[SLP][X86] Extend target coverage for PR47629

As suggested on D90445, add tests for various SSE/AVX levels and more complex gep pointer offsets

diff --git a/llvm/test/Transforms/SLPVectorizer/X86/pr47629.ll b/llvm/test/Transforms/SLPVectorizer/X86/pr47629.ll
index 024e2fb..2d910f0 100644 (file)
--- a/llvm/test/Transforms/SLPVectorizer/X86/pr47629.ll
+++ b/llvm/test/Transforms/SLPVectorizer/X86/pr47629.ll
@@ -1,5 +1,8 @@
 ; NOTE: Assertions have been autogenerated by utils/update_test_checks.py
-; RUN:  opt < %s -slp-vectorizer -instcombine -S -mtriple=x86_64-unknown-linux -march=avx512 | FileCheck %s
+; RUN:  opt < %s -slp-vectorizer -instcombine -S -mtriple=x86_64-unknown-linux -mattr=+sse2    | FileCheck %s --check-prefixes=CHECK,SSE
+; RUN:  opt < %s -slp-vectorizer -instcombine -S -mtriple=x86_64-unknown-linux -mattr=+avx     | FileCheck %s --check-prefixes=CHECK,AVX,AVX1
+; RUN:  opt < %s -slp-vectorizer -instcombine -S -mtriple=x86_64-unknown-linux -mattr=+avx2    | FileCheck %s --check-prefixes=CHECK,AVX,AVX2
+; RUN:  opt < %s -slp-vectorizer -instcombine -S -mtriple=x86_64-unknown-linux -mattr=+avx512f | FileCheck %s --check-prefixes=CHECK,AVX,AVX512F
 
 define void @gather_load(i32* %0, i32* readonly %1) {
 ; CHECK-LABEL: @gather_load(
@@ -38,3 +41,48 @@ define void @gather_load(i32* %0, i32* readonly %1) {
   store <4 x i32> %17, <4 x i32>* %18, align 4
   ret void
 }
+
+define void @gather_load_2(i32* %0, i32* readonly %1) {
+; CHECK-LABEL: @gather_load_2(
+; CHECK-NEXT:    [[TMP3:%.*]] = getelementptr inbounds i32, i32* [[TMP1:%.*]], i64 1
+; CHECK-NEXT:    [[TMP4:%.*]] = load i32, i32* [[TMP3]], align 4
+; CHECK-NEXT:    [[TMP5:%.*]] = add nsw i32 [[TMP4]], 1
+; CHECK-NEXT:    [[TMP6:%.*]] = getelementptr inbounds i32, i32* [[TMP0:%.*]], i64 1
+; CHECK-NEXT:    store i32 [[TMP5]], i32* [[TMP0]], align 4
+; CHECK-NEXT:    [[TMP7:%.*]] = getelementptr inbounds i32, i32* [[TMP1]], i64 10
+; CHECK-NEXT:    [[TMP8:%.*]] = load i32, i32* [[TMP7]], align 4
+; CHECK-NEXT:    [[TMP9:%.*]] = add nsw i32 [[TMP8]], 2
+; CHECK-NEXT:    [[TMP10:%.*]] = getelementptr inbounds i32, i32* [[TMP0]], i64 2
+; CHECK-NEXT:    store i32 [[TMP9]], i32* [[TMP6]], align 4
+; CHECK-NEXT:    [[TMP11:%.*]] = getelementptr inbounds i32, i32* [[TMP1]], i64 3
+; CHECK-NEXT:    [[TMP12:%.*]] = load i32, i32* [[TMP11]], align 4
+; CHECK-NEXT:    [[TMP13:%.*]] = add nsw i32 [[TMP12]], 3
+; CHECK-NEXT:    [[TMP14:%.*]] = getelementptr inbounds i32, i32* [[TMP0]], i64 3
+; CHECK-NEXT:    store i32 [[TMP13]], i32* [[TMP10]], align 4
+; CHECK-NEXT:    [[TMP15:%.*]] = getelementptr inbounds i32, i32* [[TMP1]], i64 5
+; CHECK-NEXT:    [[TMP16:%.*]] = load i32, i32* [[TMP15]], align 4
+; CHECK-NEXT:    [[TMP17:%.*]] = add nsw i32 [[TMP16]], 4
+; CHECK-NEXT:    store i32 [[TMP17]], i32* [[TMP14]], align 4
+; CHECK-NEXT:    ret void
+;
+  %3 = getelementptr inbounds i32, i32* %1, i64 1
+  %4 = load i32, i32* %3, align 4
+  %5 = add nsw i32 %4, 1
+  %6 = getelementptr inbounds i32, i32* %0, i64 1
+  store i32 %5, i32* %0, align 4
+  %7 = getelementptr inbounds i32, i32* %1, i64 10
+  %8 = load i32, i32* %7, align 4
+  %9 = add nsw i32 %8, 2
+  %10 = getelementptr inbounds i32, i32* %0, i64 2
+  store i32 %9, i32* %6, align 4
+  %11 = getelementptr inbounds i32, i32* %1, i64 3
+  %12 = load i32, i32* %11, align 4
+  %13 = add nsw i32 %12, 3
+  %14 = getelementptr inbounds i32, i32* %0, i64 3
+  store i32 %13, i32* %10, align 4
+  %15 = getelementptr inbounds i32, i32* %1, i64 5
+  %16 = load i32, i32* %15, align 4
+  %17 = add nsw i32 %16, 4
+  store i32 %17, i32* %14, align 4
+  ret void
+}