[SLP][NFC]Add a test with external cast and extracted operand, NFC

Author: alexey-bataev

llvm/test/Transforms/SLPVectorizer/X86/cast-operand-extracted.ll

@@ -0,0 +1,79 @@
+; NOTE: Assertions have been autogenerated by utils/update_test_checks.py UTC_ARGS: --version 5
+; RUN: opt -S --passes=slp-vectorizer -mtriple=x86_64-unknown-linux-gnu -mcpu=cascadelake < %s | FileCheck %s
+
+define void @test(ptr %0, i32 %add651) {
+; CHECK-LABEL: define void @test(
+; CHECK-SAME: ptr [[TMP0:%.*]], i32 [[ADD651:%.*]]) #[[ATTR0:[0-9]+]] {
+; CHECK-NEXT:  [[ENTRY:.*:]]
+; CHECK-NEXT:    [[PREDPEL11:%.*]] = alloca [0 x [0 x [25 x i32]]], i32 0, align 16
+; CHECK-NEXT:    [[ARRAYIDX469_6:%.*]] = getelementptr i8, ptr [[PREDPEL11]], i64 28
+; CHECK-NEXT:    [[ARRAYIDX469_7:%.*]] = getelementptr i8, ptr [[PREDPEL11]], i64 32
+; CHECK-NEXT:    [[TMP1:%.*]] = getelementptr i8, ptr [[PREDPEL11]], i64 36
+; CHECK-NEXT:    [[TMP2:%.*]] = load i32, ptr [[ARRAYIDX469_7]], align 16
+; CHECK-NEXT:    [[TMP3:%.*]] = load <2 x i32>, ptr [[ARRAYIDX469_6]], align 4
+; CHECK-NEXT:    [[CONV470_7:%.*]] = trunc i32 [[TMP2]] to i16
+; CHECK-NEXT:    store i16 [[CONV470_7]], ptr [[TMP0]], align 2
+; CHECK-NEXT:    [[TMP4:%.*]] = load ptr, ptr [[TMP0]], align 8
+; CHECK-NEXT:    [[ARRAYIDX660:%.*]] = getelementptr i8, ptr [[TMP4]], i64 7800
+; CHECK-NEXT:    [[ARRAYIDX689:%.*]] = getelementptr i8, ptr [[TMP4]], i64 7816
+; CHECK-NEXT:    [[TMP5:%.*]] = load <2 x i32>, ptr [[TMP1]], align 4
+; CHECK-NEXT:    [[TMP6:%.*]] = add <2 x i32> [[TMP3]], <i32 1, i32 1>
+; CHECK-NEXT:    [[TMP7:%.*]] = shufflevector <2 x i32> [[TMP3]], <2 x i32> [[TMP5]], <2 x i32> <i32 1, i32 2>
+; CHECK-NEXT:    [[TMP8:%.*]] = add <2 x i32> [[TMP6]], [[TMP7]]
+; CHECK-NEXT:    [[TMP9:%.*]] = shufflevector <2 x i32> [[TMP5]], <2 x i32> <i32 1, i32 poison>, <2 x i32> <i32 2, i32 1>
+; CHECK-NEXT:    [[TMP10:%.*]] = add <2 x i32> [[TMP8]], [[TMP9]]
+; CHECK-NEXT:    [[TMP11:%.*]] = insertelement <4 x i32> poison, i32 [[ADD651]], i32 0
+; CHECK-NEXT:    [[TMP12:%.*]] = shufflevector <2 x i32> [[TMP3]], <2 x i32> poison, <4 x i32> <i32 0, i32 1, i32 poison, i32 poison>
+; CHECK-NEXT:    [[TMP13:%.*]] = shufflevector <4 x i32> [[TMP11]], <4 x i32> [[TMP12]], <4 x i32> <i32 0, i32 5, i32 poison, i32 poison>
+; CHECK-NEXT:    [[TMP14:%.*]] = call <4 x i32> @llvm.vector.insert.v4i32.v2i32(<4 x i32> [[TMP13]], <2 x i32> [[TMP10]], i64 2)
+; CHECK-NEXT:    [[TMP15:%.*]] = lshr <4 x i32> [[TMP14]], <i32 1, i32 1, i32 1, i32 1>
+; CHECK-NEXT:    [[TMP16:%.*]] = trunc <4 x i32> [[TMP15]] to <4 x i16>
+; CHECK-NEXT:    [[TMP17:%.*]] = shufflevector <4 x i16> [[TMP16]], <4 x i16> poison, <2 x i32> <i32 1, i32 2>
+; CHECK-NEXT:    store <2 x i16> [[TMP17]], ptr [[ARRAYIDX689]], align 8
+; CHECK-NEXT:    [[TMP18:%.*]] = extractelement <4 x i16> [[TMP16]], i32 3
+; CHECK-NEXT:    store i16 [[TMP18]], ptr [[TMP4]], align 8
+; CHECK-NEXT:    store <4 x i16> [[TMP16]], ptr [[ARRAYIDX660]], align 8
+; CHECK-NEXT:    ret void
+;
+entry:
+  %PredPel11 = alloca [0 x [0 x [25 x i32]]], i32 0, align 16
+  %arrayidx469.6 = getelementptr i8, ptr %PredPel11, i64 28
+  %1 = load i32, ptr %arrayidx469.6, align 4
+  %arrayidx469.7 = getelementptr i8, ptr %PredPel11, i64 32
+  %2 = load i32, ptr %arrayidx469.7, align 16
+  %conv470.7 = trunc i32 %2 to i16
+  store i16 %conv470.7, ptr %0, align 2
+  %3 = getelementptr i8, ptr %PredPel11, i64 36
+  %4 = getelementptr i8, ptr %PredPel11, i64 40
+  %5 = load ptr, ptr %0, align 8
+  %add6511 = add i32 %1, 1
+  %shr656 = lshr i32 %add651, 1
+  %conv657 = trunc i32 %shr656 to i16
+  %arrayidx660 = getelementptr i8, ptr %5, i64 7800
+  store i16 %conv657, ptr %arrayidx660, align 8
+  %shr685 = lshr i32 %2, 1
+  %conv686 = trunc i32 %shr685 to i16
+  %arrayidx689 = getelementptr i8, ptr %5, i64 7816
+  store i16 %conv686, ptr %arrayidx689, align 8
+  %arrayidx694 = getelementptr i8, ptr %5, i64 7802
+  store i16 %conv686, ptr %arrayidx694, align 2
+  %6 = load i32, ptr %3, align 4
+  %add716 = add i32 %add6511, %2
+  %add717 = add i32 %add716, 1
+  %shr718 = lshr i32 %add717, 1
+  %conv719 = trunc i32 %shr718 to i16
+  %arrayidx727 = getelementptr i8, ptr %5, i64 7818
+  store i16 %conv719, ptr %arrayidx727, align 2
+  %arrayidx731 = getelementptr i8, ptr %5, i64 7804
+  store i16 %conv719, ptr %arrayidx731, align 4
+  %7 = load i32, ptr %4, align 8
+  %add750 = add i32 %2, 1
+  %add753 = add i32 %add750, %6
+  %add754 = add i32 %add753, %7
+  %shr755 = lshr i32 %add754, 1
+  %conv756 = trunc i32 %shr755 to i16
+  store i16 %conv756, ptr %5, align 8
+  %arrayidx772 = getelementptr i8, ptr %5, i64 7806
+  store i16 %conv756, ptr %arrayidx772, align 2
+  ret void
+}