[llvm] a9215ed - [InstCombine][X86] simplifyDemandedVectorEltsIntrinsic - handle avx2 per-element vector shifts
Simon Pilgrim via llvm-commits
llvm-commits at lists.llvm.org
Thu Apr 28 10:16:01 PDT 2022
Author: Simon Pilgrim
Date: 2022-04-28T18:14:54+01:00
New Revision: a9215ed9cc904f05d02dc75f9f4680bf0b33965f
URL: https://github.com/llvm/llvm-project/commit/a9215ed9cc904f05d02dc75f9f4680bf0b33965f
DIFF: https://github.com/llvm/llvm-project/commit/a9215ed9cc904f05d02dc75f9f4680bf0b33965f.diff
LOG: [InstCombine][X86] simplifyDemandedVectorEltsIntrinsic - handle avx2 per-element vector shifts
Added:
Modified:
llvm/lib/Target/X86/X86InstCombineIntrinsic.cpp
llvm/test/Transforms/InstCombine/X86/x86-avx2-inseltpoison.ll
llvm/test/Transforms/InstCombine/X86/x86-avx2.ll
Removed:
################################################################################
diff --git a/llvm/lib/Target/X86/X86InstCombineIntrinsic.cpp b/llvm/lib/Target/X86/X86InstCombineIntrinsic.cpp
index cdbb1d7584b6..4f5fe2bf2c34 100644
--- a/llvm/lib/Target/X86/X86InstCombineIntrinsic.cpp
+++ b/llvm/lib/Target/X86/X86InstCombineIntrinsic.cpp
@@ -1931,6 +1931,23 @@ Optional<Value *> X86TTIImpl::simplifyDemandedVectorEltsIntrinsic(
break;
}
+ // General per-element vector operations.
+ case Intrinsic::x86_avx2_psllv_d:
+ case Intrinsic::x86_avx2_psllv_d_256:
+ case Intrinsic::x86_avx2_psllv_q:
+ case Intrinsic::x86_avx2_psllv_q_256:
+ case Intrinsic::x86_avx2_psrlv_d:
+ case Intrinsic::x86_avx2_psrlv_d_256:
+ case Intrinsic::x86_avx2_psrlv_q:
+ case Intrinsic::x86_avx2_psrlv_q_256:
+ case Intrinsic::x86_avx2_psrav_d:
+ case Intrinsic::x86_avx2_psrav_d_256: {
+ simplifyAndSetOp(&II, 0, DemandedElts, UndefElts);
+ simplifyAndSetOp(&II, 1, DemandedElts, UndefElts2);
+ UndefElts &= UndefElts2;
+ break;
+ }
+
case Intrinsic::x86_sse2_packssdw_128:
case Intrinsic::x86_sse2_packsswb_128:
case Intrinsic::x86_sse2_packuswb_128:
diff --git a/llvm/test/Transforms/InstCombine/X86/x86-avx2-inseltpoison.ll b/llvm/test/Transforms/InstCombine/X86/x86-avx2-inseltpoison.ll
index 880b7270fff6..aa94bb8d3f54 100644
--- a/llvm/test/Transforms/InstCombine/X86/x86-avx2-inseltpoison.ll
+++ b/llvm/test/Transforms/InstCombine/X86/x86-avx2-inseltpoison.ll
@@ -108,10 +108,9 @@ define <8 x float> @elts_test_vpermps(<8 x float> %a0, <8 x i32> %a1) {
define <2 x i64> @elts_test_vpsllvq(<2 x i64> %a0, <2 x i64> %a1) {
; CHECK-LABEL: @elts_test_vpsllvq(
-; CHECK-NEXT: [[TMP1:%.*]] = insertelement <2 x i64> [[A1:%.*]], i64 0, i64 1
-; CHECK-NEXT: [[TMP2:%.*]] = tail call <2 x i64> @llvm.x86.avx2.psllv.q(<2 x i64> [[A0:%.*]], <2 x i64> [[TMP1]])
-; CHECK-NEXT: [[TMP3:%.*]] = shufflevector <2 x i64> [[TMP2]], <2 x i64> poison, <2 x i32> zeroinitializer
-; CHECK-NEXT: ret <2 x i64> [[TMP3]]
+; CHECK-NEXT: [[TMP1:%.*]] = tail call <2 x i64> @llvm.x86.avx2.psllv.q(<2 x i64> [[A0:%.*]], <2 x i64> [[A1:%.*]])
+; CHECK-NEXT: [[TMP2:%.*]] = shufflevector <2 x i64> [[TMP1]], <2 x i64> poison, <2 x i32> zeroinitializer
+; CHECK-NEXT: ret <2 x i64> [[TMP2]]
;
%1 = insertelement <2 x i64> %a1, i64 0, i64 1
%2 = tail call <2 x i64> @llvm.x86.avx2.psllv.q(<2 x i64> %a0, <2 x i64> %1)
@@ -121,10 +120,9 @@ define <2 x i64> @elts_test_vpsllvq(<2 x i64> %a0, <2 x i64> %a1) {
define <2 x i64> @elts_test_vpsrlvq(<2 x i64> %a0, <2 x i64> %a1) {
; CHECK-LABEL: @elts_test_vpsrlvq(
-; CHECK-NEXT: [[TMP1:%.*]] = insertelement <2 x i64> [[A1:%.*]], i64 0, i64 1
-; CHECK-NEXT: [[TMP2:%.*]] = tail call <2 x i64> @llvm.x86.avx2.psrlv.q(<2 x i64> [[A0:%.*]], <2 x i64> [[TMP1]])
-; CHECK-NEXT: [[TMP3:%.*]] = shufflevector <2 x i64> [[TMP2]], <2 x i64> poison, <2 x i32> zeroinitializer
-; CHECK-NEXT: ret <2 x i64> [[TMP3]]
+; CHECK-NEXT: [[TMP1:%.*]] = tail call <2 x i64> @llvm.x86.avx2.psrlv.q(<2 x i64> [[A0:%.*]], <2 x i64> [[A1:%.*]])
+; CHECK-NEXT: [[TMP2:%.*]] = shufflevector <2 x i64> [[TMP1]], <2 x i64> poison, <2 x i32> zeroinitializer
+; CHECK-NEXT: ret <2 x i64> [[TMP2]]
;
%1 = insertelement <2 x i64> %a1, i64 0, i64 1
%2 = tail call <2 x i64> @llvm.x86.avx2.psrlv.q(<2 x i64> %a0, <2 x i64> %1)
@@ -134,10 +132,9 @@ define <2 x i64> @elts_test_vpsrlvq(<2 x i64> %a0, <2 x i64> %a1) {
define <4 x i64> @elts_test_vpsllvq_256(<4 x i64> %a0, <4 x i64> %a1) {
; CHECK-LABEL: @elts_test_vpsllvq_256(
-; CHECK-NEXT: [[TMP1:%.*]] = insertelement <4 x i64> [[A1:%.*]], i64 0, i64 2
-; CHECK-NEXT: [[TMP2:%.*]] = tail call <4 x i64> @llvm.x86.avx2.psllv.q.256(<4 x i64> [[A0:%.*]], <4 x i64> [[TMP1]])
-; CHECK-NEXT: [[TMP3:%.*]] = shufflevector <4 x i64> [[TMP2]], <4 x i64> poison, <4 x i32> zeroinitializer
-; CHECK-NEXT: ret <4 x i64> [[TMP3]]
+; CHECK-NEXT: [[TMP1:%.*]] = tail call <4 x i64> @llvm.x86.avx2.psllv.q.256(<4 x i64> [[A0:%.*]], <4 x i64> [[A1:%.*]])
+; CHECK-NEXT: [[TMP2:%.*]] = shufflevector <4 x i64> [[TMP1]], <4 x i64> poison, <4 x i32> zeroinitializer
+; CHECK-NEXT: ret <4 x i64> [[TMP2]]
;
%1 = insertelement <4 x i64> %a1, i64 0, i64 2
%2 = tail call <4 x i64> @llvm.x86.avx2.psllv.q.256(<4 x i64> %a0, <4 x i64> %1)
@@ -147,10 +144,9 @@ define <4 x i64> @elts_test_vpsllvq_256(<4 x i64> %a0, <4 x i64> %a1) {
define <4 x i64> @elts_test_vpsrlvq_256(<4 x i64> %a0, <4 x i64> %a1) {
; CHECK-LABEL: @elts_test_vpsrlvq_256(
-; CHECK-NEXT: [[TMP1:%.*]] = insertelement <4 x i64> [[A1:%.*]], i64 0, i64 3
-; CHECK-NEXT: [[TMP2:%.*]] = tail call <4 x i64> @llvm.x86.avx2.psrlv.q.256(<4 x i64> [[A0:%.*]], <4 x i64> [[TMP1]])
-; CHECK-NEXT: [[TMP3:%.*]] = shufflevector <4 x i64> [[TMP2]], <4 x i64> poison, <4 x i32> zeroinitializer
-; CHECK-NEXT: ret <4 x i64> [[TMP3]]
+; CHECK-NEXT: [[TMP1:%.*]] = tail call <4 x i64> @llvm.x86.avx2.psrlv.q.256(<4 x i64> [[A0:%.*]], <4 x i64> [[A1:%.*]])
+; CHECK-NEXT: [[TMP2:%.*]] = shufflevector <4 x i64> [[TMP1]], <4 x i64> poison, <4 x i32> zeroinitializer
+; CHECK-NEXT: ret <4 x i64> [[TMP2]]
;
%1 = insertelement <4 x i64> %a1, i64 0, i64 3
%2 = tail call <4 x i64> @llvm.x86.avx2.psrlv.q.256(<4 x i64> %a0, <4 x i64> %1)
@@ -160,10 +156,9 @@ define <4 x i64> @elts_test_vpsrlvq_256(<4 x i64> %a0, <4 x i64> %a1) {
define <4 x i32> @elts_test_vpsllvd(<4 x i32> %a0, <4 x i32> %a1) {
; CHECK-LABEL: @elts_test_vpsllvd(
-; CHECK-NEXT: [[TMP1:%.*]] = insertelement <4 x i32> [[A1:%.*]], i32 0, i64 3
-; CHECK-NEXT: [[TMP2:%.*]] = tail call <4 x i32> @llvm.x86.avx2.psllv.d(<4 x i32> [[A0:%.*]], <4 x i32> [[TMP1]])
-; CHECK-NEXT: [[TMP3:%.*]] = shufflevector <4 x i32> [[TMP2]], <4 x i32> poison, <4 x i32> zeroinitializer
-; CHECK-NEXT: ret <4 x i32> [[TMP3]]
+; CHECK-NEXT: [[TMP1:%.*]] = tail call <4 x i32> @llvm.x86.avx2.psllv.d(<4 x i32> [[A0:%.*]], <4 x i32> [[A1:%.*]])
+; CHECK-NEXT: [[TMP2:%.*]] = shufflevector <4 x i32> [[TMP1]], <4 x i32> poison, <4 x i32> zeroinitializer
+; CHECK-NEXT: ret <4 x i32> [[TMP2]]
;
%1 = insertelement <4 x i32> %a1, i32 0, i64 3
%2 = tail call <4 x i32> @llvm.x86.avx2.psllv.d(<4 x i32> %a0, <4 x i32> %1)
@@ -173,10 +168,9 @@ define <4 x i32> @elts_test_vpsllvd(<4 x i32> %a0, <4 x i32> %a1) {
define <4 x i32> @elts_test_vpsravd(<4 x i32> %a0, <4 x i32> %a1) {
; CHECK-LABEL: @elts_test_vpsravd(
-; CHECK-NEXT: [[TMP1:%.*]] = insertelement <4 x i32> [[A1:%.*]], i32 0, i64 1
-; CHECK-NEXT: [[TMP2:%.*]] = tail call <4 x i32> @llvm.x86.avx2.psrav.d(<4 x i32> [[A0:%.*]], <4 x i32> [[TMP1]])
-; CHECK-NEXT: [[TMP3:%.*]] = shufflevector <4 x i32> [[TMP2]], <4 x i32> poison, <4 x i32> zeroinitializer
-; CHECK-NEXT: ret <4 x i32> [[TMP3]]
+; CHECK-NEXT: [[TMP1:%.*]] = tail call <4 x i32> @llvm.x86.avx2.psrav.d(<4 x i32> [[A0:%.*]], <4 x i32> [[A1:%.*]])
+; CHECK-NEXT: [[TMP2:%.*]] = shufflevector <4 x i32> [[TMP1]], <4 x i32> poison, <4 x i32> zeroinitializer
+; CHECK-NEXT: ret <4 x i32> [[TMP2]]
;
%1 = insertelement <4 x i32> %a1, i32 0, i64 1
%2 = tail call <4 x i32> @llvm.x86.avx2.psrav.d(<4 x i32> %a0, <4 x i32> %1)
@@ -186,10 +180,9 @@ define <4 x i32> @elts_test_vpsravd(<4 x i32> %a0, <4 x i32> %a1) {
define <4 x i32> @elts_test_vpsrlvd(<4 x i32> %a0, <4 x i32> %a1) {
; CHECK-LABEL: @elts_test_vpsrlvd(
-; CHECK-NEXT: [[TMP1:%.*]] = insertelement <4 x i32> [[A1:%.*]], i32 0, i64 2
-; CHECK-NEXT: [[TMP2:%.*]] = tail call <4 x i32> @llvm.x86.avx2.psrlv.d(<4 x i32> [[A0:%.*]], <4 x i32> [[TMP1]])
-; CHECK-NEXT: [[TMP3:%.*]] = shufflevector <4 x i32> [[TMP2]], <4 x i32> poison, <4 x i32> zeroinitializer
-; CHECK-NEXT: ret <4 x i32> [[TMP3]]
+; CHECK-NEXT: [[TMP1:%.*]] = tail call <4 x i32> @llvm.x86.avx2.psrlv.d(<4 x i32> [[A0:%.*]], <4 x i32> [[A1:%.*]])
+; CHECK-NEXT: [[TMP2:%.*]] = shufflevector <4 x i32> [[TMP1]], <4 x i32> poison, <4 x i32> zeroinitializer
+; CHECK-NEXT: ret <4 x i32> [[TMP2]]
;
%1 = insertelement <4 x i32> %a1, i32 0, i64 2
%2 = tail call <4 x i32> @llvm.x86.avx2.psrlv.d(<4 x i32> %a0, <4 x i32> %1)
@@ -199,10 +192,9 @@ define <4 x i32> @elts_test_vpsrlvd(<4 x i32> %a0, <4 x i32> %a1) {
define <8 x i32> @elts_test_vpsllvd_256(<8 x i32> %a0, <8 x i32> %a1) {
; CHECK-LABEL: @elts_test_vpsllvd_256(
-; CHECK-NEXT: [[TMP1:%.*]] = insertelement <8 x i32> [[A1:%.*]], i32 0, i64 3
-; CHECK-NEXT: [[TMP2:%.*]] = tail call <8 x i32> @llvm.x86.avx2.psllv.d.256(<8 x i32> [[A0:%.*]], <8 x i32> [[TMP1]])
-; CHECK-NEXT: [[TMP3:%.*]] = shufflevector <8 x i32> [[TMP2]], <8 x i32> poison, <8 x i32> zeroinitializer
-; CHECK-NEXT: ret <8 x i32> [[TMP3]]
+; CHECK-NEXT: [[TMP1:%.*]] = tail call <8 x i32> @llvm.x86.avx2.psllv.d.256(<8 x i32> [[A0:%.*]], <8 x i32> [[A1:%.*]])
+; CHECK-NEXT: [[TMP2:%.*]] = shufflevector <8 x i32> [[TMP1]], <8 x i32> poison, <8 x i32> zeroinitializer
+; CHECK-NEXT: ret <8 x i32> [[TMP2]]
;
%1 = insertelement <8 x i32> %a1, i32 0, i64 3
%2 = tail call <8 x i32> @llvm.x86.avx2.psllv.d.256(<8 x i32> %a0, <8 x i32> %1)
@@ -212,10 +204,9 @@ define <8 x i32> @elts_test_vpsllvd_256(<8 x i32> %a0, <8 x i32> %a1) {
define <8 x i32> @elts_test_vpsravd_256(<8 x i32> %a0, <8 x i32> %a1) {
; CHECK-LABEL: @elts_test_vpsravd_256(
-; CHECK-NEXT: [[TMP1:%.*]] = insertelement <8 x i32> [[A1:%.*]], i32 0, i64 4
-; CHECK-NEXT: [[TMP2:%.*]] = tail call <8 x i32> @llvm.x86.avx2.psrav.d.256(<8 x i32> [[A0:%.*]], <8 x i32> [[TMP1]])
-; CHECK-NEXT: [[TMP3:%.*]] = shufflevector <8 x i32> [[TMP2]], <8 x i32> poison, <8 x i32> zeroinitializer
-; CHECK-NEXT: ret <8 x i32> [[TMP3]]
+; CHECK-NEXT: [[TMP1:%.*]] = tail call <8 x i32> @llvm.x86.avx2.psrav.d.256(<8 x i32> [[A0:%.*]], <8 x i32> [[A1:%.*]])
+; CHECK-NEXT: [[TMP2:%.*]] = shufflevector <8 x i32> [[TMP1]], <8 x i32> poison, <8 x i32> zeroinitializer
+; CHECK-NEXT: ret <8 x i32> [[TMP2]]
;
%1 = insertelement <8 x i32> %a1, i32 0, i64 4
%2 = tail call <8 x i32> @llvm.x86.avx2.psrav.d.256(<8 x i32> %a0, <8 x i32> %1)
@@ -225,10 +216,9 @@ define <8 x i32> @elts_test_vpsravd_256(<8 x i32> %a0, <8 x i32> %a1) {
define <8 x i32> @elts_test_vpsrlvd_256(<8 x i32> %a0, <8 x i32> %a1) {
; CHECK-LABEL: @elts_test_vpsrlvd_256(
-; CHECK-NEXT: [[TMP1:%.*]] = insertelement <8 x i32> [[A1:%.*]], i32 0, i64 5
-; CHECK-NEXT: [[TMP2:%.*]] = tail call <8 x i32> @llvm.x86.avx2.psrlv.d.256(<8 x i32> [[A0:%.*]], <8 x i32> [[TMP1]])
-; CHECK-NEXT: [[TMP3:%.*]] = shufflevector <8 x i32> [[TMP2]], <8 x i32> poison, <8 x i32> zeroinitializer
-; CHECK-NEXT: ret <8 x i32> [[TMP3]]
+; CHECK-NEXT: [[TMP1:%.*]] = tail call <8 x i32> @llvm.x86.avx2.psrlv.d.256(<8 x i32> [[A0:%.*]], <8 x i32> [[A1:%.*]])
+; CHECK-NEXT: [[TMP2:%.*]] = shufflevector <8 x i32> [[TMP1]], <8 x i32> poison, <8 x i32> zeroinitializer
+; CHECK-NEXT: ret <8 x i32> [[TMP2]]
;
%1 = insertelement <8 x i32> %a1, i32 0, i64 5
%2 = tail call <8 x i32> @llvm.x86.avx2.psrlv.d.256(<8 x i32> %a0, <8 x i32> %1)
diff --git a/llvm/test/Transforms/InstCombine/X86/x86-avx2.ll b/llvm/test/Transforms/InstCombine/X86/x86-avx2.ll
index f0e0e6d2f580..0d1d3150b509 100644
--- a/llvm/test/Transforms/InstCombine/X86/x86-avx2.ll
+++ b/llvm/test/Transforms/InstCombine/X86/x86-avx2.ll
@@ -108,10 +108,9 @@ define <8 x float> @elts_test_vpermps(<8 x float> %a0, <8 x i32> %a1) {
define <2 x i64> @elts_test_vpsllvq(<2 x i64> %a0, <2 x i64> %a1) {
; CHECK-LABEL: @elts_test_vpsllvq(
-; CHECK-NEXT: [[TMP1:%.*]] = insertelement <2 x i64> [[A1:%.*]], i64 0, i64 1
-; CHECK-NEXT: [[TMP2:%.*]] = tail call <2 x i64> @llvm.x86.avx2.psllv.q(<2 x i64> [[A0:%.*]], <2 x i64> [[TMP1]])
-; CHECK-NEXT: [[TMP3:%.*]] = shufflevector <2 x i64> [[TMP2]], <2 x i64> undef, <2 x i32> zeroinitializer
-; CHECK-NEXT: ret <2 x i64> [[TMP3]]
+; CHECK-NEXT: [[TMP1:%.*]] = tail call <2 x i64> @llvm.x86.avx2.psllv.q(<2 x i64> [[A0:%.*]], <2 x i64> [[A1:%.*]])
+; CHECK-NEXT: [[TMP2:%.*]] = shufflevector <2 x i64> [[TMP1]], <2 x i64> undef, <2 x i32> zeroinitializer
+; CHECK-NEXT: ret <2 x i64> [[TMP2]]
;
%1 = insertelement <2 x i64> %a1, i64 0, i64 1
%2 = tail call <2 x i64> @llvm.x86.avx2.psllv.q(<2 x i64> %a0, <2 x i64> %1)
@@ -121,10 +120,9 @@ define <2 x i64> @elts_test_vpsllvq(<2 x i64> %a0, <2 x i64> %a1) {
define <2 x i64> @elts_test_vpsrlvq(<2 x i64> %a0, <2 x i64> %a1) {
; CHECK-LABEL: @elts_test_vpsrlvq(
-; CHECK-NEXT: [[TMP1:%.*]] = insertelement <2 x i64> [[A1:%.*]], i64 0, i64 1
-; CHECK-NEXT: [[TMP2:%.*]] = tail call <2 x i64> @llvm.x86.avx2.psrlv.q(<2 x i64> [[A0:%.*]], <2 x i64> [[TMP1]])
-; CHECK-NEXT: [[TMP3:%.*]] = shufflevector <2 x i64> [[TMP2]], <2 x i64> undef, <2 x i32> zeroinitializer
-; CHECK-NEXT: ret <2 x i64> [[TMP3]]
+; CHECK-NEXT: [[TMP1:%.*]] = tail call <2 x i64> @llvm.x86.avx2.psrlv.q(<2 x i64> [[A0:%.*]], <2 x i64> [[A1:%.*]])
+; CHECK-NEXT: [[TMP2:%.*]] = shufflevector <2 x i64> [[TMP1]], <2 x i64> undef, <2 x i32> zeroinitializer
+; CHECK-NEXT: ret <2 x i64> [[TMP2]]
;
%1 = insertelement <2 x i64> %a1, i64 0, i64 1
%2 = tail call <2 x i64> @llvm.x86.avx2.psrlv.q(<2 x i64> %a0, <2 x i64> %1)
@@ -134,10 +132,9 @@ define <2 x i64> @elts_test_vpsrlvq(<2 x i64> %a0, <2 x i64> %a1) {
define <4 x i64> @elts_test_vpsllvq_256(<4 x i64> %a0, <4 x i64> %a1) {
; CHECK-LABEL: @elts_test_vpsllvq_256(
-; CHECK-NEXT: [[TMP1:%.*]] = insertelement <4 x i64> [[A1:%.*]], i64 0, i64 2
-; CHECK-NEXT: [[TMP2:%.*]] = tail call <4 x i64> @llvm.x86.avx2.psllv.q.256(<4 x i64> [[A0:%.*]], <4 x i64> [[TMP1]])
-; CHECK-NEXT: [[TMP3:%.*]] = shufflevector <4 x i64> [[TMP2]], <4 x i64> undef, <4 x i32> zeroinitializer
-; CHECK-NEXT: ret <4 x i64> [[TMP3]]
+; CHECK-NEXT: [[TMP1:%.*]] = tail call <4 x i64> @llvm.x86.avx2.psllv.q.256(<4 x i64> [[A0:%.*]], <4 x i64> [[A1:%.*]])
+; CHECK-NEXT: [[TMP2:%.*]] = shufflevector <4 x i64> [[TMP1]], <4 x i64> undef, <4 x i32> zeroinitializer
+; CHECK-NEXT: ret <4 x i64> [[TMP2]]
;
%1 = insertelement <4 x i64> %a1, i64 0, i64 2
%2 = tail call <4 x i64> @llvm.x86.avx2.psllv.q.256(<4 x i64> %a0, <4 x i64> %1)
@@ -147,10 +144,9 @@ define <4 x i64> @elts_test_vpsllvq_256(<4 x i64> %a0, <4 x i64> %a1) {
define <4 x i64> @elts_test_vpsrlvq_256(<4 x i64> %a0, <4 x i64> %a1) {
; CHECK-LABEL: @elts_test_vpsrlvq_256(
-; CHECK-NEXT: [[TMP1:%.*]] = insertelement <4 x i64> [[A1:%.*]], i64 0, i64 3
-; CHECK-NEXT: [[TMP2:%.*]] = tail call <4 x i64> @llvm.x86.avx2.psrlv.q.256(<4 x i64> [[A0:%.*]], <4 x i64> [[TMP1]])
-; CHECK-NEXT: [[TMP3:%.*]] = shufflevector <4 x i64> [[TMP2]], <4 x i64> undef, <4 x i32> zeroinitializer
-; CHECK-NEXT: ret <4 x i64> [[TMP3]]
+; CHECK-NEXT: [[TMP1:%.*]] = tail call <4 x i64> @llvm.x86.avx2.psrlv.q.256(<4 x i64> [[A0:%.*]], <4 x i64> [[A1:%.*]])
+; CHECK-NEXT: [[TMP2:%.*]] = shufflevector <4 x i64> [[TMP1]], <4 x i64> undef, <4 x i32> zeroinitializer
+; CHECK-NEXT: ret <4 x i64> [[TMP2]]
;
%1 = insertelement <4 x i64> %a1, i64 0, i64 3
%2 = tail call <4 x i64> @llvm.x86.avx2.psrlv.q.256(<4 x i64> %a0, <4 x i64> %1)
@@ -160,10 +156,9 @@ define <4 x i64> @elts_test_vpsrlvq_256(<4 x i64> %a0, <4 x i64> %a1) {
define <4 x i32> @elts_test_vpsllvd(<4 x i32> %a0, <4 x i32> %a1) {
; CHECK-LABEL: @elts_test_vpsllvd(
-; CHECK-NEXT: [[TMP1:%.*]] = insertelement <4 x i32> [[A1:%.*]], i32 0, i64 3
-; CHECK-NEXT: [[TMP2:%.*]] = tail call <4 x i32> @llvm.x86.avx2.psllv.d(<4 x i32> [[A0:%.*]], <4 x i32> [[TMP1]])
-; CHECK-NEXT: [[TMP3:%.*]] = shufflevector <4 x i32> [[TMP2]], <4 x i32> undef, <4 x i32> zeroinitializer
-; CHECK-NEXT: ret <4 x i32> [[TMP3]]
+; CHECK-NEXT: [[TMP1:%.*]] = tail call <4 x i32> @llvm.x86.avx2.psllv.d(<4 x i32> [[A0:%.*]], <4 x i32> [[A1:%.*]])
+; CHECK-NEXT: [[TMP2:%.*]] = shufflevector <4 x i32> [[TMP1]], <4 x i32> undef, <4 x i32> zeroinitializer
+; CHECK-NEXT: ret <4 x i32> [[TMP2]]
;
%1 = insertelement <4 x i32> %a1, i32 0, i64 3
%2 = tail call <4 x i32> @llvm.x86.avx2.psllv.d(<4 x i32> %a0, <4 x i32> %1)
@@ -173,10 +168,9 @@ define <4 x i32> @elts_test_vpsllvd(<4 x i32> %a0, <4 x i32> %a1) {
define <4 x i32> @elts_test_vpsravd(<4 x i32> %a0, <4 x i32> %a1) {
; CHECK-LABEL: @elts_test_vpsravd(
-; CHECK-NEXT: [[TMP1:%.*]] = insertelement <4 x i32> [[A1:%.*]], i32 0, i64 1
-; CHECK-NEXT: [[TMP2:%.*]] = tail call <4 x i32> @llvm.x86.avx2.psrav.d(<4 x i32> [[A0:%.*]], <4 x i32> [[TMP1]])
-; CHECK-NEXT: [[TMP3:%.*]] = shufflevector <4 x i32> [[TMP2]], <4 x i32> undef, <4 x i32> zeroinitializer
-; CHECK-NEXT: ret <4 x i32> [[TMP3]]
+; CHECK-NEXT: [[TMP1:%.*]] = tail call <4 x i32> @llvm.x86.avx2.psrav.d(<4 x i32> [[A0:%.*]], <4 x i32> [[A1:%.*]])
+; CHECK-NEXT: [[TMP2:%.*]] = shufflevector <4 x i32> [[TMP1]], <4 x i32> undef, <4 x i32> zeroinitializer
+; CHECK-NEXT: ret <4 x i32> [[TMP2]]
;
%1 = insertelement <4 x i32> %a1, i32 0, i64 1
%2 = tail call <4 x i32> @llvm.x86.avx2.psrav.d(<4 x i32> %a0, <4 x i32> %1)
@@ -186,10 +180,9 @@ define <4 x i32> @elts_test_vpsravd(<4 x i32> %a0, <4 x i32> %a1) {
define <4 x i32> @elts_test_vpsrlvd(<4 x i32> %a0, <4 x i32> %a1) {
; CHECK-LABEL: @elts_test_vpsrlvd(
-; CHECK-NEXT: [[TMP1:%.*]] = insertelement <4 x i32> [[A1:%.*]], i32 0, i64 2
-; CHECK-NEXT: [[TMP2:%.*]] = tail call <4 x i32> @llvm.x86.avx2.psrlv.d(<4 x i32> [[A0:%.*]], <4 x i32> [[TMP1]])
-; CHECK-NEXT: [[TMP3:%.*]] = shufflevector <4 x i32> [[TMP2]], <4 x i32> undef, <4 x i32> zeroinitializer
-; CHECK-NEXT: ret <4 x i32> [[TMP3]]
+; CHECK-NEXT: [[TMP1:%.*]] = tail call <4 x i32> @llvm.x86.avx2.psrlv.d(<4 x i32> [[A0:%.*]], <4 x i32> [[A1:%.*]])
+; CHECK-NEXT: [[TMP2:%.*]] = shufflevector <4 x i32> [[TMP1]], <4 x i32> undef, <4 x i32> zeroinitializer
+; CHECK-NEXT: ret <4 x i32> [[TMP2]]
;
%1 = insertelement <4 x i32> %a1, i32 0, i64 2
%2 = tail call <4 x i32> @llvm.x86.avx2.psrlv.d(<4 x i32> %a0, <4 x i32> %1)
@@ -199,10 +192,9 @@ define <4 x i32> @elts_test_vpsrlvd(<4 x i32> %a0, <4 x i32> %a1) {
define <8 x i32> @elts_test_vpsllvd_256(<8 x i32> %a0, <8 x i32> %a1) {
; CHECK-LABEL: @elts_test_vpsllvd_256(
-; CHECK-NEXT: [[TMP1:%.*]] = insertelement <8 x i32> [[A1:%.*]], i32 0, i64 3
-; CHECK-NEXT: [[TMP2:%.*]] = tail call <8 x i32> @llvm.x86.avx2.psllv.d.256(<8 x i32> [[A0:%.*]], <8 x i32> [[TMP1]])
-; CHECK-NEXT: [[TMP3:%.*]] = shufflevector <8 x i32> [[TMP2]], <8 x i32> undef, <8 x i32> zeroinitializer
-; CHECK-NEXT: ret <8 x i32> [[TMP3]]
+; CHECK-NEXT: [[TMP1:%.*]] = tail call <8 x i32> @llvm.x86.avx2.psllv.d.256(<8 x i32> [[A0:%.*]], <8 x i32> [[A1:%.*]])
+; CHECK-NEXT: [[TMP2:%.*]] = shufflevector <8 x i32> [[TMP1]], <8 x i32> undef, <8 x i32> zeroinitializer
+; CHECK-NEXT: ret <8 x i32> [[TMP2]]
;
%1 = insertelement <8 x i32> %a1, i32 0, i64 3
%2 = tail call <8 x i32> @llvm.x86.avx2.psllv.d.256(<8 x i32> %a0, <8 x i32> %1)
@@ -212,10 +204,9 @@ define <8 x i32> @elts_test_vpsllvd_256(<8 x i32> %a0, <8 x i32> %a1) {
define <8 x i32> @elts_test_vpsravd_256(<8 x i32> %a0, <8 x i32> %a1) {
; CHECK-LABEL: @elts_test_vpsravd_256(
-; CHECK-NEXT: [[TMP1:%.*]] = insertelement <8 x i32> [[A1:%.*]], i32 0, i64 4
-; CHECK-NEXT: [[TMP2:%.*]] = tail call <8 x i32> @llvm.x86.avx2.psrav.d.256(<8 x i32> [[A0:%.*]], <8 x i32> [[TMP1]])
-; CHECK-NEXT: [[TMP3:%.*]] = shufflevector <8 x i32> [[TMP2]], <8 x i32> undef, <8 x i32> zeroinitializer
-; CHECK-NEXT: ret <8 x i32> [[TMP3]]
+; CHECK-NEXT: [[TMP1:%.*]] = tail call <8 x i32> @llvm.x86.avx2.psrav.d.256(<8 x i32> [[A0:%.*]], <8 x i32> [[A1:%.*]])
+; CHECK-NEXT: [[TMP2:%.*]] = shufflevector <8 x i32> [[TMP1]], <8 x i32> undef, <8 x i32> zeroinitializer
+; CHECK-NEXT: ret <8 x i32> [[TMP2]]
;
%1 = insertelement <8 x i32> %a1, i32 0, i64 4
%2 = tail call <8 x i32> @llvm.x86.avx2.psrav.d.256(<8 x i32> %a0, <8 x i32> %1)
@@ -225,10 +216,9 @@ define <8 x i32> @elts_test_vpsravd_256(<8 x i32> %a0, <8 x i32> %a1) {
define <8 x i32> @elts_test_vpsrlvd_256(<8 x i32> %a0, <8 x i32> %a1) {
; CHECK-LABEL: @elts_test_vpsrlvd_256(
-; CHECK-NEXT: [[TMP1:%.*]] = insertelement <8 x i32> [[A1:%.*]], i32 0, i64 5
-; CHECK-NEXT: [[TMP2:%.*]] = tail call <8 x i32> @llvm.x86.avx2.psrlv.d.256(<8 x i32> [[A0:%.*]], <8 x i32> [[TMP1]])
-; CHECK-NEXT: [[TMP3:%.*]] = shufflevector <8 x i32> [[TMP2]], <8 x i32> undef, <8 x i32> zeroinitializer
-; CHECK-NEXT: ret <8 x i32> [[TMP3]]
+; CHECK-NEXT: [[TMP1:%.*]] = tail call <8 x i32> @llvm.x86.avx2.psrlv.d.256(<8 x i32> [[A0:%.*]], <8 x i32> [[A1:%.*]])
+; CHECK-NEXT: [[TMP2:%.*]] = shufflevector <8 x i32> [[TMP1]], <8 x i32> undef, <8 x i32> zeroinitializer
+; CHECK-NEXT: ret <8 x i32> [[TMP2]]
;
%1 = insertelement <8 x i32> %a1, i32 0, i64 5
%2 = tail call <8 x i32> @llvm.x86.avx2.psrlv.d.256(<8 x i32> %a0, <8 x i32> %1)
More information about the llvm-commits
mailing list