[llvm] b598c2c - [VectorCombine] add tests for scalar fneg with insert/extract; NFC

[Sanjay Patel via llvm-commits]

Author: Sanjay Patel
Date: 2022-10-06T08:57:55-04:00
New Revision: b598c2c24a0eb2cc6b5fa2ab3278be6cde43bb10

URL: https://github.com/llvm/llvm-project/commit/b598c2c24a0eb2cc6b5fa2ab3278be6cde43bb10
DIFF: https://github.com/llvm/llvm-project/commit/b598c2c24a0eb2cc6b5fa2ab3278be6cde43bb10.diff

LOG: [VectorCombine] add tests for scalar fneg with insert/extract; NFC

Added: 
    llvm/test/Transforms/VectorCombine/X86/extract-fneg-insert.ll

Modified: 
    

Removed: 
    


################################################################################
diff  --git a/llvm/test/Transforms/VectorCombine/X86/extract-fneg-insert.ll b/llvm/test/Transforms/VectorCombine/X86/extract-fneg-insert.ll
new file mode 100644
index 0000000000000..05670cabace62
--- /dev/null
+++ b/llvm/test/Transforms/VectorCombine/X86/extract-fneg-insert.ll
@@ -0,0 +1,117 @@
+; NOTE: Assertions have been autogenerated by utils/update_test_checks.py
+; RUN: opt < %s -passes=vector-combine -S -mtriple=x86_64-- -mattr=SSE2 | FileCheck %s --check-prefixes=CHECK,SSE
+; RUN: opt < %s -passes=vector-combine -S -mtriple=x86_64-- -mattr=AVX2 | FileCheck %s --check-prefixes=CHECK,AVX
+
+declare void @use(float)
+
+define <4 x float> @ext0_v4f32(<4 x float> %x, <4 x float> %y) {
+; CHECK-LABEL: @ext0_v4f32(
+; CHECK-NEXT:    [[E:%.*]] = extractelement <4 x float> [[X:%.*]], i32 0
+; CHECK-NEXT:    [[N:%.*]] = fneg float [[E]]
+; CHECK-NEXT:    [[R:%.*]] = insertelement <4 x float> [[Y:%.*]], float [[N]], i32 0
+; CHECK-NEXT:    ret <4 x float> [[R]]
+;
+  %e = extractelement <4 x float> %x, i32 0
+  %n = fneg float %e
+  %r = insertelement <4 x float> %y, float %n, i32 0
+  ret <4 x float> %r
+}
+
+; Eliminating extract/insert is profitable.
+
+define <4 x float> @ext2_v4f32(<4 x float> %x, <4 x float> %y) {
+; CHECK-LABEL: @ext2_v4f32(
+; CHECK-NEXT:    [[E:%.*]] = extractelement <4 x float> [[X:%.*]], i32 2
+; CHECK-NEXT:    [[N:%.*]] = fneg float [[E]]
+; CHECK-NEXT:    [[R:%.*]] = insertelement <4 x float> [[Y:%.*]], float [[N]], i32 2
+; CHECK-NEXT:    ret <4 x float> [[R]]
+;
+  %e = extractelement <4 x float> %x, i32 2
+  %n = fneg float %e
+  %r = insertelement <4 x float> %y, float %n, i32 2
+  ret <4 x float> %r
+}
+
+; Eliminating extract/insert is still profitable. Flags propagate.
+
+define <2 x double> @ext1_v2f64(<2 x double> %x, <2 x double> %y) {
+; CHECK-LABEL: @ext1_v2f64(
+; CHECK-NEXT:    [[E:%.*]] = extractelement <2 x double> [[X:%.*]], i32 1
+; CHECK-NEXT:    [[N:%.*]] = fneg nsz double [[E]]
+; CHECK-NEXT:    [[R:%.*]] = insertelement <2 x double> [[Y:%.*]], double [[N]], i32 1
+; CHECK-NEXT:    ret <2 x double> [[R]]
+;
+  %e = extractelement <2 x double> %x, i32 1
+  %n = fneg nsz double %e
+  %r = insertelement <2 x double> %y, double %n, i32 1
+  ret <2 x double> %r
+}
+
+define <8 x float> @ext7_v8f32(<8 x float> %x, <8 x float> %y) {
+; CHECK-LABEL: @ext7_v8f32(
+; CHECK-NEXT:    [[E:%.*]] = extractelement <8 x float> [[X:%.*]], i32 7
+; CHECK-NEXT:    [[N:%.*]] = fneg float [[E]]
+; CHECK-NEXT:    [[R:%.*]] = insertelement <8 x float> [[Y:%.*]], float [[N]], i32 7
+; CHECK-NEXT:    ret <8 x float> [[R]]
+;
+  %e = extractelement <8 x float> %x, i32 7
+  %n = fneg float %e
+  %r = insertelement <8 x float> %y, float %n, i32 7
+  ret <8 x float> %r
+}
+
+define <8 x float> @ext7_v8f32_use1(<8 x float> %x, <8 x float> %y) {
+; CHECK-LABEL: @ext7_v8f32_use1(
+; CHECK-NEXT:    [[E:%.*]] = extractelement <8 x float> [[X:%.*]], i32 5
+; CHECK-NEXT:    call void @use(float [[E]])
+; CHECK-NEXT:    [[N:%.*]] = fneg float [[E]]
+; CHECK-NEXT:    [[R:%.*]] = insertelement <8 x float> [[Y:%.*]], float [[N]], i32 5
+; CHECK-NEXT:    ret <8 x float> [[R]]
+;
+  %e = extractelement <8 x float> %x, i32 5
+  call void @use(float %e)
+  %n = fneg float %e
+  %r = insertelement <8 x float> %y, float %n, i32 5
+  ret <8 x float> %r
+}
+
+define <8 x float> @ext7_v8f32_use2(<8 x float> %x, <8 x float> %y) {
+; CHECK-LABEL: @ext7_v8f32_use2(
+; CHECK-NEXT:    [[E:%.*]] = extractelement <8 x float> [[X:%.*]], i32 3
+; CHECK-NEXT:    [[N:%.*]] = fneg float [[E]]
+; CHECK-NEXT:    call void @use(float [[N]])
+; CHECK-NEXT:    [[R:%.*]] = insertelement <8 x float> [[Y:%.*]], float [[N]], i32 3
+; CHECK-NEXT:    ret <8 x float> [[R]]
+;
+  %e = extractelement <8 x float> %x, i32 3
+  %n = fneg float %e
+  call void @use(float %n)
+  %r = insertelement <8 x float> %y, float %n, i32 3
+  ret <8 x float> %r
+}
+
+define <2 x double> @ext_index_var_v2f64(<2 x double> %x, <2 x double> %y, i32 %index) {
+; CHECK-LABEL: @ext_index_var_v2f64(
+; CHECK-NEXT:    [[E:%.*]] = extractelement <2 x double> [[X:%.*]], i32 [[INDEX:%.*]]
+; CHECK-NEXT:    [[N:%.*]] = fneg nsz double [[E]]
+; CHECK-NEXT:    [[R:%.*]] = insertelement <2 x double> [[Y:%.*]], double [[N]], i32 [[INDEX]]
+; CHECK-NEXT:    ret <2 x double> [[R]]
+;
+  %e = extractelement <2 x double> %x, i32 %index
+  %n = fneg nsz double %e
+  %r = insertelement <2 x double> %y, double %n, i32 %index
+  ret <2 x double> %r
+}
+
+define <2 x double> @ext1_v2f64_ins0(<2 x double> %x, <2 x double> %y) {
+; CHECK-LABEL: @ext1_v2f64_ins0(
+; CHECK-NEXT:    [[E:%.*]] = extractelement <2 x double> [[X:%.*]], i32 1
+; CHECK-NEXT:    [[N:%.*]] = fneg nsz double [[E]]
+; CHECK-NEXT:    [[R:%.*]] = insertelement <2 x double> [[Y:%.*]], double [[N]], i32 0
+; CHECK-NEXT:    ret <2 x double> [[R]]
+;
+  %e = extractelement <2 x double> %x, i32 1
+  %n = fneg nsz double %e
+  %r = insertelement <2 x double> %y, double %n, i32 0
+  ret <2 x double> %r
+}