[llvm] fc01b9b - [CostModel][X86] Align v4i64 MUL costs on AVX1 targets with worst case
Simon Pilgrim via llvm-commits
llvm-commits at lists.llvm.org
Sat May 22 12:11:37 PDT 2021
Author: Simon Pilgrim
Date: 2021-05-22T20:07:55+01:00
New Revision: fc01b9bdf8b55f6b09f1dcaedf78dad62ff205c1
URL: https://github.com/llvm/llvm-project/commit/fc01b9bdf8b55f6b09f1dcaedf78dad62ff205c1
DIFF: https://github.com/llvm/llvm-project/commit/fc01b9bdf8b55f6b09f1dcaedf78dad62ff205c1.diff
LOG: [CostModel][X86] Align v4i64 MUL costs on AVX1 targets with worst case
Based on worst case of sandybridge (vs btver2 + bdver2) llvm-mca analysis - which is a lot less than what we were predicting (I think based off total uop count).
Added:
Modified:
llvm/lib/Target/X86/X86TargetTransformInfo.cpp
llvm/test/Analysis/CostModel/X86/arith-fix.ll
llvm/test/Analysis/CostModel/X86/arith-overflow.ll
llvm/test/Analysis/CostModel/X86/arith.ll
llvm/test/Analysis/CostModel/X86/reduce-mul.ll
llvm/test/Analysis/CostModel/X86/rem.ll
llvm/test/Transforms/SLPVectorizer/X86/arith-fix.ll
llvm/test/Transforms/SLPVectorizer/X86/pr46983.ll
Removed:
################################################################################
diff --git a/llvm/lib/Target/X86/X86TargetTransformInfo.cpp b/llvm/lib/Target/X86/X86TargetTransformInfo.cpp
index f77de4e8f754..cbf57ab5c0b5 100644
--- a/llvm/lib/Target/X86/X86TargetTransformInfo.cpp
+++ b/llvm/lib/Target/X86/X86TargetTransformInfo.cpp
@@ -815,6 +815,8 @@ InstructionCost X86TTIImpl::getArithmeticInstrCost(
// Two ops + 1 extract + 1 insert = 4.
{ ISD::MUL, MVT::v16i16, 4 },
{ ISD::MUL, MVT::v8i32, 5 }, // BTVER2 from http://www.agner.org/
+ { ISD::MUL, MVT::v4i64, 12 },
+
{ ISD::SUB, MVT::v32i8, 4 },
{ ISD::ADD, MVT::v32i8, 4 },
{ ISD::SUB, MVT::v16i16, 4 },
@@ -824,13 +826,6 @@ InstructionCost X86TTIImpl::getArithmeticInstrCost(
{ ISD::SUB, MVT::v4i64, 4 },
{ ISD::ADD, MVT::v4i64, 4 },
- // A v4i64 multiply is custom lowered as two split v2i64 vectors that then
- // are lowered as a series of long multiplies(3), shifts(3) and adds(2)
- // Because we believe v4i64 to be a legal type, we must also include the
- // extract+insert in the cost table. Therefore, the cost here is 18
- // instead of 8.
- { ISD::MUL, MVT::v4i64, 18 },
-
{ ISD::FNEG, MVT::v4f64, 2 }, // BTVER2 from http://www.agner.org/
{ ISD::FNEG, MVT::v8f32, 2 }, // BTVER2 from http://www.agner.org/
diff --git a/llvm/test/Analysis/CostModel/X86/arith-fix.ll b/llvm/test/Analysis/CostModel/X86/arith-fix.ll
index d7ca25e368ec..c2aa7aa77c2b 100644
--- a/llvm/test/Analysis/CostModel/X86/arith-fix.ll
+++ b/llvm/test/Analysis/CostModel/X86/arith-fix.ll
@@ -76,9 +76,9 @@ define i32 @smul(i32 %arg) {
; AVX1-NEXT: Cost Model: Found an estimated cost of 45 for instruction: %V4I64 = call <4 x i64> @llvm.smul.fix.v4i64(<4 x i64> undef, <4 x i64> undef, i32 3)
; AVX1-NEXT: Cost Model: Found an estimated cost of 90 for instruction: %V8I64 = call <8 x i64> @llvm.smul.fix.v8i64(<8 x i64> undef, <8 x i64> undef, i32 3)
; AVX1-NEXT: Cost Model: Found an estimated cost of 6 for instruction: %I32 = call i32 @llvm.smul.fix.i32(i32 undef, i32 undef, i32 3)
-; AVX1-NEXT: Cost Model: Found an estimated cost of 33 for instruction: %V4I32 = call <4 x i32> @llvm.smul.fix.v4i32(<4 x i32> undef, <4 x i32> undef, i32 3)
-; AVX1-NEXT: Cost Model: Found an estimated cost of 69 for instruction: %V8I32 = call <8 x i32> @llvm.smul.fix.v8i32(<8 x i32> undef, <8 x i32> undef, i32 3)
-; AVX1-NEXT: Cost Model: Found an estimated cost of 138 for instruction: %V16I32 = call <16 x i32> @llvm.smul.fix.v16i32(<16 x i32> undef, <16 x i32> undef, i32 3)
+; AVX1-NEXT: Cost Model: Found an estimated cost of 27 for instruction: %V4I32 = call <4 x i32> @llvm.smul.fix.v4i32(<4 x i32> undef, <4 x i32> undef, i32 3)
+; AVX1-NEXT: Cost Model: Found an estimated cost of 57 for instruction: %V8I32 = call <8 x i32> @llvm.smul.fix.v8i32(<8 x i32> undef, <8 x i32> undef, i32 3)
+; AVX1-NEXT: Cost Model: Found an estimated cost of 114 for instruction: %V16I32 = call <16 x i32> @llvm.smul.fix.v16i32(<16 x i32> undef, <16 x i32> undef, i32 3)
; AVX1-NEXT: Cost Model: Found an estimated cost of 6 for instruction: %I16 = call i16 @llvm.smul.fix.i16(i16 undef, i16 undef, i32 3)
; AVX1-NEXT: Cost Model: Found an estimated cost of 26 for instruction: %V8I16 = call <8 x i16> @llvm.smul.fix.v8i16(<8 x i16> undef, <8 x i16> undef, i32 3)
; AVX1-NEXT: Cost Model: Found an estimated cost of 39 for instruction: %V16I16 = call <16 x i16> @llvm.smul.fix.v16i16(<16 x i16> undef, <16 x i16> undef, i32 3)
@@ -209,9 +209,9 @@ define i32 @smul(i32 %arg) {
; BTVER2-NEXT: Cost Model: Found an estimated cost of 45 for instruction: %V4I64 = call <4 x i64> @llvm.smul.fix.v4i64(<4 x i64> undef, <4 x i64> undef, i32 3)
; BTVER2-NEXT: Cost Model: Found an estimated cost of 90 for instruction: %V8I64 = call <8 x i64> @llvm.smul.fix.v8i64(<8 x i64> undef, <8 x i64> undef, i32 3)
; BTVER2-NEXT: Cost Model: Found an estimated cost of 6 for instruction: %I32 = call i32 @llvm.smul.fix.i32(i32 undef, i32 undef, i32 3)
-; BTVER2-NEXT: Cost Model: Found an estimated cost of 33 for instruction: %V4I32 = call <4 x i32> @llvm.smul.fix.v4i32(<4 x i32> undef, <4 x i32> undef, i32 3)
-; BTVER2-NEXT: Cost Model: Found an estimated cost of 69 for instruction: %V8I32 = call <8 x i32> @llvm.smul.fix.v8i32(<8 x i32> undef, <8 x i32> undef, i32 3)
-; BTVER2-NEXT: Cost Model: Found an estimated cost of 138 for instruction: %V16I32 = call <16 x i32> @llvm.smul.fix.v16i32(<16 x i32> undef, <16 x i32> undef, i32 3)
+; BTVER2-NEXT: Cost Model: Found an estimated cost of 27 for instruction: %V4I32 = call <4 x i32> @llvm.smul.fix.v4i32(<4 x i32> undef, <4 x i32> undef, i32 3)
+; BTVER2-NEXT: Cost Model: Found an estimated cost of 57 for instruction: %V8I32 = call <8 x i32> @llvm.smul.fix.v8i32(<8 x i32> undef, <8 x i32> undef, i32 3)
+; BTVER2-NEXT: Cost Model: Found an estimated cost of 114 for instruction: %V16I32 = call <16 x i32> @llvm.smul.fix.v16i32(<16 x i32> undef, <16 x i32> undef, i32 3)
; BTVER2-NEXT: Cost Model: Found an estimated cost of 6 for instruction: %I16 = call i16 @llvm.smul.fix.i16(i16 undef, i16 undef, i32 3)
; BTVER2-NEXT: Cost Model: Found an estimated cost of 26 for instruction: %V8I16 = call <8 x i16> @llvm.smul.fix.v8i16(<8 x i16> undef, <8 x i16> undef, i32 3)
; BTVER2-NEXT: Cost Model: Found an estimated cost of 39 for instruction: %V16I16 = call <16 x i16> @llvm.smul.fix.v16i16(<16 x i16> undef, <16 x i16> undef, i32 3)
@@ -310,9 +310,9 @@ define i32 @umul(i32 %arg) {
; AVX1-NEXT: Cost Model: Found an estimated cost of 45 for instruction: %V4I64 = call <4 x i64> @llvm.umul.fix.v4i64(<4 x i64> undef, <4 x i64> undef, i32 3)
; AVX1-NEXT: Cost Model: Found an estimated cost of 90 for instruction: %V8I64 = call <8 x i64> @llvm.umul.fix.v8i64(<8 x i64> undef, <8 x i64> undef, i32 3)
; AVX1-NEXT: Cost Model: Found an estimated cost of 4 for instruction: %I32 = call i32 @llvm.umul.fix.i32(i32 undef, i32 undef, i32 3)
-; AVX1-NEXT: Cost Model: Found an estimated cost of 33 for instruction: %V4I32 = call <4 x i32> @llvm.umul.fix.v4i32(<4 x i32> undef, <4 x i32> undef, i32 3)
-; AVX1-NEXT: Cost Model: Found an estimated cost of 69 for instruction: %V8I32 = call <8 x i32> @llvm.umul.fix.v8i32(<8 x i32> undef, <8 x i32> undef, i32 3)
-; AVX1-NEXT: Cost Model: Found an estimated cost of 138 for instruction: %V16I32 = call <16 x i32> @llvm.umul.fix.v16i32(<16 x i32> undef, <16 x i32> undef, i32 3)
+; AVX1-NEXT: Cost Model: Found an estimated cost of 27 for instruction: %V4I32 = call <4 x i32> @llvm.umul.fix.v4i32(<4 x i32> undef, <4 x i32> undef, i32 3)
+; AVX1-NEXT: Cost Model: Found an estimated cost of 57 for instruction: %V8I32 = call <8 x i32> @llvm.umul.fix.v8i32(<8 x i32> undef, <8 x i32> undef, i32 3)
+; AVX1-NEXT: Cost Model: Found an estimated cost of 114 for instruction: %V16I32 = call <16 x i32> @llvm.umul.fix.v16i32(<16 x i32> undef, <16 x i32> undef, i32 3)
; AVX1-NEXT: Cost Model: Found an estimated cost of 6 for instruction: %I16 = call i16 @llvm.umul.fix.i16(i16 undef, i16 undef, i32 3)
; AVX1-NEXT: Cost Model: Found an estimated cost of 26 for instruction: %V8I16 = call <8 x i16> @llvm.umul.fix.v8i16(<8 x i16> undef, <8 x i16> undef, i32 3)
; AVX1-NEXT: Cost Model: Found an estimated cost of 39 for instruction: %V16I16 = call <16 x i16> @llvm.umul.fix.v16i16(<16 x i16> undef, <16 x i16> undef, i32 3)
@@ -443,9 +443,9 @@ define i32 @umul(i32 %arg) {
; BTVER2-NEXT: Cost Model: Found an estimated cost of 45 for instruction: %V4I64 = call <4 x i64> @llvm.umul.fix.v4i64(<4 x i64> undef, <4 x i64> undef, i32 3)
; BTVER2-NEXT: Cost Model: Found an estimated cost of 90 for instruction: %V8I64 = call <8 x i64> @llvm.umul.fix.v8i64(<8 x i64> undef, <8 x i64> undef, i32 3)
; BTVER2-NEXT: Cost Model: Found an estimated cost of 4 for instruction: %I32 = call i32 @llvm.umul.fix.i32(i32 undef, i32 undef, i32 3)
-; BTVER2-NEXT: Cost Model: Found an estimated cost of 33 for instruction: %V4I32 = call <4 x i32> @llvm.umul.fix.v4i32(<4 x i32> undef, <4 x i32> undef, i32 3)
-; BTVER2-NEXT: Cost Model: Found an estimated cost of 69 for instruction: %V8I32 = call <8 x i32> @llvm.umul.fix.v8i32(<8 x i32> undef, <8 x i32> undef, i32 3)
-; BTVER2-NEXT: Cost Model: Found an estimated cost of 138 for instruction: %V16I32 = call <16 x i32> @llvm.umul.fix.v16i32(<16 x i32> undef, <16 x i32> undef, i32 3)
+; BTVER2-NEXT: Cost Model: Found an estimated cost of 27 for instruction: %V4I32 = call <4 x i32> @llvm.umul.fix.v4i32(<4 x i32> undef, <4 x i32> undef, i32 3)
+; BTVER2-NEXT: Cost Model: Found an estimated cost of 57 for instruction: %V8I32 = call <8 x i32> @llvm.umul.fix.v8i32(<8 x i32> undef, <8 x i32> undef, i32 3)
+; BTVER2-NEXT: Cost Model: Found an estimated cost of 114 for instruction: %V16I32 = call <16 x i32> @llvm.umul.fix.v16i32(<16 x i32> undef, <16 x i32> undef, i32 3)
; BTVER2-NEXT: Cost Model: Found an estimated cost of 6 for instruction: %I16 = call i16 @llvm.umul.fix.i16(i16 undef, i16 undef, i32 3)
; BTVER2-NEXT: Cost Model: Found an estimated cost of 26 for instruction: %V8I16 = call <8 x i16> @llvm.umul.fix.v8i16(<8 x i16> undef, <8 x i16> undef, i32 3)
; BTVER2-NEXT: Cost Model: Found an estimated cost of 39 for instruction: %V16I16 = call <16 x i16> @llvm.umul.fix.v16i16(<16 x i16> undef, <16 x i16> undef, i32 3)
diff --git a/llvm/test/Analysis/CostModel/X86/arith-overflow.ll b/llvm/test/Analysis/CostModel/X86/arith-overflow.ll
index 801bf3f0d284..8326ef053967 100644
--- a/llvm/test/Analysis/CostModel/X86/arith-overflow.ll
+++ b/llvm/test/Analysis/CostModel/X86/arith-overflow.ll
@@ -1032,9 +1032,9 @@ define i32 @smul(i32 %arg) {
; AVX1-NEXT: Cost Model: Found an estimated cost of 54 for instruction: %V4I64 = call { <4 x i64>, <4 x i1> } @llvm.smul.with.overflow.v4i64(<4 x i64> undef, <4 x i64> undef)
; AVX1-NEXT: Cost Model: Found an estimated cost of 108 for instruction: %V8I64 = call { <8 x i64>, <8 x i1> } @llvm.smul.with.overflow.v8i64(<8 x i64> undef, <8 x i64> undef)
; AVX1-NEXT: Cost Model: Found an estimated cost of 2 for instruction: %I32 = call { i32, i1 } @llvm.smul.with.overflow.i32(i32 undef, i32 undef)
-; AVX1-NEXT: Cost Model: Found an estimated cost of 33 for instruction: %V4I32 = call { <4 x i32>, <4 x i1> } @llvm.smul.with.overflow.v4i32(<4 x i32> undef, <4 x i32> undef)
-; AVX1-NEXT: Cost Model: Found an estimated cost of 72 for instruction: %V8I32 = call { <8 x i32>, <8 x i1> } @llvm.smul.with.overflow.v8i32(<8 x i32> undef, <8 x i32> undef)
-; AVX1-NEXT: Cost Model: Found an estimated cost of 144 for instruction: %V16I32 = call { <16 x i32>, <16 x i1> } @llvm.smul.with.overflow.v16i32(<16 x i32> undef, <16 x i32> undef)
+; AVX1-NEXT: Cost Model: Found an estimated cost of 27 for instruction: %V4I32 = call { <4 x i32>, <4 x i1> } @llvm.smul.with.overflow.v4i32(<4 x i32> undef, <4 x i32> undef)
+; AVX1-NEXT: Cost Model: Found an estimated cost of 60 for instruction: %V8I32 = call { <8 x i32>, <8 x i1> } @llvm.smul.with.overflow.v8i32(<8 x i32> undef, <8 x i32> undef)
+; AVX1-NEXT: Cost Model: Found an estimated cost of 120 for instruction: %V16I32 = call { <16 x i32>, <16 x i1> } @llvm.smul.with.overflow.v16i32(<16 x i32> undef, <16 x i32> undef)
; AVX1-NEXT: Cost Model: Found an estimated cost of 2 for instruction: %I16 = call { i16, i1 } @llvm.smul.with.overflow.i16(i16 undef, i16 undef)
; AVX1-NEXT: Cost Model: Found an estimated cost of 26 for instruction: %V8I16 = call { <8 x i16>, <8 x i1> } @llvm.smul.with.overflow.v8i16(<8 x i16> undef, <8 x i16> undef)
; AVX1-NEXT: Cost Model: Found an estimated cost of 42 for instruction: %V16I16 = call { <16 x i16>, <16 x i1> } @llvm.smul.with.overflow.v16i16(<16 x i16> undef, <16 x i16> undef)
@@ -1165,9 +1165,9 @@ define i32 @smul(i32 %arg) {
; BTVER2-NEXT: Cost Model: Found an estimated cost of 54 for instruction: %V4I64 = call { <4 x i64>, <4 x i1> } @llvm.smul.with.overflow.v4i64(<4 x i64> undef, <4 x i64> undef)
; BTVER2-NEXT: Cost Model: Found an estimated cost of 108 for instruction: %V8I64 = call { <8 x i64>, <8 x i1> } @llvm.smul.with.overflow.v8i64(<8 x i64> undef, <8 x i64> undef)
; BTVER2-NEXT: Cost Model: Found an estimated cost of 2 for instruction: %I32 = call { i32, i1 } @llvm.smul.with.overflow.i32(i32 undef, i32 undef)
-; BTVER2-NEXT: Cost Model: Found an estimated cost of 33 for instruction: %V4I32 = call { <4 x i32>, <4 x i1> } @llvm.smul.with.overflow.v4i32(<4 x i32> undef, <4 x i32> undef)
-; BTVER2-NEXT: Cost Model: Found an estimated cost of 72 for instruction: %V8I32 = call { <8 x i32>, <8 x i1> } @llvm.smul.with.overflow.v8i32(<8 x i32> undef, <8 x i32> undef)
-; BTVER2-NEXT: Cost Model: Found an estimated cost of 144 for instruction: %V16I32 = call { <16 x i32>, <16 x i1> } @llvm.smul.with.overflow.v16i32(<16 x i32> undef, <16 x i32> undef)
+; BTVER2-NEXT: Cost Model: Found an estimated cost of 27 for instruction: %V4I32 = call { <4 x i32>, <4 x i1> } @llvm.smul.with.overflow.v4i32(<4 x i32> undef, <4 x i32> undef)
+; BTVER2-NEXT: Cost Model: Found an estimated cost of 60 for instruction: %V8I32 = call { <8 x i32>, <8 x i1> } @llvm.smul.with.overflow.v8i32(<8 x i32> undef, <8 x i32> undef)
+; BTVER2-NEXT: Cost Model: Found an estimated cost of 120 for instruction: %V16I32 = call { <16 x i32>, <16 x i1> } @llvm.smul.with.overflow.v16i32(<16 x i32> undef, <16 x i32> undef)
; BTVER2-NEXT: Cost Model: Found an estimated cost of 2 for instruction: %I16 = call { i16, i1 } @llvm.smul.with.overflow.i16(i16 undef, i16 undef)
; BTVER2-NEXT: Cost Model: Found an estimated cost of 26 for instruction: %V8I16 = call { <8 x i16>, <8 x i1> } @llvm.smul.with.overflow.v8i16(<8 x i16> undef, <8 x i16> undef)
; BTVER2-NEXT: Cost Model: Found an estimated cost of 42 for instruction: %V16I16 = call { <16 x i16>, <16 x i1> } @llvm.smul.with.overflow.v16i16(<16 x i16> undef, <16 x i16> undef)
@@ -1270,9 +1270,9 @@ define i32 @umul(i32 %arg) {
; AVX1-NEXT: Cost Model: Found an estimated cost of 44 for instruction: %V4I64 = call { <4 x i64>, <4 x i1> } @llvm.umul.with.overflow.v4i64(<4 x i64> undef, <4 x i64> undef)
; AVX1-NEXT: Cost Model: Found an estimated cost of 88 for instruction: %V8I64 = call { <8 x i64>, <8 x i1> } @llvm.umul.with.overflow.v8i64(<8 x i64> undef, <8 x i64> undef)
; AVX1-NEXT: Cost Model: Found an estimated cost of 2 for instruction: %I32 = call { i32, i1 } @llvm.umul.with.overflow.i32(i32 undef, i32 undef)
-; AVX1-NEXT: Cost Model: Found an estimated cost of 32 for instruction: %V4I32 = call { <4 x i32>, <4 x i1> } @llvm.umul.with.overflow.v4i32(<4 x i32> undef, <4 x i32> undef)
-; AVX1-NEXT: Cost Model: Found an estimated cost of 68 for instruction: %V8I32 = call { <8 x i32>, <8 x i1> } @llvm.umul.with.overflow.v8i32(<8 x i32> undef, <8 x i32> undef)
-; AVX1-NEXT: Cost Model: Found an estimated cost of 136 for instruction: %V16I32 = call { <16 x i32>, <16 x i1> } @llvm.umul.with.overflow.v16i32(<16 x i32> undef, <16 x i32> undef)
+; AVX1-NEXT: Cost Model: Found an estimated cost of 26 for instruction: %V4I32 = call { <4 x i32>, <4 x i1> } @llvm.umul.with.overflow.v4i32(<4 x i32> undef, <4 x i32> undef)
+; AVX1-NEXT: Cost Model: Found an estimated cost of 56 for instruction: %V8I32 = call { <8 x i32>, <8 x i1> } @llvm.umul.with.overflow.v8i32(<8 x i32> undef, <8 x i32> undef)
+; AVX1-NEXT: Cost Model: Found an estimated cost of 112 for instruction: %V16I32 = call { <16 x i32>, <16 x i1> } @llvm.umul.with.overflow.v16i32(<16 x i32> undef, <16 x i32> undef)
; AVX1-NEXT: Cost Model: Found an estimated cost of 2 for instruction: %I16 = call { i16, i1 } @llvm.umul.with.overflow.i16(i16 undef, i16 undef)
; AVX1-NEXT: Cost Model: Found an estimated cost of 25 for instruction: %V8I16 = call { <8 x i16>, <8 x i1> } @llvm.umul.with.overflow.v8i16(<8 x i16> undef, <8 x i16> undef)
; AVX1-NEXT: Cost Model: Found an estimated cost of 38 for instruction: %V16I16 = call { <16 x i16>, <16 x i1> } @llvm.umul.with.overflow.v16i16(<16 x i16> undef, <16 x i16> undef)
@@ -1403,9 +1403,9 @@ define i32 @umul(i32 %arg) {
; BTVER2-NEXT: Cost Model: Found an estimated cost of 44 for instruction: %V4I64 = call { <4 x i64>, <4 x i1> } @llvm.umul.with.overflow.v4i64(<4 x i64> undef, <4 x i64> undef)
; BTVER2-NEXT: Cost Model: Found an estimated cost of 88 for instruction: %V8I64 = call { <8 x i64>, <8 x i1> } @llvm.umul.with.overflow.v8i64(<8 x i64> undef, <8 x i64> undef)
; BTVER2-NEXT: Cost Model: Found an estimated cost of 2 for instruction: %I32 = call { i32, i1 } @llvm.umul.with.overflow.i32(i32 undef, i32 undef)
-; BTVER2-NEXT: Cost Model: Found an estimated cost of 32 for instruction: %V4I32 = call { <4 x i32>, <4 x i1> } @llvm.umul.with.overflow.v4i32(<4 x i32> undef, <4 x i32> undef)
-; BTVER2-NEXT: Cost Model: Found an estimated cost of 68 for instruction: %V8I32 = call { <8 x i32>, <8 x i1> } @llvm.umul.with.overflow.v8i32(<8 x i32> undef, <8 x i32> undef)
-; BTVER2-NEXT: Cost Model: Found an estimated cost of 136 for instruction: %V16I32 = call { <16 x i32>, <16 x i1> } @llvm.umul.with.overflow.v16i32(<16 x i32> undef, <16 x i32> undef)
+; BTVER2-NEXT: Cost Model: Found an estimated cost of 26 for instruction: %V4I32 = call { <4 x i32>, <4 x i1> } @llvm.umul.with.overflow.v4i32(<4 x i32> undef, <4 x i32> undef)
+; BTVER2-NEXT: Cost Model: Found an estimated cost of 56 for instruction: %V8I32 = call { <8 x i32>, <8 x i1> } @llvm.umul.with.overflow.v8i32(<8 x i32> undef, <8 x i32> undef)
+; BTVER2-NEXT: Cost Model: Found an estimated cost of 112 for instruction: %V16I32 = call { <16 x i32>, <16 x i1> } @llvm.umul.with.overflow.v16i32(<16 x i32> undef, <16 x i32> undef)
; BTVER2-NEXT: Cost Model: Found an estimated cost of 2 for instruction: %I16 = call { i16, i1 } @llvm.umul.with.overflow.i16(i16 undef, i16 undef)
; BTVER2-NEXT: Cost Model: Found an estimated cost of 25 for instruction: %V8I16 = call { <8 x i16>, <8 x i1> } @llvm.umul.with.overflow.v8i16(<8 x i16> undef, <8 x i16> undef)
; BTVER2-NEXT: Cost Model: Found an estimated cost of 38 for instruction: %V16I16 = call { <16 x i16>, <16 x i1> } @llvm.umul.with.overflow.v16i16(<16 x i16> undef, <16 x i16> undef)
diff --git a/llvm/test/Analysis/CostModel/X86/arith.ll b/llvm/test/Analysis/CostModel/X86/arith.ll
index 3e92c36541e4..a7eb9926abfa 100644
--- a/llvm/test/Analysis/CostModel/X86/arith.ll
+++ b/llvm/test/Analysis/CostModel/X86/arith.ll
@@ -976,8 +976,8 @@ define i32 @mul(i32 %arg) {
; AVX1-LABEL: 'mul'
; AVX1-NEXT: Cost Model: Found an estimated cost of 1 for instruction: %I64 = mul i64 undef, undef
; AVX1-NEXT: Cost Model: Found an estimated cost of 8 for instruction: %V2I64 = mul <2 x i64> undef, undef
-; AVX1-NEXT: Cost Model: Found an estimated cost of 18 for instruction: %V4I64 = mul <4 x i64> undef, undef
-; AVX1-NEXT: Cost Model: Found an estimated cost of 36 for instruction: %V8I64 = mul <8 x i64> undef, undef
+; AVX1-NEXT: Cost Model: Found an estimated cost of 12 for instruction: %V4I64 = mul <4 x i64> undef, undef
+; AVX1-NEXT: Cost Model: Found an estimated cost of 24 for instruction: %V8I64 = mul <8 x i64> undef, undef
; AVX1-NEXT: Cost Model: Found an estimated cost of 1 for instruction: %I32 = mul i32 undef, undef
; AVX1-NEXT: Cost Model: Found an estimated cost of 2 for instruction: %V4I32 = mul <4 x i32> undef, undef
; AVX1-NEXT: Cost Model: Found an estimated cost of 5 for instruction: %V8I32 = mul <8 x i32> undef, undef
@@ -1130,8 +1130,8 @@ define i32 @mul(i32 %arg) {
; BTVER2-LABEL: 'mul'
; BTVER2-NEXT: Cost Model: Found an estimated cost of 1 for instruction: %I64 = mul i64 undef, undef
; BTVER2-NEXT: Cost Model: Found an estimated cost of 8 for instruction: %V2I64 = mul <2 x i64> undef, undef
-; BTVER2-NEXT: Cost Model: Found an estimated cost of 18 for instruction: %V4I64 = mul <4 x i64> undef, undef
-; BTVER2-NEXT: Cost Model: Found an estimated cost of 36 for instruction: %V8I64 = mul <8 x i64> undef, undef
+; BTVER2-NEXT: Cost Model: Found an estimated cost of 12 for instruction: %V4I64 = mul <4 x i64> undef, undef
+; BTVER2-NEXT: Cost Model: Found an estimated cost of 24 for instruction: %V8I64 = mul <8 x i64> undef, undef
; BTVER2-NEXT: Cost Model: Found an estimated cost of 1 for instruction: %I32 = mul i32 undef, undef
; BTVER2-NEXT: Cost Model: Found an estimated cost of 2 for instruction: %V4I32 = mul <4 x i32> undef, undef
; BTVER2-NEXT: Cost Model: Found an estimated cost of 5 for instruction: %V8I32 = mul <8 x i32> undef, undef
diff --git a/llvm/test/Analysis/CostModel/X86/reduce-mul.ll b/llvm/test/Analysis/CostModel/X86/reduce-mul.ll
index e2d3ebc1d92b..06eb8ed36b07 100644
--- a/llvm/test/Analysis/CostModel/X86/reduce-mul.ll
+++ b/llvm/test/Analysis/CostModel/X86/reduce-mul.ll
@@ -21,8 +21,8 @@ define i32 @reduce_i64(i32 %arg) {
; AVX1-NEXT: Cost Model: Found an estimated cost of 0 for instruction: %V1 = call i64 @llvm.vector.reduce.mul.v1i64(<1 x i64> undef)
; AVX1-NEXT: Cost Model: Found an estimated cost of 10 for instruction: %V2 = call i64 @llvm.vector.reduce.mul.v2i64(<2 x i64> undef)
; AVX1-NEXT: Cost Model: Found an estimated cost of 19 for instruction: %V4 = call i64 @llvm.vector.reduce.mul.v4i64(<4 x i64> undef)
-; AVX1-NEXT: Cost Model: Found an estimated cost of 37 for instruction: %V8 = call i64 @llvm.vector.reduce.mul.v8i64(<8 x i64> undef)
-; AVX1-NEXT: Cost Model: Found an estimated cost of 73 for instruction: %V16 = call i64 @llvm.vector.reduce.mul.v16i64(<16 x i64> undef)
+; AVX1-NEXT: Cost Model: Found an estimated cost of 31 for instruction: %V8 = call i64 @llvm.vector.reduce.mul.v8i64(<8 x i64> undef)
+; AVX1-NEXT: Cost Model: Found an estimated cost of 55 for instruction: %V16 = call i64 @llvm.vector.reduce.mul.v16i64(<16 x i64> undef)
; AVX1-NEXT: Cost Model: Found an estimated cost of 0 for instruction: ret i32 undef
;
; AVX2-LABEL: 'reduce_i64'
diff --git a/llvm/test/Analysis/CostModel/X86/rem.ll b/llvm/test/Analysis/CostModel/X86/rem.ll
index 8c34117ce1d1..f0b5772efede 100644
--- a/llvm/test/Analysis/CostModel/X86/rem.ll
+++ b/llvm/test/Analysis/CostModel/X86/rem.ll
@@ -786,8 +786,8 @@ define i32 @srem_constpow2() {
; AVX1-LABEL: 'srem_constpow2'
; AVX1-NEXT: Cost Model: Found an estimated cost of 6 for instruction: %I64 = srem i64 undef, 16
; AVX1-NEXT: Cost Model: Found an estimated cost of 38 for instruction: %V2i64 = srem <2 x i64> undef, <i64 8, i64 16>
-; AVX1-NEXT: Cost Model: Found an estimated cost of 88 for instruction: %V4i64 = srem <4 x i64> undef, <i64 2, i64 4, i64 8, i64 16>
-; AVX1-NEXT: Cost Model: Found an estimated cost of 176 for instruction: %V8i64 = srem <8 x i64> undef, <i64 2, i64 4, i64 8, i64 16, i64 32, i64 64, i64 128, i64 256>
+; AVX1-NEXT: Cost Model: Found an estimated cost of 82 for instruction: %V4i64 = srem <4 x i64> undef, <i64 2, i64 4, i64 8, i64 16>
+; AVX1-NEXT: Cost Model: Found an estimated cost of 164 for instruction: %V8i64 = srem <8 x i64> undef, <i64 2, i64 4, i64 8, i64 16, i64 32, i64 64, i64 128, i64 256>
; AVX1-NEXT: Cost Model: Found an estimated cost of 6 for instruction: %I32 = srem i32 undef, 16
; AVX1-NEXT: Cost Model: Found an estimated cost of 39 for instruction: %V4i32 = srem <4 x i32> undef, <i32 2, i32 4, i32 8, i32 16>
; AVX1-NEXT: Cost Model: Found an estimated cost of 89 for instruction: %V8i32 = srem <8 x i32> undef, <i32 2, i32 4, i32 8, i32 16, i32 32, i32 64, i32 128, i32 256>
@@ -900,8 +900,8 @@ define i32 @srem_constpow2() {
; BTVER2-LABEL: 'srem_constpow2'
; BTVER2-NEXT: Cost Model: Found an estimated cost of 6 for instruction: %I64 = srem i64 undef, 16
; BTVER2-NEXT: Cost Model: Found an estimated cost of 38 for instruction: %V2i64 = srem <2 x i64> undef, <i64 8, i64 16>
-; BTVER2-NEXT: Cost Model: Found an estimated cost of 88 for instruction: %V4i64 = srem <4 x i64> undef, <i64 2, i64 4, i64 8, i64 16>
-; BTVER2-NEXT: Cost Model: Found an estimated cost of 176 for instruction: %V8i64 = srem <8 x i64> undef, <i64 2, i64 4, i64 8, i64 16, i64 32, i64 64, i64 128, i64 256>
+; BTVER2-NEXT: Cost Model: Found an estimated cost of 82 for instruction: %V4i64 = srem <4 x i64> undef, <i64 2, i64 4, i64 8, i64 16>
+; BTVER2-NEXT: Cost Model: Found an estimated cost of 164 for instruction: %V8i64 = srem <8 x i64> undef, <i64 2, i64 4, i64 8, i64 16, i64 32, i64 64, i64 128, i64 256>
; BTVER2-NEXT: Cost Model: Found an estimated cost of 6 for instruction: %I32 = srem i32 undef, 16
; BTVER2-NEXT: Cost Model: Found an estimated cost of 39 for instruction: %V4i32 = srem <4 x i32> undef, <i32 2, i32 4, i32 8, i32 16>
; BTVER2-NEXT: Cost Model: Found an estimated cost of 89 for instruction: %V8i32 = srem <8 x i32> undef, <i32 2, i32 4, i32 8, i32 16, i32 32, i32 64, i32 128, i32 256>
@@ -1081,8 +1081,8 @@ define i32 @srem_uniformconstpow2() {
; AVX1-LABEL: 'srem_uniformconstpow2'
; AVX1-NEXT: Cost Model: Found an estimated cost of 6 for instruction: %I64 = srem i64 undef, 16
; AVX1-NEXT: Cost Model: Found an estimated cost of 19 for instruction: %V2i64 = srem <2 x i64> undef, <i64 16, i64 16>
-; AVX1-NEXT: Cost Model: Found an estimated cost of 50 for instruction: %V4i64 = srem <4 x i64> undef, <i64 16, i64 16, i64 16, i64 16>
-; AVX1-NEXT: Cost Model: Found an estimated cost of 100 for instruction: %V8i64 = srem <8 x i64> undef, <i64 16, i64 16, i64 16, i64 16, i64 16, i64 16, i64 16, i64 16>
+; AVX1-NEXT: Cost Model: Found an estimated cost of 44 for instruction: %V4i64 = srem <4 x i64> undef, <i64 16, i64 16, i64 16, i64 16>
+; AVX1-NEXT: Cost Model: Found an estimated cost of 88 for instruction: %V8i64 = srem <8 x i64> undef, <i64 16, i64 16, i64 16, i64 16, i64 16, i64 16, i64 16, i64 16>
; AVX1-NEXT: Cost Model: Found an estimated cost of 6 for instruction: %I32 = srem i32 undef, 16
; AVX1-NEXT: Cost Model: Found an estimated cost of 7 for instruction: %V4i32 = srem <4 x i32> undef, <i32 16, i32 16, i32 16, i32 16>
; AVX1-NEXT: Cost Model: Found an estimated cost of 25 for instruction: %V8i32 = srem <8 x i32> undef, <i32 16, i32 16, i32 16, i32 16, i32 16, i32 16, i32 16, i32 16>
@@ -1195,8 +1195,8 @@ define i32 @srem_uniformconstpow2() {
; BTVER2-LABEL: 'srem_uniformconstpow2'
; BTVER2-NEXT: Cost Model: Found an estimated cost of 6 for instruction: %I64 = srem i64 undef, 16
; BTVER2-NEXT: Cost Model: Found an estimated cost of 19 for instruction: %V2i64 = srem <2 x i64> undef, <i64 16, i64 16>
-; BTVER2-NEXT: Cost Model: Found an estimated cost of 50 for instruction: %V4i64 = srem <4 x i64> undef, <i64 16, i64 16, i64 16, i64 16>
-; BTVER2-NEXT: Cost Model: Found an estimated cost of 100 for instruction: %V8i64 = srem <8 x i64> undef, <i64 16, i64 16, i64 16, i64 16, i64 16, i64 16, i64 16, i64 16>
+; BTVER2-NEXT: Cost Model: Found an estimated cost of 44 for instruction: %V4i64 = srem <4 x i64> undef, <i64 16, i64 16, i64 16, i64 16>
+; BTVER2-NEXT: Cost Model: Found an estimated cost of 88 for instruction: %V8i64 = srem <8 x i64> undef, <i64 16, i64 16, i64 16, i64 16, i64 16, i64 16, i64 16, i64 16>
; BTVER2-NEXT: Cost Model: Found an estimated cost of 6 for instruction: %I32 = srem i32 undef, 16
; BTVER2-NEXT: Cost Model: Found an estimated cost of 7 for instruction: %V4i32 = srem <4 x i32> undef, <i32 16, i32 16, i32 16, i32 16>
; BTVER2-NEXT: Cost Model: Found an estimated cost of 25 for instruction: %V8i32 = srem <8 x i32> undef, <i32 16, i32 16, i32 16, i32 16, i32 16, i32 16, i32 16, i32 16>
diff --git a/llvm/test/Transforms/SLPVectorizer/X86/arith-fix.ll b/llvm/test/Transforms/SLPVectorizer/X86/arith-fix.ll
index 11ee76ef8c21..375fae877534 100644
--- a/llvm/test/Transforms/SLPVectorizer/X86/arith-fix.ll
+++ b/llvm/test/Transforms/SLPVectorizer/X86/arith-fix.ll
@@ -282,83 +282,16 @@ define void @smul_v16i32() {
; SLM-NEXT: store i32 [[R15]], i32* getelementptr inbounds ([16 x i32], [16 x i32]* @c32, i32 0, i64 15), align 4
; SLM-NEXT: ret void
;
-; AVX1-LABEL: @smul_v16i32(
-; AVX1-NEXT: [[A0:%.*]] = load i32, i32* getelementptr inbounds ([16 x i32], [16 x i32]* @a32, i32 0, i64 0), align 4
-; AVX1-NEXT: [[A1:%.*]] = load i32, i32* getelementptr inbounds ([16 x i32], [16 x i32]* @a32, i32 0, i64 1), align 4
-; AVX1-NEXT: [[A2:%.*]] = load i32, i32* getelementptr inbounds ([16 x i32], [16 x i32]* @a32, i32 0, i64 2), align 4
-; AVX1-NEXT: [[A3:%.*]] = load i32, i32* getelementptr inbounds ([16 x i32], [16 x i32]* @a32, i32 0, i64 3), align 4
-; AVX1-NEXT: [[A4:%.*]] = load i32, i32* getelementptr inbounds ([16 x i32], [16 x i32]* @a32, i32 0, i64 4), align 4
-; AVX1-NEXT: [[A5:%.*]] = load i32, i32* getelementptr inbounds ([16 x i32], [16 x i32]* @a32, i32 0, i64 5), align 4
-; AVX1-NEXT: [[A6:%.*]] = load i32, i32* getelementptr inbounds ([16 x i32], [16 x i32]* @a32, i32 0, i64 6), align 4
-; AVX1-NEXT: [[A7:%.*]] = load i32, i32* getelementptr inbounds ([16 x i32], [16 x i32]* @a32, i32 0, i64 7), align 4
-; AVX1-NEXT: [[A8:%.*]] = load i32, i32* getelementptr inbounds ([16 x i32], [16 x i32]* @a32, i32 0, i64 8), align 4
-; AVX1-NEXT: [[A9:%.*]] = load i32, i32* getelementptr inbounds ([16 x i32], [16 x i32]* @a32, i32 0, i64 9), align 4
-; AVX1-NEXT: [[A10:%.*]] = load i32, i32* getelementptr inbounds ([16 x i32], [16 x i32]* @a32, i32 0, i64 10), align 4
-; AVX1-NEXT: [[A11:%.*]] = load i32, i32* getelementptr inbounds ([16 x i32], [16 x i32]* @a32, i32 0, i64 11), align 4
-; AVX1-NEXT: [[A12:%.*]] = load i32, i32* getelementptr inbounds ([16 x i32], [16 x i32]* @a32, i32 0, i64 12), align 4
-; AVX1-NEXT: [[A13:%.*]] = load i32, i32* getelementptr inbounds ([16 x i32], [16 x i32]* @a32, i32 0, i64 13), align 4
-; AVX1-NEXT: [[A14:%.*]] = load i32, i32* getelementptr inbounds ([16 x i32], [16 x i32]* @a32, i32 0, i64 14), align 4
-; AVX1-NEXT: [[A15:%.*]] = load i32, i32* getelementptr inbounds ([16 x i32], [16 x i32]* @a32, i32 0, i64 15), align 4
-; AVX1-NEXT: [[B0:%.*]] = load i32, i32* getelementptr inbounds ([16 x i32], [16 x i32]* @b32, i32 0, i64 0), align 4
-; AVX1-NEXT: [[B1:%.*]] = load i32, i32* getelementptr inbounds ([16 x i32], [16 x i32]* @b32, i32 0, i64 1), align 4
-; AVX1-NEXT: [[B2:%.*]] = load i32, i32* getelementptr inbounds ([16 x i32], [16 x i32]* @b32, i32 0, i64 2), align 4
-; AVX1-NEXT: [[B3:%.*]] = load i32, i32* getelementptr inbounds ([16 x i32], [16 x i32]* @b32, i32 0, i64 3), align 4
-; AVX1-NEXT: [[B4:%.*]] = load i32, i32* getelementptr inbounds ([16 x i32], [16 x i32]* @b32, i32 0, i64 4), align 4
-; AVX1-NEXT: [[B5:%.*]] = load i32, i32* getelementptr inbounds ([16 x i32], [16 x i32]* @b32, i32 0, i64 5), align 4
-; AVX1-NEXT: [[B6:%.*]] = load i32, i32* getelementptr inbounds ([16 x i32], [16 x i32]* @b32, i32 0, i64 6), align 4
-; AVX1-NEXT: [[B7:%.*]] = load i32, i32* getelementptr inbounds ([16 x i32], [16 x i32]* @b32, i32 0, i64 7), align 4
-; AVX1-NEXT: [[B8:%.*]] = load i32, i32* getelementptr inbounds ([16 x i32], [16 x i32]* @b32, i32 0, i64 8), align 4
-; AVX1-NEXT: [[B9:%.*]] = load i32, i32* getelementptr inbounds ([16 x i32], [16 x i32]* @b32, i32 0, i64 9), align 4
-; AVX1-NEXT: [[B10:%.*]] = load i32, i32* getelementptr inbounds ([16 x i32], [16 x i32]* @b32, i32 0, i64 10), align 4
-; AVX1-NEXT: [[B11:%.*]] = load i32, i32* getelementptr inbounds ([16 x i32], [16 x i32]* @b32, i32 0, i64 11), align 4
-; AVX1-NEXT: [[B12:%.*]] = load i32, i32* getelementptr inbounds ([16 x i32], [16 x i32]* @b32, i32 0, i64 12), align 4
-; AVX1-NEXT: [[B13:%.*]] = load i32, i32* getelementptr inbounds ([16 x i32], [16 x i32]* @b32, i32 0, i64 13), align 4
-; AVX1-NEXT: [[B14:%.*]] = load i32, i32* getelementptr inbounds ([16 x i32], [16 x i32]* @b32, i32 0, i64 14), align 4
-; AVX1-NEXT: [[B15:%.*]] = load i32, i32* getelementptr inbounds ([16 x i32], [16 x i32]* @b32, i32 0, i64 15), align 4
-; AVX1-NEXT: [[R0:%.*]] = call i32 @llvm.smul.fix.i32(i32 [[A0]], i32 [[B0]], i32 3)
-; AVX1-NEXT: [[R1:%.*]] = call i32 @llvm.smul.fix.i32(i32 [[A1]], i32 [[B1]], i32 3)
-; AVX1-NEXT: [[R2:%.*]] = call i32 @llvm.smul.fix.i32(i32 [[A2]], i32 [[B2]], i32 3)
-; AVX1-NEXT: [[R3:%.*]] = call i32 @llvm.smul.fix.i32(i32 [[A3]], i32 [[B3]], i32 3)
-; AVX1-NEXT: [[R4:%.*]] = call i32 @llvm.smul.fix.i32(i32 [[A4]], i32 [[B4]], i32 3)
-; AVX1-NEXT: [[R5:%.*]] = call i32 @llvm.smul.fix.i32(i32 [[A5]], i32 [[B5]], i32 3)
-; AVX1-NEXT: [[R6:%.*]] = call i32 @llvm.smul.fix.i32(i32 [[A6]], i32 [[B6]], i32 3)
-; AVX1-NEXT: [[R7:%.*]] = call i32 @llvm.smul.fix.i32(i32 [[A7]], i32 [[B7]], i32 3)
-; AVX1-NEXT: [[R8:%.*]] = call i32 @llvm.smul.fix.i32(i32 [[A8]], i32 [[B8]], i32 3)
-; AVX1-NEXT: [[R9:%.*]] = call i32 @llvm.smul.fix.i32(i32 [[A9]], i32 [[B9]], i32 3)
-; AVX1-NEXT: [[R10:%.*]] = call i32 @llvm.smul.fix.i32(i32 [[A10]], i32 [[B10]], i32 3)
-; AVX1-NEXT: [[R11:%.*]] = call i32 @llvm.smul.fix.i32(i32 [[A11]], i32 [[B11]], i32 3)
-; AVX1-NEXT: [[R12:%.*]] = call i32 @llvm.smul.fix.i32(i32 [[A12]], i32 [[B12]], i32 3)
-; AVX1-NEXT: [[R13:%.*]] = call i32 @llvm.smul.fix.i32(i32 [[A13]], i32 [[B13]], i32 3)
-; AVX1-NEXT: [[R14:%.*]] = call i32 @llvm.smul.fix.i32(i32 [[A14]], i32 [[B14]], i32 3)
-; AVX1-NEXT: [[R15:%.*]] = call i32 @llvm.smul.fix.i32(i32 [[A15]], i32 [[B15]], i32 3)
-; AVX1-NEXT: store i32 [[R0]], i32* getelementptr inbounds ([16 x i32], [16 x i32]* @c32, i32 0, i64 0), align 4
-; AVX1-NEXT: store i32 [[R1]], i32* getelementptr inbounds ([16 x i32], [16 x i32]* @c32, i32 0, i64 1), align 4
-; AVX1-NEXT: store i32 [[R2]], i32* getelementptr inbounds ([16 x i32], [16 x i32]* @c32, i32 0, i64 2), align 4
-; AVX1-NEXT: store i32 [[R3]], i32* getelementptr inbounds ([16 x i32], [16 x i32]* @c32, i32 0, i64 3), align 4
-; AVX1-NEXT: store i32 [[R4]], i32* getelementptr inbounds ([16 x i32], [16 x i32]* @c32, i32 0, i64 4), align 4
-; AVX1-NEXT: store i32 [[R5]], i32* getelementptr inbounds ([16 x i32], [16 x i32]* @c32, i32 0, i64 5), align 4
-; AVX1-NEXT: store i32 [[R6]], i32* getelementptr inbounds ([16 x i32], [16 x i32]* @c32, i32 0, i64 6), align 4
-; AVX1-NEXT: store i32 [[R7]], i32* getelementptr inbounds ([16 x i32], [16 x i32]* @c32, i32 0, i64 7), align 4
-; AVX1-NEXT: store i32 [[R8]], i32* getelementptr inbounds ([16 x i32], [16 x i32]* @c32, i32 0, i64 8), align 4
-; AVX1-NEXT: store i32 [[R9]], i32* getelementptr inbounds ([16 x i32], [16 x i32]* @c32, i32 0, i64 9), align 4
-; AVX1-NEXT: store i32 [[R10]], i32* getelementptr inbounds ([16 x i32], [16 x i32]* @c32, i32 0, i64 10), align 4
-; AVX1-NEXT: store i32 [[R11]], i32* getelementptr inbounds ([16 x i32], [16 x i32]* @c32, i32 0, i64 11), align 4
-; AVX1-NEXT: store i32 [[R12]], i32* getelementptr inbounds ([16 x i32], [16 x i32]* @c32, i32 0, i64 12), align 4
-; AVX1-NEXT: store i32 [[R13]], i32* getelementptr inbounds ([16 x i32], [16 x i32]* @c32, i32 0, i64 13), align 4
-; AVX1-NEXT: store i32 [[R14]], i32* getelementptr inbounds ([16 x i32], [16 x i32]* @c32, i32 0, i64 14), align 4
-; AVX1-NEXT: store i32 [[R15]], i32* getelementptr inbounds ([16 x i32], [16 x i32]* @c32, i32 0, i64 15), align 4
-; AVX1-NEXT: ret void
-;
-; AVX2-LABEL: @smul_v16i32(
-; AVX2-NEXT: [[TMP1:%.*]] = load <8 x i32>, <8 x i32>* bitcast ([16 x i32]* @a32 to <8 x i32>*), align 4
-; AVX2-NEXT: [[TMP2:%.*]] = load <8 x i32>, <8 x i32>* bitcast (i32* getelementptr inbounds ([16 x i32], [16 x i32]* @a32, i32 0, i64 8) to <8 x i32>*), align 4
-; AVX2-NEXT: [[TMP3:%.*]] = load <8 x i32>, <8 x i32>* bitcast ([16 x i32]* @b32 to <8 x i32>*), align 4
-; AVX2-NEXT: [[TMP4:%.*]] = load <8 x i32>, <8 x i32>* bitcast (i32* getelementptr inbounds ([16 x i32], [16 x i32]* @b32, i32 0, i64 8) to <8 x i32>*), align 4
-; AVX2-NEXT: [[TMP5:%.*]] = call <8 x i32> @llvm.smul.fix.v8i32(<8 x i32> [[TMP1]], <8 x i32> [[TMP3]], i32 3)
-; AVX2-NEXT: [[TMP6:%.*]] = call <8 x i32> @llvm.smul.fix.v8i32(<8 x i32> [[TMP2]], <8 x i32> [[TMP4]], i32 3)
-; AVX2-NEXT: store <8 x i32> [[TMP5]], <8 x i32>* bitcast ([16 x i32]* @c32 to <8 x i32>*), align 4
-; AVX2-NEXT: store <8 x i32> [[TMP6]], <8 x i32>* bitcast (i32* getelementptr inbounds ([16 x i32], [16 x i32]* @c32, i32 0, i64 8) to <8 x i32>*), align 4
-; AVX2-NEXT: ret void
+; AVX-LABEL: @smul_v16i32(
+; AVX-NEXT: [[TMP1:%.*]] = load <8 x i32>, <8 x i32>* bitcast ([16 x i32]* @a32 to <8 x i32>*), align 4
+; AVX-NEXT: [[TMP2:%.*]] = load <8 x i32>, <8 x i32>* bitcast (i32* getelementptr inbounds ([16 x i32], [16 x i32]* @a32, i32 0, i64 8) to <8 x i32>*), align 4
+; AVX-NEXT: [[TMP3:%.*]] = load <8 x i32>, <8 x i32>* bitcast ([16 x i32]* @b32 to <8 x i32>*), align 4
+; AVX-NEXT: [[TMP4:%.*]] = load <8 x i32>, <8 x i32>* bitcast (i32* getelementptr inbounds ([16 x i32], [16 x i32]* @b32, i32 0, i64 8) to <8 x i32>*), align 4
+; AVX-NEXT: [[TMP5:%.*]] = call <8 x i32> @llvm.smul.fix.v8i32(<8 x i32> [[TMP1]], <8 x i32> [[TMP3]], i32 3)
+; AVX-NEXT: [[TMP6:%.*]] = call <8 x i32> @llvm.smul.fix.v8i32(<8 x i32> [[TMP2]], <8 x i32> [[TMP4]], i32 3)
+; AVX-NEXT: store <8 x i32> [[TMP5]], <8 x i32>* bitcast ([16 x i32]* @c32 to <8 x i32>*), align 4
+; AVX-NEXT: store <8 x i32> [[TMP6]], <8 x i32>* bitcast (i32* getelementptr inbounds ([16 x i32], [16 x i32]* @c32, i32 0, i64 8) to <8 x i32>*), align 4
+; AVX-NEXT: ret void
;
; AVX512-LABEL: @smul_v16i32(
; AVX512-NEXT: [[TMP1:%.*]] = load <16 x i32>, <16 x i32>* bitcast ([16 x i32]* @a32 to <16 x i32>*), align 4
@@ -366,17 +299,6 @@ define void @smul_v16i32() {
; AVX512-NEXT: [[TMP3:%.*]] = call <16 x i32> @llvm.smul.fix.v16i32(<16 x i32> [[TMP1]], <16 x i32> [[TMP2]], i32 3)
; AVX512-NEXT: store <16 x i32> [[TMP3]], <16 x i32>* bitcast ([16 x i32]* @c32 to <16 x i32>*), align 4
; AVX512-NEXT: ret void
-;
-; AVX256BW-LABEL: @smul_v16i32(
-; AVX256BW-NEXT: [[TMP1:%.*]] = load <8 x i32>, <8 x i32>* bitcast ([16 x i32]* @a32 to <8 x i32>*), align 4
-; AVX256BW-NEXT: [[TMP2:%.*]] = load <8 x i32>, <8 x i32>* bitcast (i32* getelementptr inbounds ([16 x i32], [16 x i32]* @a32, i32 0, i64 8) to <8 x i32>*), align 4
-; AVX256BW-NEXT: [[TMP3:%.*]] = load <8 x i32>, <8 x i32>* bitcast ([16 x i32]* @b32 to <8 x i32>*), align 4
-; AVX256BW-NEXT: [[TMP4:%.*]] = load <8 x i32>, <8 x i32>* bitcast (i32* getelementptr inbounds ([16 x i32], [16 x i32]* @b32, i32 0, i64 8) to <8 x i32>*), align 4
-; AVX256BW-NEXT: [[TMP5:%.*]] = call <8 x i32> @llvm.smul.fix.v8i32(<8 x i32> [[TMP1]], <8 x i32> [[TMP3]], i32 3)
-; AVX256BW-NEXT: [[TMP6:%.*]] = call <8 x i32> @llvm.smul.fix.v8i32(<8 x i32> [[TMP2]], <8 x i32> [[TMP4]], i32 3)
-; AVX256BW-NEXT: store <8 x i32> [[TMP5]], <8 x i32>* bitcast ([16 x i32]* @c32 to <8 x i32>*), align 4
-; AVX256BW-NEXT: store <8 x i32> [[TMP6]], <8 x i32>* bitcast (i32* getelementptr inbounds ([16 x i32], [16 x i32]* @c32, i32 0, i64 8) to <8 x i32>*), align 4
-; AVX256BW-NEXT: ret void
;
%a0 = load i32, i32* getelementptr inbounds ([16 x i32], [16 x i32]* @a32, i32 0, i64 0 ), align 4
%a1 = load i32, i32* getelementptr inbounds ([16 x i32], [16 x i32]* @a32, i32 0, i64 1 ), align 4
diff --git a/llvm/test/Transforms/SLPVectorizer/X86/pr46983.ll b/llvm/test/Transforms/SLPVectorizer/X86/pr46983.ll
index b1cf6ad6a1f6..e6edc82496bc 100644
--- a/llvm/test/Transforms/SLPVectorizer/X86/pr46983.ll
+++ b/llvm/test/Transforms/SLPVectorizer/X86/pr46983.ll
@@ -1,14 +1,14 @@
; NOTE: Assertions have been autogenerated by utils/update_test_checks.py
; RUN: opt < %s -slp-vectorizer -instcombine -S -mtriple=x86_64-unknown-linux-gnu -mattr=+sse2 | FileCheck %s --check-prefixes=CHECK,SSE
; RUN: opt < %s -slp-vectorizer -instcombine -S -mtriple=x86_64-unknown-linux-gnu -mattr=+sse4.2 | FileCheck %s --check-prefixes=CHECK,SSE
-; RUN: opt < %s -slp-vectorizer -instcombine -S -mtriple=x86_64-unknown-linux-gnu -mattr=+avx | FileCheck %s --check-prefixes=CHECK,AVX1
-; RUN: opt < %s -slp-vectorizer -instcombine -S -mtriple=x86_64-unknown-linux-gnu -mattr=+avx2 | FileCheck %s --check-prefixes=CHECK,AVX2
-; RUN: opt < %s -slp-vectorizer -instcombine -S -mtriple=x86_64-unknown-linux-gnu -mattr=+avx512bw,+avx512vl | FileCheck %s --check-prefixes=CHECK,AVX2
+; RUN: opt < %s -slp-vectorizer -instcombine -S -mtriple=x86_64-unknown-linux-gnu -mattr=+avx | FileCheck %s --check-prefixes=CHECK,AVX
+; RUN: opt < %s -slp-vectorizer -instcombine -S -mtriple=x86_64-unknown-linux-gnu -mattr=+avx2 | FileCheck %s --check-prefixes=CHECK,AVX
+; RUN: opt < %s -slp-vectorizer -instcombine -S -mtriple=x86_64-unknown-linux-gnu -mattr=+avx512bw,+avx512vl | FileCheck %s --check-prefixes=CHECK,AVX
define void @store_i32(i32* nocapture %0, i32 %1, i32 %2) {
; CHECK-LABEL: @store_i32(
; CHECK-NEXT: [[TMP4:%.*]] = bitcast i32* [[TMP0:%.*]] to <4 x i32>*
-; CHECK-NEXT: [[TMP5:%.*]] = load <4 x i32>, <4 x i32>* [[TMP4]], align 4, [[TBAA0:!tbaa !.*]]
+; CHECK-NEXT: [[TMP5:%.*]] = load <4 x i32>, <4 x i32>* [[TMP4]], align 4, !tbaa [[TBAA0:![0-9]+]]
; CHECK-NEXT: [[TMP6:%.*]] = insertelement <4 x i32> poison, i32 [[TMP1:%.*]], i32 0
; CHECK-NEXT: [[TMP7:%.*]] = shufflevector <4 x i32> [[TMP6]], <4 x i32> undef, <4 x i32> zeroinitializer
; CHECK-NEXT: [[TMP8:%.*]] = mul <4 x i32> [[TMP5]], [[TMP7]]
@@ -16,7 +16,7 @@ define void @store_i32(i32* nocapture %0, i32 %1, i32 %2) {
; CHECK-NEXT: [[TMP10:%.*]] = icmp ult <4 x i32> [[TMP9]], <i32 255, i32 255, i32 255, i32 255>
; CHECK-NEXT: [[TMP11:%.*]] = select <4 x i1> [[TMP10]], <4 x i32> [[TMP9]], <4 x i32> <i32 255, i32 255, i32 255, i32 255>
; CHECK-NEXT: [[TMP12:%.*]] = bitcast i32* [[TMP0]] to <4 x i32>*
-; CHECK-NEXT: store <4 x i32> [[TMP11]], <4 x i32>* [[TMP12]], align 4, [[TBAA0]]
+; CHECK-NEXT: store <4 x i32> [[TMP11]], <4 x i32>* [[TMP12]], align 4, !tbaa [[TBAA0]]
; CHECK-NEXT: ret void
;
%4 = load i32, i32* %0, align 4, !tbaa !2
@@ -52,7 +52,7 @@ define void @store_i32(i32* nocapture %0, i32 %1, i32 %2) {
define void @store_i8(i8* nocapture %0, i32 %1, i32 %2) {
; CHECK-LABEL: @store_i8(
; CHECK-NEXT: [[TMP4:%.*]] = bitcast i8* [[TMP0:%.*]] to <4 x i8>*
-; CHECK-NEXT: [[TMP5:%.*]] = load <4 x i8>, <4 x i8>* [[TMP4]], align 1, [[TBAA4:!tbaa !.*]]
+; CHECK-NEXT: [[TMP5:%.*]] = load <4 x i8>, <4 x i8>* [[TMP4]], align 1, !tbaa [[TBAA4:![0-9]+]]
; CHECK-NEXT: [[TMP6:%.*]] = zext <4 x i8> [[TMP5]] to <4 x i32>
; CHECK-NEXT: [[TMP7:%.*]] = insertelement <4 x i32> poison, i32 [[TMP1:%.*]], i32 0
; CHECK-NEXT: [[TMP8:%.*]] = shufflevector <4 x i32> [[TMP7]], <4 x i32> undef, <4 x i32> zeroinitializer
@@ -62,7 +62,7 @@ define void @store_i8(i8* nocapture %0, i32 %1, i32 %2) {
; CHECK-NEXT: [[TMP12:%.*]] = select <4 x i1> [[TMP11]], <4 x i32> [[TMP10]], <4 x i32> <i32 255, i32 255, i32 255, i32 255>
; CHECK-NEXT: [[TMP13:%.*]] = trunc <4 x i32> [[TMP12]] to <4 x i8>
; CHECK-NEXT: [[TMP14:%.*]] = bitcast i8* [[TMP0]] to <4 x i8>*
-; CHECK-NEXT: store <4 x i8> [[TMP13]], <4 x i8>* [[TMP14]], align 1, [[TBAA4]]
+; CHECK-NEXT: store <4 x i8> [[TMP13]], <4 x i8>* [[TMP14]], align 1, !tbaa [[TBAA4]]
; CHECK-NEXT: ret void
;
%4 = load i8, i8* %0, align 1, !tbaa !6
@@ -106,97 +106,58 @@ define void @store_i8(i8* nocapture %0, i32 %1, i32 %2) {
define void @store_i64(i64* nocapture %0, i32 %1, i32 %2) {
; SSE-LABEL: @store_i64(
; SSE-NEXT: [[TMP4:%.*]] = zext i32 [[TMP1:%.*]] to i64
-; SSE-NEXT: [[TMP5:%.*]] = load i64, i64* [[TMP0:%.*]], align 8, [[TBAA5:!tbaa !.*]]
+; SSE-NEXT: [[TMP5:%.*]] = load i64, i64* [[TMP0:%.*]], align 8, !tbaa [[TBAA5:![0-9]+]]
; SSE-NEXT: [[TMP6:%.*]] = mul i64 [[TMP5]], [[TMP4]]
; SSE-NEXT: [[TMP7:%.*]] = lshr i64 [[TMP6]], 15
; SSE-NEXT: [[TMP8:%.*]] = trunc i64 [[TMP7]] to i32
; SSE-NEXT: [[TMP9:%.*]] = icmp ult i32 [[TMP8]], 255
; SSE-NEXT: [[TMP10:%.*]] = and i64 [[TMP7]], 4294967295
; SSE-NEXT: [[TMP11:%.*]] = select i1 [[TMP9]], i64 [[TMP10]], i64 255
-; SSE-NEXT: store i64 [[TMP11]], i64* [[TMP0]], align 8, [[TBAA5]]
+; SSE-NEXT: store i64 [[TMP11]], i64* [[TMP0]], align 8, !tbaa [[TBAA5]]
; SSE-NEXT: [[TMP12:%.*]] = getelementptr inbounds i64, i64* [[TMP0]], i64 1
-; SSE-NEXT: [[TMP13:%.*]] = load i64, i64* [[TMP12]], align 8, [[TBAA5]]
+; SSE-NEXT: [[TMP13:%.*]] = load i64, i64* [[TMP12]], align 8, !tbaa [[TBAA5]]
; SSE-NEXT: [[TMP14:%.*]] = mul i64 [[TMP13]], [[TMP4]]
; SSE-NEXT: [[TMP15:%.*]] = lshr i64 [[TMP14]], 15
; SSE-NEXT: [[TMP16:%.*]] = trunc i64 [[TMP15]] to i32
; SSE-NEXT: [[TMP17:%.*]] = icmp ult i32 [[TMP16]], 255
; SSE-NEXT: [[TMP18:%.*]] = and i64 [[TMP15]], 4294967295
; SSE-NEXT: [[TMP19:%.*]] = select i1 [[TMP17]], i64 [[TMP18]], i64 255
-; SSE-NEXT: store i64 [[TMP19]], i64* [[TMP12]], align 8, [[TBAA5]]
+; SSE-NEXT: store i64 [[TMP19]], i64* [[TMP12]], align 8, !tbaa [[TBAA5]]
; SSE-NEXT: [[TMP20:%.*]] = getelementptr inbounds i64, i64* [[TMP0]], i64 2
-; SSE-NEXT: [[TMP21:%.*]] = load i64, i64* [[TMP20]], align 8, [[TBAA5]]
+; SSE-NEXT: [[TMP21:%.*]] = load i64, i64* [[TMP20]], align 8, !tbaa [[TBAA5]]
; SSE-NEXT: [[TMP22:%.*]] = mul i64 [[TMP21]], [[TMP4]]
; SSE-NEXT: [[TMP23:%.*]] = lshr i64 [[TMP22]], 15
; SSE-NEXT: [[TMP24:%.*]] = trunc i64 [[TMP23]] to i32
; SSE-NEXT: [[TMP25:%.*]] = icmp ult i32 [[TMP24]], 255
; SSE-NEXT: [[TMP26:%.*]] = and i64 [[TMP23]], 4294967295
; SSE-NEXT: [[TMP27:%.*]] = select i1 [[TMP25]], i64 [[TMP26]], i64 255
-; SSE-NEXT: store i64 [[TMP27]], i64* [[TMP20]], align 8, [[TBAA5]]
+; SSE-NEXT: store i64 [[TMP27]], i64* [[TMP20]], align 8, !tbaa [[TBAA5]]
; SSE-NEXT: [[TMP28:%.*]] = getelementptr inbounds i64, i64* [[TMP0]], i64 3
-; SSE-NEXT: [[TMP29:%.*]] = load i64, i64* [[TMP28]], align 8, [[TBAA5]]
+; SSE-NEXT: [[TMP29:%.*]] = load i64, i64* [[TMP28]], align 8, !tbaa [[TBAA5]]
; SSE-NEXT: [[TMP30:%.*]] = mul i64 [[TMP29]], [[TMP4]]
; SSE-NEXT: [[TMP31:%.*]] = lshr i64 [[TMP30]], 15
; SSE-NEXT: [[TMP32:%.*]] = trunc i64 [[TMP31]] to i32
; SSE-NEXT: [[TMP33:%.*]] = icmp ult i32 [[TMP32]], 255
; SSE-NEXT: [[TMP34:%.*]] = and i64 [[TMP31]], 4294967295
; SSE-NEXT: [[TMP35:%.*]] = select i1 [[TMP33]], i64 [[TMP34]], i64 255
-; SSE-NEXT: store i64 [[TMP35]], i64* [[TMP28]], align 8, [[TBAA5]]
+; SSE-NEXT: store i64 [[TMP35]], i64* [[TMP28]], align 8, !tbaa [[TBAA5]]
; SSE-NEXT: ret void
;
-; AVX1-LABEL: @store_i64(
-; AVX1-NEXT: [[TMP4:%.*]] = zext i32 [[TMP1:%.*]] to i64
-; AVX1-NEXT: [[TMP5:%.*]] = load i64, i64* [[TMP0:%.*]], align 8, [[TBAA5:!tbaa !.*]]
-; AVX1-NEXT: [[TMP6:%.*]] = mul i64 [[TMP5]], [[TMP4]]
-; AVX1-NEXT: [[TMP7:%.*]] = lshr i64 [[TMP6]], 15
-; AVX1-NEXT: [[TMP8:%.*]] = trunc i64 [[TMP7]] to i32
-; AVX1-NEXT: [[TMP9:%.*]] = icmp ult i32 [[TMP8]], 255
-; AVX1-NEXT: [[TMP10:%.*]] = and i64 [[TMP7]], 4294967295
-; AVX1-NEXT: [[TMP11:%.*]] = select i1 [[TMP9]], i64 [[TMP10]], i64 255
-; AVX1-NEXT: store i64 [[TMP11]], i64* [[TMP0]], align 8, [[TBAA5]]
-; AVX1-NEXT: [[TMP12:%.*]] = getelementptr inbounds i64, i64* [[TMP0]], i64 1
-; AVX1-NEXT: [[TMP13:%.*]] = load i64, i64* [[TMP12]], align 8, [[TBAA5]]
-; AVX1-NEXT: [[TMP14:%.*]] = mul i64 [[TMP13]], [[TMP4]]
-; AVX1-NEXT: [[TMP15:%.*]] = lshr i64 [[TMP14]], 15
-; AVX1-NEXT: [[TMP16:%.*]] = trunc i64 [[TMP15]] to i32
-; AVX1-NEXT: [[TMP17:%.*]] = icmp ult i32 [[TMP16]], 255
-; AVX1-NEXT: [[TMP18:%.*]] = and i64 [[TMP15]], 4294967295
-; AVX1-NEXT: [[TMP19:%.*]] = select i1 [[TMP17]], i64 [[TMP18]], i64 255
-; AVX1-NEXT: store i64 [[TMP19]], i64* [[TMP12]], align 8, [[TBAA5]]
-; AVX1-NEXT: [[TMP20:%.*]] = getelementptr inbounds i64, i64* [[TMP0]], i64 2
-; AVX1-NEXT: [[TMP21:%.*]] = load i64, i64* [[TMP20]], align 8, [[TBAA5]]
-; AVX1-NEXT: [[TMP22:%.*]] = mul i64 [[TMP21]], [[TMP4]]
-; AVX1-NEXT: [[TMP23:%.*]] = lshr i64 [[TMP22]], 15
-; AVX1-NEXT: [[TMP24:%.*]] = trunc i64 [[TMP23]] to i32
-; AVX1-NEXT: [[TMP25:%.*]] = icmp ult i32 [[TMP24]], 255
-; AVX1-NEXT: [[TMP26:%.*]] = and i64 [[TMP23]], 4294967295
-; AVX1-NEXT: [[TMP27:%.*]] = select i1 [[TMP25]], i64 [[TMP26]], i64 255
-; AVX1-NEXT: store i64 [[TMP27]], i64* [[TMP20]], align 8, [[TBAA5]]
-; AVX1-NEXT: [[TMP28:%.*]] = getelementptr inbounds i64, i64* [[TMP0]], i64 3
-; AVX1-NEXT: [[TMP29:%.*]] = load i64, i64* [[TMP28]], align 8, [[TBAA5]]
-; AVX1-NEXT: [[TMP30:%.*]] = mul i64 [[TMP29]], [[TMP4]]
-; AVX1-NEXT: [[TMP31:%.*]] = lshr i64 [[TMP30]], 15
-; AVX1-NEXT: [[TMP32:%.*]] = trunc i64 [[TMP31]] to i32
-; AVX1-NEXT: [[TMP33:%.*]] = icmp ult i32 [[TMP32]], 255
-; AVX1-NEXT: [[TMP34:%.*]] = and i64 [[TMP31]], 4294967295
-; AVX1-NEXT: [[TMP35:%.*]] = select i1 [[TMP33]], i64 [[TMP34]], i64 255
-; AVX1-NEXT: store i64 [[TMP35]], i64* [[TMP28]], align 8, [[TBAA5]]
-; AVX1-NEXT: ret void
-;
-; AVX2-LABEL: @store_i64(
-; AVX2-NEXT: [[TMP4:%.*]] = zext i32 [[TMP1:%.*]] to i64
-; AVX2-NEXT: [[TMP5:%.*]] = bitcast i64* [[TMP0:%.*]] to <4 x i64>*
-; AVX2-NEXT: [[TMP6:%.*]] = load <4 x i64>, <4 x i64>* [[TMP5]], align 8, [[TBAA5:!tbaa !.*]]
-; AVX2-NEXT: [[TMP7:%.*]] = insertelement <4 x i64> poison, i64 [[TMP4]], i32 0
-; AVX2-NEXT: [[TMP8:%.*]] = shufflevector <4 x i64> [[TMP7]], <4 x i64> undef, <4 x i32> zeroinitializer
-; AVX2-NEXT: [[TMP9:%.*]] = mul <4 x i64> [[TMP6]], [[TMP8]]
-; AVX2-NEXT: [[TMP10:%.*]] = lshr <4 x i64> [[TMP9]], <i64 15, i64 15, i64 15, i64 15>
-; AVX2-NEXT: [[TMP11:%.*]] = trunc <4 x i64> [[TMP10]] to <4 x i32>
-; AVX2-NEXT: [[TMP12:%.*]] = icmp ult <4 x i32> [[TMP11]], <i32 255, i32 255, i32 255, i32 255>
-; AVX2-NEXT: [[TMP13:%.*]] = and <4 x i64> [[TMP10]], <i64 4294967295, i64 4294967295, i64 4294967295, i64 4294967295>
-; AVX2-NEXT: [[TMP14:%.*]] = select <4 x i1> [[TMP12]], <4 x i64> [[TMP13]], <4 x i64> <i64 255, i64 255, i64 255, i64 255>
-; AVX2-NEXT: [[TMP15:%.*]] = bitcast i64* [[TMP0]] to <4 x i64>*
-; AVX2-NEXT: store <4 x i64> [[TMP14]], <4 x i64>* [[TMP15]], align 8, [[TBAA5]]
-; AVX2-NEXT: ret void
+; AVX-LABEL: @store_i64(
+; AVX-NEXT: [[TMP4:%.*]] = zext i32 [[TMP1:%.*]] to i64
+; AVX-NEXT: [[TMP5:%.*]] = bitcast i64* [[TMP0:%.*]] to <4 x i64>*
+; AVX-NEXT: [[TMP6:%.*]] = load <4 x i64>, <4 x i64>* [[TMP5]], align 8, !tbaa [[TBAA5:![0-9]+]]
+; AVX-NEXT: [[TMP7:%.*]] = insertelement <4 x i64> poison, i64 [[TMP4]], i32 0
+; AVX-NEXT: [[TMP8:%.*]] = shufflevector <4 x i64> [[TMP7]], <4 x i64> undef, <4 x i32> zeroinitializer
+; AVX-NEXT: [[TMP9:%.*]] = mul <4 x i64> [[TMP6]], [[TMP8]]
+; AVX-NEXT: [[TMP10:%.*]] = lshr <4 x i64> [[TMP9]], <i64 15, i64 15, i64 15, i64 15>
+; AVX-NEXT: [[TMP11:%.*]] = trunc <4 x i64> [[TMP10]] to <4 x i32>
+; AVX-NEXT: [[TMP12:%.*]] = icmp ult <4 x i32> [[TMP11]], <i32 255, i32 255, i32 255, i32 255>
+; AVX-NEXT: [[TMP13:%.*]] = and <4 x i64> [[TMP10]], <i64 4294967295, i64 4294967295, i64 4294967295, i64 4294967295>
+; AVX-NEXT: [[TMP14:%.*]] = select <4 x i1> [[TMP12]], <4 x i64> [[TMP13]], <4 x i64> <i64 255, i64 255, i64 255, i64 255>
+; AVX-NEXT: [[TMP15:%.*]] = bitcast i64* [[TMP0]] to <4 x i64>*
+; AVX-NEXT: store <4 x i64> [[TMP14]], <4 x i64>* [[TMP15]], align 8, !tbaa [[TBAA5]]
+; AVX-NEXT: ret void
;
%4 = zext i32 %1 to i64
%5 = load i64, i64* %0, align 8, !tbaa !7
More information about the llvm-commits
mailing list