[VectorCombine][X86] Add some tests showing failure to fold shuffle(cast(x),cast(y)) -> cast(shuffle(x,y))

Part of #67803

1 files changed, 189 insertions, 0 deletions

diff --git a/llvm/test/Transforms/VectorCombine/X86/shuffle-of-casts.ll b/llvm/test/Transforms/VectorCombine/X86/shuffle-of-casts.ll
new file mode 100644
index 0000000..3a7c331
--- /dev/null
+++ b/llvm/test/Transforms/VectorCombine/X86/shuffle-of-casts.ll
@@ -0,0 +1,189 @@
+; 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
+
+; standard vector concatenations
+
+define <16 x i32> @concat_zext_v8i16_v16i32(<8 x i16> %a0, <8 x i16> %a1) {
+; CHECK-LABEL: @concat_zext_v8i16_v16i32(
+; CHECK-NEXT:    [[X0:%.*]] = zext <8 x i16> [[A0:%.*]] to <8 x i32>
+; CHECK-NEXT:    [[X1:%.*]] = zext <8 x i16> [[A1:%.*]] to <8 x i32>
+; CHECK-NEXT:    [[R:%.*]] = shufflevector <8 x i32> [[X0]], <8 x i32> [[X1]], <16 x i32> <i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7, i32 8, i32 9, i32 10, i32 11, i32 12, i32 13, i32 14, i32 15>
+; CHECK-NEXT:    ret <16 x i32> [[R]]
+;
+  %x0 = zext <8 x i16> %a0 to <8 x i32>
+  %x1 = zext <8 x i16> %a1 to <8 x i32>
+  %r = shufflevector <8 x i32> %x0, <8 x i32> %x1, <16 x i32> <i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7, i32 8, i32 9, i32 10, i32 11, i32 12, i32 13, i32 14, i32 15>
+  ret <16 x i32> %r
+}
+
+define <16 x i32> @concat_sext_v8i16_v16i32(<8 x i16> %a0, <8 x i16> %a1) {
+; CHECK-LABEL: @concat_sext_v8i16_v16i32(
+; CHECK-NEXT:    [[X0:%.*]] = sext <8 x i16> [[A0:%.*]] to <8 x i32>
+; CHECK-NEXT:    [[X1:%.*]] = sext <8 x i16> [[A1:%.*]] to <8 x i32>
+; CHECK-NEXT:    [[R:%.*]] = shufflevector <8 x i32> [[X0]], <8 x i32> [[X1]], <16 x i32> <i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7, i32 8, i32 9, i32 10, i32 11, i32 12, i32 13, i32 14, i32 15>
+; CHECK-NEXT:    ret <16 x i32> [[R]]
+;
+  %x0 = sext <8 x i16> %a0 to <8 x i32>
+  %x1 = sext <8 x i16> %a1 to <8 x i32>
+  %r = shufflevector <8 x i32> %x0, <8 x i32> %x1, <16 x i32> <i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7, i32 8, i32 9, i32 10, i32 11, i32 12, i32 13, i32 14, i32 15>
+  ret <16 x i32> %r
+}
+
+define <8 x i32> @concat_sext_v4i1_v8i32(<4 x i1> %a0, <4 x i1> %a1) {
+; CHECK-LABEL: @concat_sext_v4i1_v8i32(
+; CHECK-NEXT:    [[X0:%.*]] = sext <4 x i1> [[A0:%.*]] to <4 x i32>
+; CHECK-NEXT:    [[X1:%.*]] = sext <4 x i1> [[A1:%.*]] to <4 x i32>
+; CHECK-NEXT:    [[R:%.*]] = shufflevector <4 x i32> [[X0]], <4 x i32> [[X1]], <8 x i32> <i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7>
+; CHECK-NEXT:    ret <8 x i32> [[R]]
+;
+  %x0 = sext <4 x i1> %a0 to <4 x i32>
+  %x1 = sext <4 x i1> %a1 to <4 x i32>
+  %r = shufflevector <4 x i32> %x0, <4 x i32> %x1, <8 x i32> <i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7>
+  ret <8 x i32> %r
+}
+
+define <8 x i16> @concat_trunc_v4i32_v8i16(<4 x i32> %a0, <4 x i32> %a1) {
+; CHECK-LABEL: @concat_trunc_v4i32_v8i16(
+; CHECK-NEXT:    [[X0:%.*]] = trunc <4 x i32> [[A0:%.*]] to <4 x i16>
+; CHECK-NEXT:    [[X1:%.*]] = trunc <4 x i32> [[A1:%.*]] to <4 x i16>
+; CHECK-NEXT:    [[R:%.*]] = shufflevector <4 x i16> [[X0]], <4 x i16> [[X1]], <8 x i32> <i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7>
+; CHECK-NEXT:    ret <8 x i16> [[R]]
+;
+  %x0 = trunc <4 x i32> %a0 to <4 x i16>
+  %x1 = trunc <4 x i32> %a1 to <4 x i16>
+  %r = shufflevector <4 x i16> %x0, <4 x i16> %x1, <8 x i32> <i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7>
+  ret <8 x i16> %r
+}
+
+define <8 x ptr> @concat_inttoptr_v4i32_v8iptr(<4 x i32> %a0, <4 x i32> %a1) {
+; CHECK-LABEL: @concat_inttoptr_v4i32_v8iptr(
+; CHECK-NEXT:    [[X0:%.*]] = inttoptr <4 x i32> [[A0:%.*]] to <4 x ptr>
+; CHECK-NEXT:    [[X1:%.*]] = inttoptr <4 x i32> [[A1:%.*]] to <4 x ptr>
+; CHECK-NEXT:    [[R:%.*]] = shufflevector <4 x ptr> [[X0]], <4 x ptr> [[X1]], <8 x i32> <i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7>
+; CHECK-NEXT:    ret <8 x ptr> [[R]]
+;
+  %x0 = inttoptr <4 x i32> %a0 to <4 x ptr>
+  %x1 = inttoptr <4 x i32> %a1 to <4 x ptr>
+  %r = shufflevector <4 x ptr> %x0, <4 x ptr> %x1, <8 x i32> <i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7>
+  ret <8 x ptr> %r
+}
+
+define <16 x i64> @concat_ptrtoint_v8i16_v16i32(<8 x ptr> %a0, <8 x ptr> %a1) {
+; CHECK-LABEL: @concat_ptrtoint_v8i16_v16i32(
+; CHECK-NEXT:    [[X0:%.*]] = ptrtoint <8 x ptr> [[A0:%.*]] to <8 x i64>
+; CHECK-NEXT:    [[X1:%.*]] = ptrtoint <8 x ptr> [[A1:%.*]] to <8 x i64>
+; CHECK-NEXT:    [[R:%.*]] = shufflevector <8 x i64> [[X0]], <8 x i64> [[X1]], <16 x i32> <i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7, i32 8, i32 9, i32 10, i32 11, i32 12, i32 13, i32 14, i32 15>
+; CHECK-NEXT:    ret <16 x i64> [[R]]
+;
+  %x0 = ptrtoint <8 x ptr> %a0 to <8 x i64>
+  %x1 = ptrtoint <8 x ptr> %a1 to <8 x i64>
+  %r = shufflevector <8 x i64> %x0, <8 x i64> %x1, <16 x i32> <i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7, i32 8, i32 9, i32 10, i32 11, i32 12, i32 13, i32 14, i32 15>
+  ret <16 x i64> %r
+}
+
+define <8 x double> @concat_fpext_v4f32_v8f64(<4 x float> %a0, <4 x float> %a1) {
+; CHECK-LABEL: @concat_fpext_v4f32_v8f64(
+; CHECK-NEXT:    [[X0:%.*]] = fpext <4 x float> [[A0:%.*]] to <4 x double>
+; CHECK-NEXT:    [[X1:%.*]] = fpext <4 x float> [[A1:%.*]] to <4 x double>
+; CHECK-NEXT:    [[R:%.*]] = shufflevector <4 x double> [[X0]], <4 x double> [[X1]], <8 x i32> <i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7>
+; CHECK-NEXT:    ret <8 x double> [[R]]
+;
+  %x0 = fpext <4 x float> %a0 to <4 x double>
+  %x1 = fpext <4 x float> %a1 to <4 x double>
+  %r = shufflevector <4 x double> %x0, <4 x double> %x1, <8 x i32> <i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7>
+  ret <8 x double> %r
+}
+
+define <16 x float> @concat_fptrunc_v8f64_v16f32(<8 x double> %a0, <8 x double> %a1) {
+; CHECK-LABEL: @concat_fptrunc_v8f64_v16f32(
+; CHECK-NEXT:    [[X0:%.*]] = fptrunc <8 x double> [[A0:%.*]] to <8 x float>
+; CHECK-NEXT:    [[X1:%.*]] = fptrunc <8 x double> [[A1:%.*]] to <8 x float>
+; CHECK-NEXT:    [[R:%.*]] = shufflevector <8 x float> [[X0]], <8 x float> [[X1]], <16 x i32> <i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7, i32 8, i32 9, i32 10, i32 11, i32 12, i32 13, i32 14, i32 15>
+; CHECK-NEXT:    ret <16 x float> [[R]]
+;
+  %x0 = fptrunc <8 x double> %a0 to <8 x float>
+  %x1 = fptrunc <8 x double> %a1 to <8 x float>
+  %r = shufflevector <8 x float> %x0, <8 x float> %x1, <16 x i32> <i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7, i32 8, i32 9, i32 10, i32 11, i32 12, i32 13, i32 14, i32 15>
+  ret <16 x float> %r
+}
+
+; commuted vector concatenation
+
+define <16 x i32> @rconcat_sext_v8i16_v16i32(<8 x i16> %a0, <8 x i16> %a1) {
+; CHECK-LABEL: @rconcat_sext_v8i16_v16i32(
+; CHECK-NEXT:    [[X0:%.*]] = sext <8 x i16> [[A0:%.*]] to <8 x i32>
+; CHECK-NEXT:    [[X1:%.*]] = sext <8 x i16> [[A1:%.*]] to <8 x i32>
+; CHECK-NEXT:    [[R:%.*]] = shufflevector <8 x i32> [[X0]], <8 x i32> [[X1]], <16 x i32> <i32 8, i32 9, i32 10, i32 11, i32 12, i32 13, i32 14, i32 15, i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7>
+; CHECK-NEXT:    ret <16 x i32> [[R]]
+;
+  %x0 = sext <8 x i16> %a0 to <8 x i32>
+  %x1 = sext <8 x i16> %a1 to <8 x i32>
+  %r = shufflevector <8 x i32> %x0, <8 x i32> %x1, <16 x i32> <i32 8, i32 9, i32 10, i32 11, i32 12, i32 13, i32 14, i32 15, i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7>
+  ret <16 x i32> %r
+}
+
+; interleaved shuffle
+
+define <8 x double> @interleave_fpext_v4f32_v8f64(<4 x float> %a0, <4 x float> %a1) {
+; CHECK-LABEL: @interleave_fpext_v4f32_v8f64(
+; CHECK-NEXT:    [[X0:%.*]] = fpext <4 x float> [[A0:%.*]] to <4 x double>
+; CHECK-NEXT:    [[X1:%.*]] = fpext <4 x float> [[A1:%.*]] to <4 x double>
+; CHECK-NEXT:    [[R:%.*]] = shufflevector <4 x double> [[X0]], <4 x double> [[X1]], <8 x i32> <i32 0, i32 4, i32 1, i32 5, i32 2, i32 6, i32 3, i32 7>
+; CHECK-NEXT:    ret <8 x double> [[R]]
+;
+  %x0 = fpext <4 x float> %a0 to <4 x double>
+  %x1 = fpext <4 x float> %a1 to <4 x double>
+  %r = shufflevector <4 x double> %x0, <4 x double> %x1, <8 x i32> <i32 0, i32 4, i32 1, i32 5, i32 2, i32 6, i32 3, i32 7>
+  ret <8 x double> %r
+}
+
+; negative - multiuse
+
+define <8 x i16> @concat_trunc_v4i32_v8i16_multiuse(<4 x i32> %a0, <4 x i32> %a1, ptr %a2) {
+; CHECK-LABEL: @concat_trunc_v4i32_v8i16_multiuse(
+; CHECK-NEXT:    [[X0:%.*]] = trunc <4 x i32> [[A0:%.*]] to <4 x i16>
+; CHECK-NEXT:    [[X1:%.*]] = trunc <4 x i32> [[A1:%.*]] to <4 x i16>
+; CHECK-NEXT:    [[R:%.*]] = shufflevector <4 x i16> [[X0]], <4 x i16> [[X1]], <8 x i32> <i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7>
+; CHECK-NEXT:    store <4 x i16> [[X0]], ptr [[A2:%.*]], align 8
+; CHECK-NEXT:    ret <8 x i16> [[R]]
+;
+  %x0 = trunc <4 x i32> %a0 to <4 x i16>
+  %x1 = trunc <4 x i32> %a1 to <4 x i16>
+  %r = shufflevector <4 x i16> %x0, <4 x i16> %x1, <8 x i32> <i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7>
+  store <4 x i16> %x0, ptr %a2
+  ret <8 x i16> %r
+}
+
+; negative - bitcasts
+
+define <8 x float> @concat_bitcast_v4i32_v8f32(<4 x i32> %a0, <4 x i32> %a1) {
+; CHECK-LABEL: @concat_bitcast_v4i32_v8f32(
+; CHECK-NEXT:    [[X0:%.*]] = bitcast <4 x i32> [[A0:%.*]] to <4 x float>
+; CHECK-NEXT:    [[X1:%.*]] = bitcast <4 x i32> [[A1:%.*]] to <4 x float>
+; CHECK-NEXT:    [[R:%.*]] = shufflevector <4 x float> [[X0]], <4 x float> [[X1]], <8 x i32> <i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7>
+; CHECK-NEXT:    ret <8 x float> [[R]]
+;
+  %x0 = bitcast <4 x i32> %a0 to <4 x float>
+  %x1 = bitcast <4 x i32> %a1 to <4 x float>
+  %r = shufflevector <4 x float> %x0, <4 x float> %x1, <8 x i32> <i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7>
+  ret <8 x float> %r
+}
+
+; negative - castop mismatch
+
+define <16 x i32> @concat_sext_zext_v8i16_v16i32(<8 x i16> %a0, <8 x i16> %a1) {
+; CHECK-LABEL: @concat_sext_zext_v8i16_v16i32(
+; CHECK-NEXT:    [[X0:%.*]] = sext <8 x i16> [[A0:%.*]] to <8 x i32>
+; CHECK-NEXT:    [[X1:%.*]] = zext <8 x i16> [[A1:%.*]] to <8 x i32>
+; CHECK-NEXT:    [[R:%.*]] = shufflevector <8 x i32> [[X0]], <8 x i32> [[X1]], <16 x i32> <i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7, i32 8, i32 9, i32 10, i32 11, i32 12, i32 13, i32 14, i32 15>
+; CHECK-NEXT:    ret <16 x i32> [[R]]
+;
+  %x0 = sext <8 x i16> %a0 to <8 x i32>
+  %x1 = zext <8 x i16> %a1 to <8 x i32>
+  %r = shufflevector <8 x i32> %x0, <8 x i32> %x1, <16 x i32> <i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7, i32 8, i32 9, i32 10, i32 11, i32 12, i32 13, i32 14, i32 15>
+  ret <16 x i32> %r
+}
+;; NOTE: These prefixes are unused and the list is autogenerated. Do not add tests below this line:
+; AVX: {{.*}}
+; SSE: {{.*}}