX512VL-NEXT:    vpsrlw $3, %xmm0, %xmm0
+; AVX512VL-NEXT:    retq
+;
 ; X32-SSE-LABEL: splatconstant_shift_v8i16:
 ; X32-SSE:       # BB#0:
 ; X32-SSE-NEXT:    psrlw $3, %xmm0
@@ -1223,6 +1326,12 @@ define <16 x i8> @splatconstant_shift_v16i8(<16 x i8> %a) nounwind {
 ; AVX512-NEXT:    vpand {{.*}}(%rip), %xmm0, %xmm0
 ; AVX512-NEXT:    retq
 ;
+; AVX512VL-LABEL: splatconstant_shift_v16i8:
+; AVX512VL:       # BB#0:
+; AVX512VL-NEXT:    vpsrlw $3, %xmm0, %xmm0
+; AVX512VL-NEXT:    vpand {{.*}}(%rip), %xmm0, %xmm0
+; AVX512VL-NEXT:    retq
+;
 ; X32-SSE-LABEL: splatconstant_shift_v16i8:
 ; X32-SSE:       # BB#0:
 ; X32-SSE-NEXT:    psrlw $3, %xmm0

diff --git a/test/CodeGen/X86/vector-shift-lshr-256.ll b/test/CodeGen/X86/vector-shift-lshr-256.ll
index 25667e7d1661a704827073a24a4042af2bf2c693..58bb8f3e6ec012cb23bcd5c2de5f70bd5e2e4187 100644 (file)
--- a/test/CodeGen/X86/vector-shift-lshr-256.ll
+++ b/test/CodeGen/X86/vector-shift-lshr-256.ll
@@ -5,6 +5,9 @@
 ; RUN: llc < %s -mtriple=x86_64-unknown-unknown -mattr=+xop,+avx2 | FileCheck %s --check-prefix=ALL --check-prefix=XOP --check-prefix=XOPAVX2
 ; RUN: llc < %s -mtriple=x86_64-unknown-unknown -mcpu=knl -mattr=+avx512dq | FileCheck %s --check-prefix=ALL --check-prefix=AVX512 --check-prefix=AVX512DQ
 ; RUN: llc < %s -mtriple=x86_64-unknown-unknown -mcpu=knl -mattr=+avx512bw | FileCheck %s --check-prefix=ALL --check-prefix=AVX512 --check-prefix=AVX512BW
+; RUN: llc < %s -mtriple=x86_64-unknown-unknown -mcpu=knl -mattr=+avx512dq,+avx512vl | FileCheck %s --check-prefix=ALL --check-prefix=AVX512VL --check-prefix=AVX512DQVL
+; RUN: llc < %s -mtriple=x86_64-unknown-unknown -mcpu=knl -mattr=+avx512bw,+avx512vl | FileCheck %s --check-prefix=ALL --check-prefix=AVX512VL --check-prefix=AVX512BWVL
+
 ;
 ; Variable Shifts
 ;
@@ -51,6 +54,11 @@ define <4 x i64> @var_shift_v4i64(<4 x i64> %a, <4 x i64> %b) nounwind {
 ; AVX512:       # BB#0:
 ; AVX512-NEXT:    vpsrlvq %ymm1, %ymm0, %ymm0
 ; AVX512-NEXT:    retq
+;
+; AVX512VL-LABEL: var_shift_v4i64:
+; AVX512VL:       # BB#0:
+; AVX512VL-NEXT:    vpsrlvq %ymm1, %ymm0, %ymm0
+; AVX512VL-NEXT:    retq
   %shift = lshr <4 x i64> %a, %b
   ret <4 x i64> %shift
 }
@@ -112,6 +120,11 @@ define <8 x i32> @var_shift_v8i32(<8 x i32> %a, <8 x i32> %b) nounwind {
 ; AVX512:       # BB#0:
 ; AVX512-NEXT:    vpsrlvd %ymm1, %ymm0, %ymm0
 ; AVX512-NEXT:    retq
+;
+; AVX512VL-LABEL: var_shift_v8i32:
+; AVX512VL:       # BB#0:
+; AVX512VL-NEXT:    vpsrlvd %ymm1, %ymm0, %ymm0
+; AVX512VL-NEXT:    retq
   %shift = lshr <8 x i32> %a, %b
   ret <8 x i32> %shift
 }
@@ -205,6 +218,19 @@ define <16 x i16> @var_shift_v16i16(<16 x i16> %a, <16 x i16> %b) nounwind {
 ; AVX512BW-NEXT:    vpsrlvw %zmm1, %zmm0, %zmm0
 ; AVX512BW-NEXT:    # kill: %YMM0<def> %YMM0<kill> %ZMM0<kill>
 ; AVX512BW-NEXT:    retq
+;
+; AVX512DQVL-LABEL: var_shift_v16i16:
+; AVX512DQVL:       # BB#0:
+; AVX512DQVL-NEXT:    vpmovzxwd {{.*#+}} zmm1 = ymm1[0],zero,ymm1[1],zero,ymm1[2],zero,ymm1[3],zero,ymm1[4],zero,ymm1[5],zero,ymm1[6],zero,ymm1[7],zero,ymm1[8],zero,ymm1[9],zero,ymm1[10],zero,ymm1[11],zero,ymm1[12],zero,ymm1[13],zero,ymm1[14],zero,ymm1[15],zero
+; AVX512DQVL-NEXT:    vpmovzxwd {{.*#+}} zmm0 = ymm0[0],zero,ymm0[1],zero,ymm0[2],zero,ymm0[3],zero,ymm0[4],zero,ymm0[5],zero,ymm0[6],zero,ymm0[7],zero,ymm0[8],zero,ymm0[9],zero,ymm0[10],zero,ymm0[11],zero,ymm0[12],zero,ymm0[13],zero,ymm0[14],zero,ymm0[15],zero
+; AVX512DQVL-NEXT:    vpsrlvd %zmm1, %zmm0, %zmm0
+; AVX512DQVL-NEXT:    vpmovdw %zmm0, %ymm0
+; AVX512DQVL-NEXT:    retq
+;
+; AVX512BWVL-LABEL: var_shift_v16i16:
+; AVX512BWVL:       # BB#0:
+; AVX512BWVL-NEXT:    vpsrlvw %ymm1, %ymm0, %ymm0
+; AVX512BWVL-NEXT:    retq
   %shift = lshr <16 x i16> %a, %b
   ret <16 x i16> %shift
 }
@@ -307,6 +333,30 @@ define <32 x i8> @var_shift_v32i8(<32 x i8> %a, <32 x i8> %b) nounwind {
 ; AVX512BW-NEXT:    vpsrlvw %zmm1, %zmm0, %zmm0
 ; AVX512BW-NEXT:    vpmovwb %zmm0, %ymm0
 ; AVX512BW-NEXT:    retq
+;
+; AVX512DQVL-LABEL: var_shift_v32i8:
+; AVX512DQVL:       # BB#0:
+; AVX512DQVL-NEXT:    vpsllw $5, %ymm1, %ymm1
+; AVX512DQVL-NEXT:    vpsrlw $4, %ymm0, %ymm2
+; AVX512DQVL-NEXT:    vpand {{.*}}(%rip), %ymm2, %ymm2
+; AVX512DQVL-NEXT:    vpblendvb %ymm1, %ymm2, %ymm0, %ymm0
+; AVX512DQVL-NEXT:    vpsrlw $2, %ymm0, %ymm2
+; AVX512DQVL-NEXT:    vpand {{.*}}(%rip), %ymm2, %ymm2
+; AVX512DQVL-NEXT:    vpaddb %ymm1, %ymm1, %ymm1
+; AVX512DQVL-NEXT:    vpblendvb %ymm1, %ymm2, %ymm0, %ymm0
+; AVX512DQVL-NEXT:    vpsrlw $1, %ymm0, %ymm2
+; AVX512DQVL-NEXT:    vpand {{.*}}(%rip), %ymm2, %ymm2
+; AVX512DQVL-NEXT:    vpaddb %ymm1, %ymm1, %ymm1
+; AVX512DQVL-NEXT:    vpblendvb %ymm1, %ymm2, %ymm0, %ymm0
+; AVX512DQVL-NEXT:    retq
+;
+; AVX512BWVL-LABEL: var_shift_v32i8:
+; AVX512BWVL:       # BB#0:
+; AVX512BWVL-NEXT:    vpmovzxbw {{.*#+}} zmm1 = ymm1[0],zero,ymm1[1],zero,ymm1[2],zero,ymm1[3],zero,ymm1[4],zero,ymm1[5],zero,ymm1[6],zero,ymm1[7],zero,ymm1[8],zero,ymm1[9],zero,ymm1[10],zero,ymm1[11],zero,ymm1[12],zero,ymm1[13],zero,ymm1[14],zero,ymm1[15],zero,ymm1[16],zero,ymm1[17],zero,ymm1[18],zero,ymm1[19],zero,ymm1[20],zero,ymm1[21],zero,ymm1[22],zero,ymm1[23],zero,ymm1[24],zero,ymm1[25],zero,ymm1[26],zero,ymm1[27],zero,ymm1[28],zero,ymm1[29],zero,ymm1[30],zero,ymm1[31],zero
+; AVX512BWVL-NEXT:    vpmovzxbw {{.*#+}} zmm0 = ymm0[0],zero,ymm0[1],zero,ymm0[2],zero,ymm0[3],zero,ymm0[4],zero,ymm0[5],zero,ymm0[6],zero,ymm0[7],zero,ymm0[8],zero,ymm0[9],zero,ymm0[10],zero,ymm0[11],zero,ymm0[12],zero,ymm0[13],zero,ymm0[14],zero,ymm0[15],zero,ymm0[16],zero,ymm0[17],zero,ymm0[18],zero,ymm0[19],zero,ymm0[20],zero,ymm0[21],zero,ymm0[22],zero,ymm0[23],zero,ymm0[24],zero,ymm0[25],zero,ymm0[26],zero,ymm0[27],zero,ymm0[28],zero,ymm0[29],zero,ymm0[30],zero,ymm0[31],zero
+; AVX512BWVL-NEXT:    vpsrlvw %zmm1, %zmm0, %zmm0
+; AVX512BWVL-NEXT:    vpmovwb %zmm0, %ymm0
+; AVX512BWVL-NEXT:    retq
   %shift = lshr <32 x i8> %a, %b
   ret <32 x i8> %shift
 }
@@ -346,6 +396,11 @@ define <4 x i64> @splatvar_shift_v4i64(<4 x i64> %a, <4 x i64> %b) nounwind {
 ; AVX512:       # BB#0:
 ; AVX512-NEXT:    vpsrlq %xmm1, %ymm0, %ymm0
 ; AVX512-NEXT:    retq
+;
+; AVX512VL-LABEL: splatvar_shift_v4i64:
+; AVX512VL:       # BB#0:
+; AVX512VL-NEXT:    vpsrlq %xmm1, %ymm0, %ymm0
+; AVX512VL-NEXT:    retq
   %splat = shufflevector <4 x i64> %b, <4 x i64> undef, <4 x i32> zeroinitializer
   %shift = lshr <4 x i64> %a, %splat
   ret <4 x i64> %shift
@@ -387,6 +442,12 @@ define <8 x i32> @splatvar_shift_v8i32(<8 x i32> %a, <8 x i32> %b) nounwind {
 ; AVX512-NEXT:    vpmovzxdq {{.*#+}} xmm1 = xmm1[0],zero,xmm1[1],zero
 ; AVX512-NEXT:    vpsrld %xmm1, %ymm0, %ymm0
 ; AVX512-NEXT:    retq
+;
+; AVX512VL-LABEL: splatvar_shift_v8i32:
+; AVX512VL:       # BB#0:
+; AVX512VL-NEXT:    vpmovzxdq {{.*#+}} xmm1 = xmm1[0],zero,xmm1[1],zero
+; AVX512VL-NEXT:    vpsrld %xmm1, %ymm0, %ymm0
+; AVX512VL-NEXT:    retq
   %splat = shufflevector <8 x i32> %b, <8 x i32> undef, <8 x i32> zeroinitializer
   %shift = lshr <8 x i32> %a, %splat
   ret <8 x i32> %shift
@@ -428,6 +489,12 @@ define <16 x i16> @splatvar_shift_v16i16(<16 x i16> %a, <16 x i16> %b) nounwind
 ; AVX512-NEXT:    vpmovzxwq {{.*#+}} xmm1 = xmm1[0],zero,zero,zero,xmm1[1],zero,zero,zero
 ; AVX512-NEXT:    vpsrlw %xmm1, %ymm0, %ymm0
 ; AVX512-NEXT:    retq
+;
+; AVX512VL-LABEL: splatvar_shift_v16i16:
+; AVX512VL:       # BB#0:
+; AVX512VL-NEXT:    vpmovzxwq {{.*#+}} xmm1 = xmm1[0],zero,zero,zero,xmm1[1],zero,zero,zero
+; AVX512VL-NEXT:    vpsrlw %xmm1, %ymm0, %ymm0
+; AVX512VL-NEXT:    retq
   %splat = shufflevector <16 x i16> %b, <16 x i16> undef, <16 x i32> zeroinitializer
   %shift = lshr <16 x i16> %a, %splat
   ret <16 x i16> %shift
@@ -532,6 +599,32 @@ define <32 x i8> @splatvar_shift_v32i8(<32 x i8> %a, <32 x i8> %b) nounwind {
 ; AVX512BW-NEXT:    vpsrlvw %zmm1, %zmm0, %zmm0
 ; AVX512BW-NEXT:    vpmovwb %zmm0, %ymm0
 ; AVX512BW-NEXT:    retq
+;
+; AVX512DQVL-LABEL: splatvar_shift_v32i8:
+; AVX512DQVL:       # BB#0:
+; AVX512DQVL-NEXT:    vpbroadcastb %xmm1, %ymm1
+; AVX512DQVL-NEXT:    vpsrlw $4, %ymm0, %ymm2
+; AVX512DQVL-NEXT:    vpand {{.*}}(%rip), %ymm2, %ymm2
+; AVX512DQVL-NEXT:    vpsllw $5, %ymm1, %ymm1
+; AVX512DQVL-NEXT:    vpblendvb %ymm1, %ymm2, %ymm0, %ymm0
+; AVX512DQVL-NEXT:    vpsrlw $2, %ymm0, %ymm2
+; AVX512DQVL-NEXT:    vpand {{.*}}(%rip), %ymm2, %ymm2
+; AVX512DQVL-NEXT:    vpaddb %ymm1, %ymm1, %ymm1
+; AVX512DQVL-NEXT:    vpblendvb %ymm1, %ymm2, %ymm0, %ymm0
+; AVX512DQVL-NEXT:    vpsrlw $1, %ymm0, %ymm2
+; AVX512DQVL-NEXT:    vpand {{.*}}(%rip), %ymm2, %ymm2
+; AVX512DQVL-NEXT:    vpaddb %ymm1, %ymm1, %ymm1
+; AVX512DQVL-NEXT:    vpblendvb %ymm1, %ymm2, %ymm0, %ymm0
+; AVX512DQVL-NEXT:    retq
+;
+; AVX512BWVL-LABEL: splatvar_shift_v32i8:
+; AVX512BWVL:       # BB#0:
+; AVX512BWVL-NEXT:    vpbroadcastb %xmm1, %ymm1
+; AVX512BWVL-NEXT:    vpmovzxbw {{.*#+}} zmm0 = ymm0[0],zero,ymm0[1],zero,ymm0[2],zero,ymm0[3],zero,ymm0[4],zero,ymm0[5],zero,ymm0[6],zero,ymm0[7],zero,ymm0[8],zero,ymm0[9],zero,ymm0[10],zero,ymm0[11],zero,ymm0[12],zero,ymm0[13],zero,ymm0[14],zero,ymm0[15],zero,ymm0[16],zero,ymm0[17],zero,ymm0[18],zero,ymm0[19],zero,ymm0[20],zero,ymm0[21],zero,ymm0[22],zero,ymm0[23],zero,ymm0[24],zero,ymm0[25],zero,ymm0[26],zero,ymm0[27],zero,ymm0[28],zero,ymm0[29],zero,ymm0[30],zero,ymm0[31],zero
+; AVX512BWVL-NEXT:    vpmovzxbw {{.*#+}} zmm1 = ymm1[0],zero,ymm1[1],zero,ymm1[2],zero,ymm1[3],zero,ymm1[4],zero,ymm1[5],zero,ymm1[6],zero,ymm1[7],zero,ymm1[8],zero,ymm1[9],zero,ymm1[10],zero,ymm1[11],zero,ymm1[12],zero,ymm1[13],zero,ymm1[14],zero,ymm1[15],zero,ymm1[16],zero,ymm1[17],zero,ymm1[18],zero,ymm1[19],zero,ymm1[20],zero,ymm1[21],zero,ymm1[22],zero,ymm1[23],zero,ymm1[24],zero,ymm1[25],zero,ymm1[26],zero,ymm1[27],zero,ymm1[28],zero,ymm1[29],zero,ymm1[30],zero,ymm1[31],zero
+; AVX512BWVL-NEXT:    vpsrlvw %zmm1, %zmm0, %zmm0
+; AVX512BWVL-NEXT:    vpmovwb %zmm0, %ymm0
+; AVX512BWVL-NEXT:    retq
   %splat = shufflevector <32 x i8> %b, <32 x i8> undef, <32 x i32> zeroinitializer
   %shift = lshr <32 x i8> %a, %splat
   ret <32 x i8> %shift
@@ -579,6 +672,11 @@ define <4 x i64> @constant_shift_v4i64(<4 x i64> %a) nounwind {
 ; AVX512:       # BB#0:
 ; AVX512-NEXT:    vpsrlvq {{.*}}(%rip), %ymm0, %ymm0
 ; AVX512-NEXT:    retq
+;
+; AVX512VL-LABEL: constant_shift_v4i64:
+; AVX512VL:       # BB#0:
+; AVX512VL-NEXT:    vpsrlvq {{.*}}(%rip), %ymm0, %ymm0
+; AVX512VL-NEXT:    retq
   %shift = lshr <4 x i64> %a, <i64 1, i64 7, i64 31, i64 62>
   ret <4 x i64> %shift
 }
@@ -624,6 +722,11 @@ define <8 x i32> @constant_shift_v8i32(<8 x i32> %a) nounwind {
 ; AVX512:       # BB#0:
 ; AVX512-NEXT:    vpsrlvd {{.*}}(%rip), %ymm0, %ymm0
 ; AVX512-NEXT:    retq
+;
+; AVX512VL-LABEL: constant_shift_v8i32:
+; AVX512VL:       # BB#0:
+; AVX512VL-NEXT:    vpsrlvd {{.*}}(%rip), %ymm0, %ymm0
+; AVX512VL-NEXT:    retq
   %shift = lshr <8 x i32> %a, <i32 4, i32 5, i32 6, i32 7, i32 8, i32 9, i32 8, i32 7>
   ret <8 x i32> %shift
 }
@@ -699,6 +802,18 @@ define <16 x i16> @constant_shift_v16i16(<16 x i16> %a) nounwind {
 ; AVX512BW-NEXT:    vpsrlvw %zmm1, %zmm0, %zmm0
 ; AVX512BW-NEXT:    # kill: %YMM0<def> %YMM0<kill> %ZMM0<kill>
 ; AVX512BW-NEXT:    retq
+;
+; AVX512DQVL-LABEL: constant_shift_v16i16:
+; AVX512DQVL:       # BB#0:
+; AVX512DQVL-NEXT:    vpmovzxwd {{.*#+}} zmm0 = ymm0[0],zero,ymm0[1],zero,ymm0[2],zero,ymm0[3],zero,ymm0[4],zero,ymm0[5],zero,ymm0[6],zero,ymm0[7],zero,ymm0[8],zero,ymm0[9],zero,ymm0[10],zero,ymm0[11],zero,ymm0[12],zero,ymm0[13],zero,ymm0[14],zero,ymm0[15],zero
+; AVX512DQVL-NEXT:    vpsrlvd {{.*}}(%rip), %zmm0, %zmm0
+; AVX512DQVL-NEXT:    vpmovdw %zmm0, %ymm0
+; AVX512DQVL-NEXT:    retq
+;
+; AVX512BWVL-LABEL: constant_shift_v16i16:
+; AVX512BWVL:       # BB#0:
+; AVX512BWVL-NEXT:    vpsrlvw {{.*}}(%rip), %ymm0, %ymm0
+; AVX512BWVL-NEXT:    retq
   %shift = lshr <16 x i16> %a, <i16 0, i16 1, i16 2, i16 3, i16 4, i16 5, i16 6, i16 7, i16 8, i16 9, i16 10, i16 11, i16 12, i16 13, i16 14, i16 15>
   ret <16 x i16> %shift
 }
@@ -795,6 +910,30 @@ define <32 x i8> @constant_shift_v32i8(<32 x i8> %a) nounwind {
 ; AVX512BW-NEXT:    vpsrlvw {{.*}}(%rip), %zmm0, %zmm0
 ; AVX512BW-NEXT:    vpmovwb %zmm0, %ymm0
 ; AVX512BW-NEXT:    retq
+;
+; AVX512DQVL-LABEL: constant_shift_v32i8:
+; AVX512DQVL:       # BB#0:
+; AVX512DQVL-NEXT:    vmovdqa {{.*#+}} ymm1 = [0,1,2,3,4,5,6,7,7,6,5,4,3,2,1,0,0,1,2,3,4,5,6,7,7,6,5,4,3,2,1,0]
+; AVX512DQVL-NEXT:    vpsllw $5, %ymm1, %ymm1
+; AVX512DQVL-NEXT:    vpsrlw $4, %ymm0, %ymm2
+; AVX512DQVL-NEXT:    vpand {{.*}}(%rip), %ymm2, %ymm2
+; AVX512DQVL-NEXT:    vpblendvb %ymm1, %ymm2, %ymm0, %ymm0
+; AVX512DQVL-NEXT:    vpsrlw $2, %ymm0, %ymm2
+; AVX512DQVL-NEXT:    vpand {{.*}}(%rip), %ymm2, %ymm2
+; AVX512DQVL-NEXT:    vpaddb %ymm1, %ymm1, %ymm1
+; AVX512DQVL-NEXT:    vpblendvb %ymm1, %ymm2, %ymm0, %ymm0
+; AVX512DQVL-NEXT:    vpsrlw $1, %ymm0, %ymm2
+; AVX512DQVL-NEXT:    vpand {{.*}}(%rip), %ymm2, %ymm2
+; AVX512DQVL-NEXT:    vpaddb %ymm1, %ymm1, %ymm1
+; AVX512DQVL-NEXT:    vpblendvb %ymm1, %ymm2, %ymm0, %ymm0
+; AVX512DQVL-NEXT:    retq
+;
+; AVX512BWVL-LABEL: constant_shift_v32i8:
+; AVX512BWVL:       # BB#0:
+; AVX512BWVL-NEXT:    vpmovzxbw {{.*#+}} zmm0 = ymm0[0],zero,ymm0[1],zero,ymm0[2],zero,ymm0[3],zero,ymm0[4],zero,ymm0[5],zero,ymm0[6],zero,ymm0[7],zero,ymm0[8],zero,ymm0[9],zero,ymm0[10],zero,ymm0[11],zero,ymm0[12],zero,ymm0[13],zero,ymm0[14],zero,ymm0[15],zero,ymm0[16],zero,ymm0[17],zero,ymm0[18],zero,ymm0[19],zero,ymm0[20],zero,ymm0[21],zero,ymm0[22],zero,ymm0[23],zero,ymm0[24],zero,ymm0[25],zero,ymm0[26],zero,ymm0[27],zero,ymm0[28],zero,ymm0[29],zero,ymm0[30],zero,ymm0[31],zero
+; AVX512BWVL-NEXT:    vpsrlvw {{.*}}(%rip), %zmm0, %zmm0
+; AVX512BWVL-NEXT:    vpmovwb %zmm0, %ymm0
+; AVX512BWVL-NEXT:    retq
   %shift = lshr <32 x i8> %a, <i8 0, i8 1, i8 2, i8 3, i8 4, i8 5, i8 6, i8 7, i8 7, i8 6, i8 5, i8 4, i8 3, i8 2, i8 1, i8 0, i8 0, i8 1, i8 2, i8 3, i8 4, i8 5, i8 6, i8 7, i8 7, i8 6, i8 5, i8 4, i8 3, i8 2, i8 1, i8 0>
   ret <32 x i8> %shift
 }
@@ -834,6 +973,11 @@ define <4 x i64> @splatconstant_shift_v4i64(<4 x i64> %a) nounwind {
 ; AVX512:       # BB#0:
 ; AVX512-NEXT:    vpsrlq $7, %ymm0, %ymm0
 ; AVX512-NEXT:    retq
+;
+; AVX512VL-LABEL: splatconstant_shift_v4i64:
+; AVX512VL:       # BB#0:
+; AVX512VL-NEXT:    vpsrlq $7, %ymm0, %ymm0
+; AVX512VL-NEXT:    retq
   %shift = lshr <4 x i64> %a, <i64 7, i64 7, i64 7, i64 7>
   ret <4 x i64> %shift
 }
@@ -869,6 +1013,11 @@ define <8 x i32> @splatconstant_shift_v8i32(<8 x i32> %a) nounwind {
 ; AVX512:       # BB#0:
 ; AVX512-NEXT:    vpsrld $5, %ymm0, %ymm0
 ; AVX512-NEXT:    retq
+;
+; AVX512VL-LABEL: splatconstant_shift_v8i32:
+; AVX512VL:       # BB#0:
+; AVX512VL-NEXT:    vpsrld $5, %ymm0, %ymm0
+; AVX512VL-NEXT:    retq
   %shift = lshr <8 x i32> %a, <i32 5, i32 5, i32 5, i32 5, i32 5, i32 5, i32 5, i32 5>
   ret <8 x i32> %shift
 }
@@ -904,6 +1053,11 @@ define <16 x i16> @splatconstant_shift_v16i16(<16 x i16> %a) nounwind {
 ; AVX512:       # BB#0:
 ; AVX512-NEXT:    vpsrlw $3, %ymm0, %ymm0
 ; AVX512-NEXT:    retq
+;
+; AVX512VL-LABEL: splatconstant_shift_v16i16:
+; AVX512VL:       # BB#0:
+; AVX512VL-NEXT:    vpsrlw $3, %ymm0, %ymm0
+; AVX512VL-NEXT:    retq
   %shift = lshr <16 x i16> %a, <i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3>
   ret <16 x i16> %shift
 }
@@ -947,6 +1101,12 @@ define <32 x i8> @splatconstant_shift_v32i8(<32 x i8> %a) nounwind {
 ; AVX512-NEXT:    vpsrlw $3, %ymm0, %ymm0
 ; AVX512-NEXT:    vpand {{.*}}(%rip), %ymm0, %ymm0
 ; AVX512-NEXT:    retq
+;
+; AVX512VL-LABEL: splatconstant_shift_v32i8:
+; AVX512VL:       # BB#0:
+; AVX512VL-NEXT:    vpsrlw $3, %ymm0, %ymm0
+; AVX512VL-NEXT:    vpand {{.*}}(%rip), %ymm0, %ymm0
+; AVX512VL-NEXT:    retq
   %shift = lshr <32 x i8> %a, <i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3>
   ret <32 x i8> %shift
 }

diff --git a/test/CodeGen/X86/vector-shift-shl-128.ll b/test/CodeGen/X86/vector-shift-shl-128.ll
index 8706078b40c99b81d2900cfcee8e8873902fc9e0..32334420f8b2210ee99cbf09473f6db1434b1e4f 100644 (file)
--- a/test/CodeGen/X86/vector-shift-shl-128.ll
+++ b/test/CodeGen/X86/vector-shift-shl-128.ll
@@ -7,6 +7,8 @@
 ; RUN: llc < %s -mtriple=x86_64-unknown-unknown -mattr=+xop,+avx2 | FileCheck %s --check-prefix=ALL --check-prefix=XOP --check-prefix=XOPAVX2
 ; RUN: llc < %s -mtriple=x86_64-unknown-unknown -mcpu=knl -mattr=+avx512dq | FileCheck %s --check-prefix=ALL --check-prefix=AVX512 --check-prefix=AVX512DQ
 ; RUN: llc < %s -mtriple=x86_64-unknown-unknown -mcpu=knl -mattr=+avx512bw | FileCheck %s --check-prefix=ALL --check-prefix=AVX512 --check-prefix=AVX512BW
+; RUN: llc < %s -mtriple=x86_64-unknown-unknown -mcpu=knl -mattr=+avx512dq,+avx512vl | FileCheck %s --check-prefix=ALL --check-prefix=AVX512VL --check-prefix=AVX512DQVL
+; RUN: llc < %s -mtriple=x86_64-unknown-unknown -mcpu=knl -mattr=+avx512bw,+avx512vl | FileCheck %s --check-prefix=ALL --check-prefix=AVX512VL --check-prefix=AVX512BWVL
 ;
 ; Just one 32-bit run to make sure we do reasonable things for i64 shifts.
 ; RUN: llc < %s -mtriple=i686-unknown-unknown -mattr=+sse2 | FileCheck %s --check-prefix=ALL --check-prefix=X32-SSE --check-prefix=X32-SSE2
@@ -63,6 +65,11 @@ define <2 x i64> @var_shift_v2i64(<2 x i64> %a, <2 x i64> %b) nounwind {
 ; AVX512-NEXT:    vpsllvq %xmm1, %xmm0, %xmm0
 ; AVX512-NEXT:    retq
 ;
+; AVX512VL-LABEL: var_shift_v2i64:
+; AVX512VL:       # BB#0:
+; AVX512VL-NEXT:    vpsllvq %xmm1, %xmm0, %xmm0
+; AVX512VL-NEXT:    retq
+;
 ; X32-SSE-LABEL: var_shift_v2i64:
 ; X32-SSE:       # BB#0:
 ; X32-SSE-NEXT:    pshufd {{.*#+}} xmm3 = xmm1[2,3,0,1]
@@ -128,6 +135,11 @@ define <4 x i32> @var_shift_v4i32(<4 x i32> %a, <4 x i32> %b) nounwind {
 ; AVX512-NEXT:    vpsllvd %xmm1, %xmm0, %xmm0
 ; AVX512-NEXT:    retq
 ;
+; AVX512VL-LABEL: var_shift_v4i32:
+; AVX512VL:       # BB#0:
+; AVX512VL-NEXT:    vpsllvd %xmm1, %xmm0, %xmm0
+; AVX512VL-NEXT:    retq
+;
 ; X32-SSE-LABEL: var_shift_v4i32:
 ; X32-SSE:       # BB#0:
 ; X32-SSE-NEXT:    pslld $23, %xmm1
@@ -263,6 +275,19 @@ define <8 x i16> @var_shift_v8i16(<8 x i16> %a, <8 x i16> %b) nounwind {
 ; AVX512BW-NEXT:    # kill: %XMM0<def> %XMM0<kill> %ZMM0<kill>
 ; AVX512BW-NEXT:    retq
 ;
+; AVX512DQVL-LABEL: var_shift_v8i16:
+; AVX512DQVL:       # BB#0:
+; AVX512DQVL-NEXT:    vpmovzxwd {{.*#+}} ymm1 = xmm1[0],zero,xmm1[1],zero,xmm1[2],zero,xmm1[3],zero,xmm1[4],zero,xmm1[5],zero,xmm1[6],zero,xmm1[7],zero
+; AVX512DQVL-NEXT:    vpmovzxwd {{.*#+}} ymm0 = xmm0[0],zero,xmm0[1],zero,xmm0[2],zero,xmm0[3],zero,xmm0[4],zero,xmm0[5],zero,xmm0[6],zero,xmm0[7],zero
+; AVX512DQVL-NEXT:    vpsllvd %ymm1, %ymm0, %ymm0
+; AVX512DQVL-NEXT:    vpmovdw %ymm0, %xmm0
+; AVX512DQVL-NEXT:    retq
+;
+; AVX512BWVL-LABEL: var_shift_v8i16:
+; AVX512BWVL:       # BB#0:
+; AVX512BWVL-NEXT:    vpsllvw %xmm1, %xmm0, %xmm0
+; AVX512BWVL-NEXT:    retq
+;
 ; X32-SSE-LABEL: var_shift_v8i16:
 ; X32-SSE:       # BB#0:
 ; X32-SSE-NEXT:    psllw $12, %xmm1
@@ -383,6 +408,14 @@ define <16 x i8> @var_shift_v16i8(<16 x i8> %a, <16 x i8> %b) nounwind {
 ; AVX512-NEXT:    vpmovdb %zmm0, %xmm0
 ; AVX512-NEXT:    retq
 ;
+; AVX512VL-LABEL: var_shift_v16i8:
+; AVX512VL:       # BB#0:
+; AVX512VL-NEXT:    vpmovzxbd {{.*#+}} zmm1 = xmm1[0],zero,zero,zero,xmm1[1],zero,zero,zero,xmm1[2],zero,zero,zero,xmm1[3],zero,zero,zero,xmm1[4],zero,zero,zero,xmm1[5],zero,zero,zero,xmm1[6],zero,zero,zero,xmm1[7],zero,zero,zero,xmm1[8],zero,zero,zero,xmm1[9],zero,zero,zero,xmm1[10],zero,zero,zero,xmm1[11],zero,zero,zero,xmm1[12],zero,zero,zero,xmm1[13],zero,zero,zero,xmm1[14],zero,zero,zero,xmm1[15],zero,zero,zero
+; AVX512VL-NEXT:    vpmovzxbd {{.*#+}} zmm0 = xmm0[0],zero,zero,zero,xmm0[1],zero,zero,zero,xmm0[2],zero,zero,zero,xmm0[3],zero,zero,zero,xmm0[4],zero,zero,zero,xmm0[5],zero,zero,zero,xmm0[6],zero,zero,zero,xmm0[7],zero,zero,zero,xmm0[8],zero,zero,zero,xmm0[9],zero,zero,zero,xmm0[10],zero,zero,zero,xmm0[11],zero,zero,zero,xmm0[12],zero,zero,zero,xmm0[13],zero,zero,zero,xmm0[14],zero,zero,zero,xmm0[15],zero,zero,zero
+; AVX512VL-NEXT:    vpsllvd %zmm1, %zmm0, %zmm0
+; AVX512VL-NEXT:    vpmovdb %zmm0, %xmm0
+; AVX512VL-NEXT:    retq
+;
 ; X32-SSE-LABEL: var_shift_v16i8:
 ; X32-SSE:       # BB#0:
 ; X32-SSE-NEXT:    psllw $5, %xmm1
@@ -441,6 +474,11 @@ define <2 x i64> @splatvar_shift_v2i64(<2 x i64> %a, <2 x i64> %b) nounwind {
 ; AVX512-NEXT:    vpsllq %xmm1, %xmm0, %xmm0
 ; AVX512-NEXT:    retq
 ;
+; AVX512VL-LABEL: splatvar_shift_v2i64:
+; AVX512VL:       # BB#0:
+; AVX512VL-NEXT:    vpsllq %xmm1, %xmm0, %xmm0
+; AVX512VL-NEXT:    retq
+;
 ; X32-SSE-LABEL: splatvar_shift_v2i64:
 ; X32-SSE:       # BB#0:
 ; X32-SSE-NEXT:    psllq %xmm1, %xmm0
@@ -482,6 +520,12 @@ define <4 x i32> @splatvar_shift_v4i32(<4 x i32> %a, <4 x i32> %b) nounwind {
 ; AVX512-NEXT:    vpslld %xmm1, %xmm0, %xmm0
 ; AVX512-NEXT:    retq
 ;
+; AVX512VL-LABEL: splatvar_shift_v4i32:
+; AVX512VL:       # BB#0:
+; AVX512VL-NEXT:    vpmovzxdq {{.*#+}} xmm1 = xmm1[0],zero,xmm1[1],zero
+; AVX512VL-NEXT:    vpslld %xmm1, %xmm0, %xmm0
+; AVX512VL-NEXT:    retq
+;
 ; X32-SSE-LABEL: splatvar_shift_v4i32:
 ; X32-SSE:       # BB#0:
 ; X32-SSE-NEXT:    xorps %xmm2, %xmm2
@@ -525,6 +569,12 @@ define <8 x i16> @splatvar_shift_v8i16(<8 x i16> %a, <8 x i16> %b) nounwind {
 ; AVX512-NEXT:    vpsllw %xmm1, %xmm0, %xmm0
 ; AVX512-NEXT:    retq
 ;
+; AVX512VL-LABEL: splatvar_shift_v8i16:
+; AVX512VL:       # BB#0:
+; AVX512VL-NEXT:    vpmovzxwq {{.*#+}} xmm1 = xmm1[0],zero,zero,zero,xmm1[1],zero,zero,zero
+; AVX512VL-NEXT:    vpsllw %xmm1, %xmm0, %xmm0
+; AVX512VL-NEXT:    retq
+;
 ; X32-SSE-LABEL: splatvar_shift_v8i16:
 ; X32-SSE:       # BB#0:
 ; X32-SSE-NEXT:    pextrw $0, %xmm1, %eax
@@ -651,6 +701,15 @@ define <16 x i8> @splatvar_shift_v16i8(<16 x i8> %a, <16 x i8> %b) nounwind {
 ; AVX512-NEXT:    vpmovdb %zmm0, %xmm0
 ; AVX512-NEXT:    retq
 ;
+; AVX512VL-LABEL: splatvar_shift_v16i8:
+; AVX512VL:       # BB#0:
+; AVX512VL-NEXT:    vpbroadcastb %xmm1, %xmm1
+; AVX512VL-NEXT:    vpmovzxbd {{.*#+}} zmm0 = xmm0[0],zero,zero,zero,xmm0[1],zero,zero,zero,xmm0[2],zero,zero,zero,xmm0[3],zero,zero,zero,xmm0[4],zero,zero,zero,xmm0[5],zero,zero,zero,xmm0[6],zero,zero,zero,xmm0[7],zero,zero,zero,xmm0[8],zero,zero,zero,xmm0[9],zero,zero,zero,xmm0[10],zero,zero,zero,xmm0[11],zero,zero,zero,xmm0[12],zero,zero,zero,xmm0[13],zero,zero,zero,xmm0[14],zero,zero,zero,xmm0[15],zero,zero,zero
+; AVX512VL-NEXT:    vpmovzxbd {{.*#+}} zmm1 = xmm1[0],zero,zero,zero,xmm1[1],zero,zero,zero,xmm1[2],zero,zero,zero,xmm1[3],zero,zero,zero,xmm1[4],zero,zero,zero,xmm1[5],zero,zero,zero,xmm1[6],zero,zero,zero,xmm1[7],zero,zero,zero,xmm1[8],zero,zero,zero,xmm1[9],zero,zero,zero,xmm1[10],zero,zero,zero,xmm1[11],zero,zero,zero,xmm1[12],zero,zero,zero,xmm1[13],zero,zero,zero,xmm1[14],zero,zero,zero,xmm1[15],zero,zero,zero
+; AVX512VL-NEXT:    vpsllvd %zmm1, %zmm0, %zmm0
+; AVX512VL-NEXT:    vpmovdb %zmm0, %xmm0
+; AVX512VL-NEXT:    retq
+;
 ; X32-SSE-LABEL: splatvar_shift_v16i8:
 ; X32-SSE:       # BB#0:
 ; X32-SSE-NEXT:    punpcklbw {{.*#+}} xmm1 = xmm1[0,0,1,1,2,2,3,3,4,4,5,5,6,6,7,7]
@@ -737,6 +796,11 @@ define <2 x i64> @constant_shift_v2i64(<2 x i64> %a) nounwind {
 ; AVX512-NEXT:    vpsllvq {{.*}}(%rip), %xmm0, %xmm0
 ; AVX512-NEXT:    retq
 ;
+; AVX512VL-LABEL: constant_shift_v2i64:
+; AVX512VL:       # BB#0:
+; AVX512VL-NEXT:    vpsllvq {{.*}}(%rip), %xmm0, %xmm0
+; AVX512VL-NEXT:    retq
+;
 ; X32-SSE-LABEL: constant_shift_v2i64:
 ; X32-SSE:       # BB#0:
 ; X32-SSE-NEXT:    movdqa %xmm0, %xmm1
@@ -792,6 +856,11 @@ define <4 x i32> @constant_shift_v4i32(<4 x i32> %a) nounwind {
 ; AVX512-NEXT:    vpsllvd {{.*}}(%rip), %xmm0, %xmm0
 ; AVX512-NEXT:    retq
 ;
+; AVX512VL-LABEL: constant_shift_v4i32:
+; AVX512VL:       # BB#0:
+; AVX512VL-NEXT:    vpsllvd {{.*}}(%rip), %xmm0, %xmm0
+; AVX512VL-NEXT:    retq
+;
 ; X32-SSE-LABEL: constant_shift_v4i32:
 ; X32-SSE:       # BB#0:
 ; X32-SSE-NEXT:    movdqa {{.*#+}} xmm1 = [16,32,64,128]
@@ -836,6 +905,16 @@ define <8 x i16> @constant_shift_v8i16(<8 x i16> %a) nounwind {
 ; AVX512BW-NEXT:    # kill: %XMM0<def> %XMM0<kill> %ZMM0<kill>
 ; AVX512BW-NEXT:    retq
 ;
+; AVX512DQVL-LABEL: constant_shift_v8i16:
+; AVX512DQVL:       # BB#0:
+; AVX512DQVL-NEXT:    vpmullw {{.*}}(%rip), %xmm0, %xmm0
+; AVX512DQVL-NEXT:    retq
+;
+; AVX512BWVL-LABEL: constant_shift_v8i16:
+; AVX512BWVL:       # BB#0:
+; AVX512BWVL-NEXT:    vpsllvw {{.*}}(%rip), %xmm0, %xmm0
+; AVX512BWVL-NEXT:    retq
+;
 ; X32-SSE-LABEL: constant_shift_v8i16:
 ; X32-SSE:       # BB#0:
 ; X32-SSE-NEXT:    pmullw {{\.LCPI.*}}, %xmm0
@@ -925,6 +1004,13 @@ define <16 x i8> @constant_shift_v16i8(<16 x i8> %a) nounwind {
 ; AVX512-NEXT:    vpmovdb %zmm0, %xmm0
 ; AVX512-NEXT:    retq
 ;
+; AVX512VL-LABEL: constant_shift_v16i8:
+; AVX512VL:       # BB#0:
+; AVX512VL-NEXT:    vpmovzxbd {{.*#+}} zmm0 = xmm0[0],zero,zero,zero,xmm0[1],zero,zero,zero,xmm0[2],zero,zero,zero,xmm0[3],zero,zero,zero,xmm0[4],zero,zero,zero,xmm0[5],zero,zero,zero,xmm0[6],zero,zero,zero,xmm0[7],zero,zero,zero,xmm0[8],zero,zero,zero,xmm0[9],zero,zero,zero,xmm0[10],zero,zero,zero,xmm0[11],zero,zero,zero,xmm0[12],zero,zero,zero,xmm0[13],zero,zero,zero,xmm0[14],zero,zero,zero,xmm0[15],zero,zero,zero
+; AVX512VL-NEXT:    vpsllvd {{.*}}(%rip), %zmm0, %zmm0
+; AVX512VL-NEXT:    vpmovdb %zmm0, %xmm0
+; AVX512VL-NEXT:    retq
+;
 ; X32-SSE-LABEL: constant_shift_v16i8:
 ; X32-SSE:       # BB#0:
 ; X32-SSE-NEXT:    movdqa {{.*#+}} xmm2 = [0,1,2,3,4,5,6,7,7,6,5,4,3,2,1,0]
@@ -984,6 +1070,11 @@ define <2 x i64> @splatconstant_shift_v2i64(<2 x i64> %a) nounwind {
 ; AVX512-NEXT:    vpsllq $7, %xmm0, %xmm0
 ; AVX512-NEXT:    retq
 ;
+; AVX512VL-LABEL: splatconstant_shift_v2i64:
+; AVX512VL:       # BB#0:
+; AVX512VL-NEXT:    vpsllq $7, %xmm0, %xmm0
+; AVX512VL-NEXT:    retq
+;
 ; X32-SSE-LABEL: splatconstant_shift_v2i64:
 ; X32-SSE:       # BB#0:
 ; X32-SSE-NEXT:    psllq $7, %xmm0
@@ -1013,6 +1104,11 @@ define <4 x i32> @splatconstant_shift_v4i32(<4 x i32> %a) nounwind {
 ; AVX512-NEXT:    vpslld $5, %xmm0, %xmm0
 ; AVX512-NEXT:    retq
 ;
+; AVX512VL-LABEL: splatconstant_shift_v4i32:
+; AVX512VL:       # BB#0:
+; AVX512VL-NEXT:    vpslld $5, %xmm0, %xmm0
+; AVX512VL-NEXT:    retq
+;
 ; X32-SSE-LABEL: splatconstant_shift_v4i32:
 ; X32-SSE:       # BB#0:
 ; X32-SSE-NEXT:    pslld $5, %xmm0
@@ -1042,6 +1138,11 @@ define <8 x i16> @splatconstant_shift_v8i16(<8 x i16> %a) nounwind {
 ; AVX512-NEXT:    vpsllw $3, %xmm0, %xmm0
 ; AVX512-NEXT:    retq
 ;
+; AVX512VL-LABEL: splatconstant_shift_v8i16:
+; AVX512VL:       # BB#0:
+; AVX512VL-NEXT:    vpsllw $3, %xmm0, %xmm0
+; AVX512VL-NEXT:    retq
+;
 ; X32-SSE-LABEL: splatconstant_shift_v8i16:
 ; X32-SSE:       # BB#0:
 ; X32-SSE-NEXT:    psllw $3, %xmm0
@@ -1074,6 +1175,12 @@ define <16 x i8> @splatconstant_shift_v16i8(<16 x i8> %a) nounwind {
 ; AVX512-NEXT:    vpand {{.*}}(%rip), %xmm0, %xmm0
 ; AVX512-NEXT:    retq
 ;
+; AVX512VL-LABEL: splatconstant_shift_v16i8:
+; AVX512VL:       # BB#0:
+; AVX512VL-NEXT:    vpsllw $3, %xmm0, %xmm0
+; AVX512VL-NEXT:    vpand {{.*}}(%rip), %xmm0, %xmm0
+; AVX512VL-NEXT:    retq
+;
 ; X32-SSE-LABEL: splatconstant_shift_v16i8:
 ; X32-SSE:       # BB#0:
 ; X32-SSE-NEXT:    psllw $3, %xmm0

diff --git a/test/CodeGen/X86/vector-shift-shl-256.ll b/test/CodeGen/X86/vector-shift-shl-256.ll
index a1ef2791c1b030344b0bef45ab1d249733d33ca5..104fa089c74495de5085103d5902e3da50939b78 100644 (file)
--- a/test/CodeGen/X86/vector-shift-shl-256.ll
+++ b/test/CodeGen/X86/vector-shift-shl-256.ll
@@ -5,6 +5,8 @@
 ; RUN: llc < %s -mtriple=x86_64-unknown-unknown -mattr=+xop,+avx2 | FileCheck %s --check-prefix=ALL --check-prefix=XOP --check-prefix=XOPAVX2
 ; RUN: llc < %s -mtriple=x86_64-unknown-unknown -mcpu=knl -mattr=+avx512dq | FileCheck %s --check-prefix=ALL --check-prefix=AVX512 --check-prefix=AVX512DQ
 ; RUN: llc < %s -mtriple=x86_64-unknown-unknown -mcpu=knl -mattr=+avx512bw | FileCheck %s --check-prefix=ALL --check-prefix=AVX512 --check-prefix=AVX512BW
+; RUN: llc < %s -mtriple=x86_64-unknown-unknown -mcpu=knl -mattr=+avx512dq,+avx512vl | FileCheck %s --check-prefix=ALL --check-prefix=AVX512VL --check-prefix=AVX512DQVL
+; RUN: llc < %s -mtriple=x86_64-unknown-unknown -mcpu=knl -mattr=+avx512bw,+avx512vl | FileCheck %s --check-prefix=ALL --check-prefix=AVX512VL --check-prefix=AVX512BWVL
 
 ;
 ; Variable Shifts
@@ -49,6 +51,11 @@ define <4 x i64> @var_shift_v4i64(<4 x i64> %a, <4 x i64> %b) nounwind {
 ; AVX512:       # BB#0:
 ; AVX512-NEXT:    vpsllvq %ymm1, %ymm0, %ymm0
 ; AVX512-NEXT:    retq
+;
+; AVX512VL-LABEL: var_shift_v4i64:
+; AVX512VL:       # BB#0:
+; AVX512VL-NEXT:    vpsllvq %ymm1, %ymm0, %ymm0
+; AVX512VL-NEXT:    retq
   %shift = shl <4 x i64> %a, %b
   ret <4 x i64> %shift
 }
@@ -93,6 +100,11 @@ define <8 x i32> @var_shift_v8i32(<8 x i32> %a, <8 x i32> %b) nounwind {
 ; AVX512:       # BB#0:
 ; AVX512-NEXT:    vpsllvd %ymm1, %ymm0, %ymm0
 ; AVX512-NEXT:    retq
+;
+; AVX512VL-LABEL: var_shift_v8i32:
+; AVX512VL:       # BB#0:
+; AVX512VL-NEXT:    vpsllvd %ymm1, %ymm0, %ymm0
+; AVX512VL-NEXT:    retq
   %shift = shl <8 x i32> %a, %b
   ret <8 x i32> %shift
 }
@@ -180,6 +192,19 @@ define <16 x i16> @var_shift_v16i16(<16 x i16> %a, <16 x i16> %b) nounwind {
 ; AVX512BW-NEXT:    vpsllvw %zmm1, %zmm0, %zmm0
 ; AVX512BW-NEXT:    # kill: %YMM0<def> %YMM0<kill> %ZMM0<kill>
 ; AVX512BW-NEXT:    retq
+;
+; AVX512DQVL-LABEL: var_shift_v16i16:
+; AVX512DQVL:       # BB#0:
+; AVX512DQVL-NEXT:    vpmovzxwd {{.*#+}} zmm1 = ymm1[0],zero,ymm1[1],zero,ymm1[2],zero,ymm1[3],zero,ymm1[4],zero,ymm1[5],zero,ymm1[6],zero,ymm1[7],zero,ymm1[8],zero,ymm1[9],zero,ymm1[10],zero,ymm1[11],zero,ymm1[12],zero,ymm1[13],zero,ymm1[14],zero,ymm1[15],zero
+; AVX512DQVL-NEXT:    vpmovzxwd {{.*#+}} zmm0 = ymm0[0],zero,ymm0[1],zero,ymm0[2],zero,ymm0[3],zero,ymm0[4],zero,ymm0[5],zero,ymm0[6],zero,ymm0[7],zero,ymm0[8],zero,ymm0[9],zero,ymm0[10],zero,ymm0[11],zero,ymm0[12],zero,ymm0[13],zero,ymm0[14],zero,ymm0[15],zero
+; AVX512DQVL-NEXT:    vpsllvd %zmm1, %zmm0, %zmm0
+; AVX512DQVL-NEXT:    vpmovdw %zmm0, %ymm0
+; AVX512DQVL-NEXT:    retq
+;
+; AVX512BWVL-LABEL: var_shift_v16i16:
+; AVX512BWVL:       # BB#0:
+; AVX512BWVL-NEXT:    vpsllvw %ymm1, %ymm0, %ymm0
+; AVX512BWVL-NEXT:    retq
   %shift = shl <16 x i16> %a, %b
   ret <16 x i16> %shift
 }
@@ -271,6 +296,29 @@ define <32 x i8> @var_shift_v32i8(<32 x i8> %a, <32 x i8> %b) nounwind {
 ; AVX512BW-NEXT:    vpsllvw %zmm1, %zmm0, %zmm0
 ; AVX512BW-NEXT:    vpmovwb %zmm0, %ymm0
 ; AVX512BW-NEXT:    retq
+;
+; AVX512DQVL-LABEL: var_shift_v32i8:
+; AVX512DQVL:       # BB#0:
+; AVX512DQVL-NEXT:    vpsllw $5, %ymm1, %ymm1
+; AVX512DQVL-NEXT:    vpsllw $4, %ymm0, %ymm2
+; AVX512DQVL-NEXT:    vpand {{.*}}(%rip), %ymm2, %ymm2
+; AVX512DQVL-NEXT:    vpblendvb %ymm1, %ymm2, %ymm0, %ymm0
+; AVX512DQVL-NEXT:    vpsllw $2, %ymm0, %ymm2
+; AVX512DQVL-NEXT:    vpand {{.*}}(%rip), %ymm2, %ymm2
+; AVX512DQVL-NEXT:    vpaddb %ymm1, %ymm1, %ymm1
+; AVX512DQVL-NEXT:    vpblendvb %ymm1, %ymm2, %ymm0, %ymm0
+; AVX512DQVL-NEXT:    vpaddb %ymm0, %ymm0, %ymm2
+; AVX512DQVL-NEXT:    vpaddb %ymm1, %ymm1, %ymm1
+; AVX512DQVL-NEXT:    vpblendvb %ymm1, %ymm2, %ymm0, %ymm0
+; AVX512DQVL-NEXT:    retq
+;
+; AVX512BWVL-LABEL: var_shift_v32i8:
+; AVX512BWVL:       # BB#0:
+; AVX512BWVL-NEXT:    vpmovzxbw {{.*#+}} zmm1 = ymm1[0],zero,ymm1[1],zero,ymm1[2],zero,ymm1[3],zero,ymm1[4],zero,ymm1[5],zero,ymm1[6],zero,ymm1[7],zero,ymm1[8],zero,ymm1[9],zero,ymm1[10],zero,ymm1[11],zero,ymm1[12],zero,ymm1[13],zero,ymm1[14],zero,ymm1[15],zero,ymm1[16],zero,ymm1[17],zero,ymm1[18],zero,ymm1[19],zero,ymm1[20],zero,ymm1[21],zero,ymm1[22],zero,ymm1[23],zero,ymm1[24],zero,ymm1[25],zero,ymm1[26],zero,ymm1[27],zero,ymm1[28],zero,ymm1[29],zero,ymm1[30],zero,ymm1[31],zero
+; AVX512BWVL-NEXT:    vpmovzxbw {{.*#+}} zmm0 = ymm0[0],zero,ymm0[1],zero,ymm0[2],zero,ymm0[3],zero,ymm0[4],zero,ymm0[5],zero,ymm0[6],zero,ymm0[7],zero,ymm0[8],zero,ymm0[9],zero,ymm0[10],zero,ymm0[11],zero,ymm0[12],zero,ymm0[13],zero,ymm0[14],zero,ymm0[15],zero,ymm0[16],zero,ymm0[17],zero,ymm0[18],zero,ymm0[19],zero,ymm0[20],zero,ymm0[21],zero,ymm0[22],zero,ymm0[23],zero,ymm0[24],zero,ymm0[25],zero,ymm0[26],zero,ymm0[27],zero,ymm0[28],zero,ymm0[29],zero,ymm0[30],zero,ymm0[31],zero
+; AVX512BWVL-NEXT:    vpsllvw %zmm1, %zmm0, %zmm0
+; AVX512BWVL-NEXT:    vpmovwb %zmm0, %ymm0
+; AVX512BWVL-NEXT:    retq
   %shift = shl <32 x i8> %a, %b
   ret <32 x i8> %shift
 }
@@ -310,6 +358,11 @@ define <4 x i64> @splatvar_shift_v4i64(<4 x i64> %a, <4 x i64> %b) nounwind {
 ; AVX512:       # BB#0:
 ; AVX512-NEXT:    vpsllq %xmm1, %ymm0, %ymm0
 ; AVX512-NEXT:    retq
+;
+; AVX512VL-LABEL: splatvar_shift_v4i64:
+; AVX512VL:       # BB#0:
+; AVX512VL-NEXT:    vpsllq %xmm1, %ymm0, %ymm0
+; AVX512VL-NEXT:    retq
   %splat = shufflevector <4 x i64> %b, <4 x i64> undef, <4 x i32> zeroinitializer
   %shift = shl <4 x i64> %a, %splat
   ret <4 x i64> %shift
@@ -351,6 +404,12 @@ define <8 x i32> @splatvar_shift_v8i32(<8 x i32> %a, <8 x i32> %b) nounwind {
 ; AVX512-NEXT:    vpmovzxdq {{.*#+}} xmm1 = xmm1[0],zero,xmm1[1],zero
 ; AVX512-NEXT:    vpslld %xmm1, %ymm0, %ymm0
 ; AVX512-NEXT:    retq
+;
+; AVX512VL-LABEL: splatvar_shift_v8i32:
+; AVX512VL:       # BB#0:
+; AVX512VL-NEXT:    vpmovzxdq {{.*#+}} xmm1 = xmm1[0],zero,xmm1[1],zero
+; AVX512VL-NEXT:    vpslld %xmm1, %ymm0, %ymm0
+; AVX512VL-NEXT:    retq
   %splat = shufflevector <8 x i32> %b, <8 x i32> undef, <8 x i32> zeroinitializer
   %shift = shl <8 x i32> %a, %splat
   ret <8 x i32> %shift
@@ -392,6 +451,12 @@ define <16 x i16> @splatvar_shift_v16i16(<16 x i16> %a, <16 x i16> %b) nounwind
 ; AVX512-NEXT:    vpmovzxwq {{.*#+}} xmm1 = xmm1[0],zero,zero,zero,xmm1[1],zero,zero,zero
 ; AVX512-NEXT:    vpsllw %xmm1, %ymm0, %ymm0
 ; AVX512-NEXT:    retq
+;
+; AVX512VL-LABEL: splatvar_shift_v16i16:
+; AVX512VL:       # BB#0:
+; AVX512VL-NEXT:    vpmovzxwq {{.*#+}} xmm1 = xmm1[0],zero,zero,zero,xmm1[1],zero,zero,zero
+; AVX512VL-NEXT:    vpsllw %xmm1, %ymm0, %ymm0
+; AVX512VL-NEXT:    retq
   %splat = shufflevector <16 x i16> %b, <16 x i16> undef, <16 x i32> zeroinitializer
   %shift = shl <16 x i16> %a, %splat
   ret <16 x i16> %shift
@@ -487,6 +552,31 @@ define <32 x i8> @splatvar_shift_v32i8(<32 x i8> %a, <32 x i8> %b) nounwind {
 ; AVX512BW-NEXT:    vpsllvw %zmm1, %zmm0, %zmm0
 ; AVX512BW-NEXT:    vpmovwb %zmm0, %ymm0
 ; AVX512BW-NEXT:    retq
+;
+; AVX512DQVL-LABEL: splatvar_shift_v32i8:
+; AVX512DQVL:       # BB#0:
+; AVX512DQVL-NEXT:    vpbroadcastb %xmm1, %ymm1
+; AVX512DQVL-NEXT:    vpsllw $4, %ymm0, %ymm2
+; AVX512DQVL-NEXT:    vpand {{.*}}(%rip), %ymm2, %ymm2
+; AVX512DQVL-NEXT:    vpsllw $5, %ymm1, %ymm1
+; AVX512DQVL-NEXT:    vpblendvb %ymm1, %ymm2, %ymm0, %ymm0
+; AVX512DQVL-NEXT:    vpsllw $2, %ymm0, %ymm2
+; AVX512DQVL-NEXT:    vpand {{.*}}(%rip), %ymm2, %ymm2
+; AVX512DQVL-NEXT:    vpaddb %ymm1, %ymm1, %ymm1
+; AVX512DQVL-NEXT:    vpblendvb %ymm1, %ymm2, %ymm0, %ymm0
+; AVX512DQVL-NEXT:    vpaddb %ymm0, %ymm0, %ymm2
+; AVX512DQVL-NEXT:    vpaddb %ymm1, %ymm1, %ymm1
+; AVX512DQVL-NEXT:    vpblendvb %ymm1, %ymm2, %ymm0, %ymm0
+; AVX512DQVL-NEXT:    retq
+;
+; AVX512BWVL-LABEL: splatvar_shift_v32i8:
+; AVX512BWVL:       # BB#0:
+; AVX512BWVL-NEXT:    vpbroadcastb %xmm1, %ymm1
+; AVX512BWVL-NEXT:    vpmovzxbw {{.*#+}} zmm0 = ymm0[0],zero,ymm0[1],zero,ymm0[2],zero,ymm0[3],zero,ymm0[4],zero,ymm0[5],zero,ymm0[6],zero,ymm0[7],zero,ymm0[8],zero,ymm0[9],zero,ymm0[10],zero,ymm0[11],zero,ymm0[12],zero,ymm0[13],zero,ymm0[14],zero,ymm0[15],zero,ymm0[16],zero,ymm0[17],zero,ymm0[18],zero,ymm0[19],zero,ymm0[20],zero,ymm0[21],zero,ymm0[22],zero,ymm0[23],zero,ymm0[24],zero,ymm0[25],zero,ymm0[26],zero,ymm0[27],zero,ymm0[28],zero,ymm0[29],zero,ymm0[30],zero,ymm0[31],zero
+; AVX512BWVL-NEXT:    vpmovzxbw {{.*#+}} zmm1 = ymm1[0],zero,ymm1[1],zero,ymm1[2],zero,ymm1[3],zero,ymm1[4],zero,ymm1[5],zero,ymm1[6],zero,ymm1[7],zero,ymm1[8],zero,ymm1[9],zero,ymm1[10],zero,ymm1[11],zero,ymm1[12],zero,ymm1[13],zero,ymm1[14],zero,ymm1[15],zero,ymm1[16],zero,ymm1[17],zero,ymm1[18],zero,ymm1[19],zero,ymm1[20],zero,ymm1[21],zero,ymm1[22],zero,ymm1[23],zero,ymm1[24],zero,ymm1[25],zero,ymm1[26],zero,ymm1[27],zero,ymm1[28],zero,ymm1[29],zero,ymm1[30],zero,ymm1[31],zero
+; AVX512BWVL-NEXT:    vpsllvw %zmm1, %zmm0, %zmm0
+; AVX512BWVL-NEXT:    vpmovwb %zmm0, %ymm0
+; AVX512BWVL-NEXT:    retq
   %splat = shufflevector <32 x i8> %b, <32 x i8> undef, <32 x i32> zeroinitializer
   %shift = shl <32 x i8> %a, %splat
   ret <32 x i8> %shift
@@ -531,6 +621,11 @@ define <4 x i64> @constant_shift_v4i64(<4 x i64> %a) nounwind {
 ; AVX512:       # BB#0:
 ; AVX512-NEXT:    vpsllvq {{.*}}(%rip), %ymm0, %ymm0
 ; AVX512-NEXT:    retq
+;
+; AVX512VL-LABEL: constant_shift_v4i64:
+; AVX512VL:       # BB#0:
+; AVX512VL-NEXT:    vpsllvq {{.*}}(%rip), %ymm0, %ymm0
+; AVX512VL-NEXT:    retq
   %shift = shl <4 x i64> %a, <i64 1, i64 7, i64 31, i64 62>
   ret <4 x i64> %shift
 }
@@ -566,6 +661,11 @@ define <8 x i32> @constant_shift_v8i32(<8 x i32> %a) nounwind {
 ; AVX512:       # BB#0:
 ; AVX512-NEXT:    vpsllvd {{.*}}(%rip), %ymm0, %ymm0
 ; AVX512-NEXT:    retq
+;
+; AVX512VL-LABEL: constant_shift_v8i32:
+; AVX512VL:       # BB#0:
+; AVX512VL-NEXT:    vpsllvd {{.*}}(%rip), %ymm0, %ymm0
+; AVX512VL-NEXT:    retq
   %shift = shl <8 x i32> %a, <i32 4, i32 5, i32 6, i32 7, i32 8, i32 9, i32 8, i32 7>
   ret <8 x i32> %shift
 }
@@ -609,6 +709,16 @@ define <16 x i16> @constant_shift_v16i16(<16 x i16> %a) nounwind {
 ; AVX512BW-NEXT:    vpsllvw %zmm1, %zmm0, %zmm0
 ; AVX512BW-NEXT:    # kill: %YMM0<def> %YMM0<kill> %ZMM0<kill>
 ; AVX512BW-NEXT:    retq
+;
+; AVX512DQVL-LABEL: constant_shift_v16i16:
+; AVX512DQVL:       # BB#0:
+; AVX512DQVL-NEXT:    vpmullw {{.*}}(%rip), %ymm0, %ymm0
+; AVX512DQVL-NEXT:    retq
+;
+; AVX512BWVL-LABEL: constant_shift_v16i16:
+; AVX512BWVL:       # BB#0:
+; AVX512BWVL-NEXT:    vpsllvw {{.*}}(%rip), %ymm0, %ymm0
+; AVX512BWVL-NEXT:    retq
   %shift = shl <16 x i16> %a, <i16 0, i16 1, i16 2, i16 3, i16 4, i16 5, i16 6, i16 7, i16 8, i16 9, i16 10, i16 11, i16 12, i16 13, i16 14, i16 15>
   ret <16 x i16> %shift
 }
@@ -698,6 +808,29 @@ define <32 x i8> @constant_shift_v32i8(<32 x i8> %a) nounwind {
 ; AVX512BW-NEXT:    vpsllvw {{.*}}(%rip), %zmm0, %zmm0
 ; AVX512BW-NEXT:    vpmovwb %zmm0, %ymm0
 ; AVX512BW-NEXT:    retq
+;
+; AVX512DQVL-LABEL: constant_shift_v32i8:
+; AVX512DQVL:       # BB#0:
+; AVX512DQVL-NEXT:    vmovdqa {{.*#+}} ymm1 = [0,1,2,3,4,5,6,7,7,6,5,4,3,2,1,0,0,1,2,3,4,5,6,7,7,6,5,4,3,2,1,0]
+; AVX512DQVL-NEXT:    vpsllw $5, %ymm1, %ymm1
+; AVX512DQVL-NEXT:    vpsllw $4, %ymm0, %ymm2
+; AVX512DQVL-NEXT:    vpand {{.*}}(%rip), %ymm2, %ymm2
+; AVX512DQVL-NEXT:    vpblendvb %ymm1, %ymm2, %ymm0, %ymm0
+; AVX512DQVL-NEXT:    vpsllw $2, %ymm0, %ymm2
+; AVX512DQVL-NEXT:    vpand {{.*}}(%rip), %ymm2, %ymm2
+; AVX512DQVL-NEXT:    vpaddb %ymm1, %ymm1, %ymm1
+; AVX512DQVL-NEXT:    vpblendvb %ymm1, %ymm2, %ymm0, %ymm0
+; AVX512DQVL-NEXT:    vpaddb %ymm0, %ymm0, %ymm2
+; AVX512DQVL-NEXT:    vpaddb %ymm1, %ymm1, %ymm1
+; AVX512DQVL-NEXT:    vpblendvb %ymm1, %ymm2, %ymm0, %ymm0
+; AVX512DQVL-NEXT:    retq
+;
+; AVX512BWVL-LABEL: constant_shift_v32i8:
+; AVX512BWVL:       # BB#0:
+; AVX512BWVL-NEXT:    vpmovzxbw {{.*#+}} zmm0 = ymm0[0],zero,ymm0[1],zero,ymm0[2],zero,ymm0[3],zero,ymm0[4],zero,ymm0[5],zero,ymm0[6],zero,ymm0[7],zero,ymm0[8],zero,ymm0[9],zero,ymm0[10],zero,ymm0[11],zero,ymm0[12],zero,ymm0[13],zero,ymm0[14],zero,ymm0[15],zero,ymm0[16],zero,ymm0[17],zero,ymm0[18],zero,ymm0[19],zero,ymm0[20],zero,ymm0[21],zero,ymm0[22],zero,ymm0[23],zero,ymm0[24],zero,ymm0[25],zero,ymm0[26],zero,ymm0[27],zero,ymm0[28],zero,ymm0[29],zero,ymm0[30],zero,ymm0[31],zero
+; AVX512BWVL-NEXT:    vpsllvw {{.*}}(%rip), %zmm0, %zmm0
+; AVX512BWVL-NEXT:    vpmovwb %zmm0, %ymm0
+; AVX512BWVL-NEXT:    retq
   %shift = shl <32 x i8> %a, <i8 0, i8 1, i8 2, i8 3, i8 4, i8 5, i8 6, i8 7, i8 7, i8 6, i8 5, i8 4, i8 3, i8 2, i8 1, i8 0, i8 0, i8 1, i8 2, i8 3, i8 4, i8 5, i8 6, i8 7, i8 7, i8 6, i8 5, i8 4, i8 3, i8 2, i8 1, i8 0>
   ret <32 x i8> %shift
 }
@@ -737,6 +870,11 @@ define <4 x i64> @splatconstant_shift_v4i64(<4 x i64> %a) nounwind {
 ; AVX512:       # BB#0:
 ; AVX512-NEXT:    vpsllq $7, %ymm0, %ymm0
 ; AVX512-NEXT:    retq
+;
+; AVX512VL-LABEL: splatconstant_shift_v4i64:
+; AVX512VL:       # BB#0:
+; AVX512VL-NEXT:    vpsllq $7, %ymm0, %ymm0
+; AVX512VL-NEXT:    retq
   %shift = shl <4 x i64> %a, <i64 7, i64 7, i64 7, i64 7>
   ret <4 x i64> %shift
 }
@@ -772,6 +910,11 @@ define <8 x i32> @splatconstant_shift_v8i32(<8 x i32> %a) nounwind {
 ; AVX512:       # BB#0:
 ; AVX512-NEXT:    vpslld $5, %ymm0, %ymm0
 ; AVX512-NEXT:    retq
+;
+; AVX512VL-LABEL: splatconstant_shift_v8i32:
+; AVX512VL:       # BB#0:
+; AVX512VL-NEXT:    vpslld $5, %ymm0, %ymm0
+; AVX512VL-NEXT:    retq
   %shift = shl <8 x i32> %a, <i32 5, i32 5, i32 5, i32 5, i32 5, i32 5, i32 5, i32 5>
   ret <8 x i32> %shift
 }
@@ -807,6 +950,11 @@ define <16 x i16> @splatconstant_shift_v16i16(<16 x i16> %a) nounwind {
 ; AVX512:       # BB#0:
 ; AVX512-NEXT:    vpsllw $3, %ymm0, %ymm0
 ; AVX512-NEXT:    retq
+;
+; AVX512VL-LABEL: splatconstant_shift_v16i16:
+; AVX512VL:       # BB#0:
+; AVX512VL-NEXT:    vpsllw $3, %ymm0, %ymm0
+; AVX512VL-NEXT:    retq
   %shift = shl <16 x i16> %a, <i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3>
   ret <16 x i16> %shift
 }
@@ -849,6 +997,12 @@ define <32 x i8> @splatconstant_shift_v32i8(<32 x i8> %a) nounwind {
 ; AVX512-NEXT:    vpsllw $3, %ymm0, %ymm0
 ; AVX512-NEXT:    vpand {{.*}}(%rip), %ymm0, %ymm0
 ; AVX512-NEXT:    retq
+;
+; AVX512VL-LABEL: splatconstant_shift_v32i8:
+; AVX512VL:       # BB#0:
+; AVX512VL-NEXT:    vpsllw $3, %ymm0, %ymm0
+; AVX512VL-NEXT:    vpand {{.*}}(%rip), %ymm0, %ymm0
+; AVX512VL-NEXT:    retq
   %shift = shl <32 x i8> %a, <i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3>
   ret <32 x i8> %shift
 }