Lower casting operations that involve vector types.

tests_lit/llvm2ice_tests/vector-bitcast.ll

+; This file tests bitcasts of vector type. For most operations, these
+; should be lowered to a no-op on -O2.
+
+; RUN: %llvm2ice -O2 --verbose none %s | FileCheck %s
+; RUN: %llvm2ice -Om1 --verbose none %s | FileCheck %s --check-prefix=OPTM1
+; RUN: %llvm2ice --verbose none %s | FileCheck --check-prefix=ERRORS %s
+; RUN: %llvm2iceinsts %s | %szdiff %s | FileCheck --check-prefix=DUMP %s
+; RUN: %llvm2iceinsts --pnacl %s | %szdiff %s \
+; RUN:                           | FileCheck --check-prefix=DUMP %s
+
+define <16 x i8> @test_bitcast_v16i8_to_v16i8(<16 x i8> %arg) {
+entry:
+  %res = bitcast <16 x i8> %arg to <16 x i8>
+  ret <16 x i8> %res
+
+; CHECK-LABEL: test_bitcast_v16i8_to_v16i8:
+; CHECK: .L{{.*}}entry:
+; CHECK-NEXT: ret
+}
+
+define <8 x i16> @test_bitcast_v16i8_to_v8i16(<16 x i8> %arg) {
+entry:
+  %res = bitcast <16 x i8> %arg to <8 x i16>
+  ret <8 x i16> %res
+
+; CHECK-LABEL: test_bitcast_v16i8_to_v8i16:
+; CHECK: .L{{.*}}entry:
+; CHECK-NEXT: ret
+}
+
+define <4 x i32> @test_bitcast_v16i8_to_v4i32(<16 x i8> %arg) {
+entry:
+  %res = bitcast <16 x i8> %arg to <4 x i32>
+  ret <4 x i32> %res
+
+; CHECK-LABEL: test_bitcast_v16i8_to_v4i32:
+; CHECK: .L{{.*}}entry:
+; CHECK-NEXT: ret
+}
+
+define <4 x float> @test_bitcast_v16i8_to_v4f32(<16 x i8> %arg) {
+entry:
+  %res = bitcast <16 x i8> %arg to <4 x float>
+  ret <4 x float> %res
+
+; CHECK-LABEL: test_bitcast_v16i8_to_v4f32:
+; CHECK: .L{{.*}}entry:
+; CHECK-NEXT: ret
+}
+
+define <16 x i8> @test_bitcast_v8i16_to_v16i8(<8 x i16> %arg) {
+entry:
+  %res = bitcast <8 x i16> %arg to <16 x i8>
+  ret <16 x i8> %res
+
+; CHECK-LABEL: test_bitcast_v8i16_to_v16i8:
+; CHECK: .L{{.*}}entry:
+; CHECK-NEXT: ret
+}
+
+define <8 x i16> @test_bitcast_v8i16_to_v8i16(<8 x i16> %arg) {
+entry:
+  %res = bitcast <8 x i16> %arg to <8 x i16>
+  ret <8 x i16> %res
+
+; CHECK-LABEL: test_bitcast_v8i16_to_v8i16:
+; CHECK: .L{{.*}}entry:
+; CHECK-NEXT: ret
+}
+
+define <4 x i32> @test_bitcast_v8i16_to_v4i32(<8 x i16> %arg) {
+entry:
+  %res = bitcast <8 x i16> %arg to <4 x i32>
+  ret <4 x i32> %res
+
+; CHECK-LABEL: test_bitcast_v8i16_to_v4i32:
+; CHECK: .L{{.*}}entry:
+; CHECK-NEXT: ret
+}
+
+define <4 x float> @test_bitcast_v8i16_to_v4f32(<8 x i16> %arg) {
+entry:
+  %res = bitcast <8 x i16> %arg to <4 x float>
+  ret <4 x float> %res
+
+; CHECK-LABEL: test_bitcast_v8i16_to_v4f32:
+; CHECK: .L{{.*}}entry:
+; CHECK-NEXT: ret
+}
+
+define <16 x i8> @test_bitcast_v4i32_to_v16i8(<4 x i32> %arg) {
+entry:
+  %res = bitcast <4 x i32> %arg to <16 x i8>
+  ret <16 x i8> %res
+
+; CHECK-LABEL: test_bitcast_v4i32_to_v16i8:
+; CHECK: .L{{.*}}entry:
+; CHECK-NEXT: ret
+}
+
+define <8 x i16> @test_bitcast_v4i32_to_v8i16(<4 x i32> %arg) {
+entry:
+  %res = bitcast <4 x i32> %arg to <8 x i16>
+  ret <8 x i16> %res
+
+; CHECK-LABEL: test_bitcast_v4i32_to_v8i16:
+; CHECK: .L{{.*}}entry:
+; CHECK-NEXT: ret
+}
+
+define <4 x i32> @test_bitcast_v4i32_to_v4i32(<4 x i32> %arg) {
+entry:
+  %res = bitcast <4 x i32> %arg to <4 x i32>
+  ret <4 x i32> %res
+
+; CHECK-LABEL: test_bitcast_v4i32_to_v4i32:
+; CHECK: .L{{.*}}entry:
+; CHECK-NEXT: ret
+}
+
+define <4 x float> @test_bitcast_v4i32_to_v4f32(<4 x i32> %arg) {
+entry:
+  %res = bitcast <4 x i32> %arg to <4 x float>
+  ret <4 x float> %res
+
+; CHECK-LABEL: test_bitcast_v4i32_to_v4f32:
+; CHECK: .L{{.*}}entry:
+; CHECK-NEXT: ret
+}
+
+define <16 x i8> @test_bitcast_v4f32_to_v16i8(<4 x float> %arg) {
+entry:
+  %res = bitcast <4 x float> %arg to <16 x i8>
+  ret <16 x i8> %res
+
+; CHECK-LABEL: test_bitcast_v4f32_to_v16i8:
+; CHECK: .L{{.*}}entry:
+; CHECK-NEXT: ret
+}
+
+define <8 x i16> @test_bitcast_v4f32_to_v8i16(<4 x float> %arg) {
+entry:
+  %res = bitcast <4 x float> %arg to <8 x i16>
+  ret <8 x i16> %res
+
+; CHECK-LABEL: test_bitcast_v4f32_to_v8i16:
+; CHECK: .L{{.*}}entry:
+; CHECK-NEXT: ret
+}
+
+define <4 x i32> @test_bitcast_v4f32_to_v4i32(<4 x float> %arg) {
+entry:
+  %res = bitcast <4 x float> %arg to <4 x i32>
+  ret <4 x i32> %res
+
+; CHECK-LABEL: test_bitcast_v4f32_to_v4i32:
+; CHECK: .L{{.*}}entry:
+; CHECK-NEXT: ret
+}
+
+define <4 x float> @test_bitcast_v4f32_to_v4f32(<4 x float> %arg) {
+entry:
+  %res = bitcast <4 x float> %arg to <4 x float>
+  ret <4 x float> %res
+
+; CHECK-LABEL: test_bitcast_v4f32_to_v4f32:
+; CHECK: .L{{.*}}entry:
+; CHECK-NEXT: ret
+}
+
+define i8 @test_bitcast_v8i1_to_i8(<8 x i1> %arg) {
+entry:
+  %res = bitcast <8 x i1> %arg to i8
+  ret i8 %res
+
+; CHECK-LABEL: test_bitcast_v8i1_to_i8:
+; CHECK: call Sz_bitcast_v8i1_to_i8
+
+; OPTM1-LABEL: test_bitcast_v8i1_to_i8:
+; OPMT1: call Sz_bitcast_v8i1_to_i8
+}
+
+define i16 @test_bitcast_v16i1_to_i16(<16 x i1> %arg) {
+entry:
+  %res = bitcast <16 x i1> %arg to i16
+  ret i16 %res
+
+; CHECK-LABEL: test_bitcast_v16i1_to_i16:
+; CHECK: call Sz_bitcast_v16i1_to_i16

[jvoung (off chromium), 2014/07/16 00:01:38]

Could add the OPTM1 check here too?

[wala, 2014/07/16 17:19:01]

Done.

+}
+
+define <8 x i1> @test_bitcast_i8_to_v8i1(i8 %arg) {

[jvoung (off chromium), 2014/07/16 00:01:38]

Technically, the PNaCl bitcode ABI doesn't allow arguments and return values to
be smaller than i32.

This mostly causes a problem on a function call (push), but it might be good to
make this test conform to that ABI.

[wala, 2014/07/16 17:19:01]

Done.

+entry:
+  %res = bitcast i8 %arg to <8 x i1>
+  ret <8 x i1> %res
+
+; CHECK-LABEL: test_bitcast_i8_to_v8i1:
+; CHECK: call Sz_bitcast_i8_to_v8i1
+
+; OPTM1-LABEL: test_bitcast_i8_to_v8i1:
+; OPTM1: call Sz_bitcast_i8_to_v8i1
+}
+
+define <16 x i1> @test_bitcast_i16_to_v16i1(i16 %arg) {
+entry:
+  %res = bitcast i16 %arg to <16 x i1>
+  ret <16 x i1> %res
+
+; CHECK-LABEL: test_bitcast_i16_to_v16i1:
+; CHECK: call Sz_bitcast_i16_to_v16i1
+
+; OPTM1-LABEL: test_bitcast_i16_to_v16i1:
+; OPTM1: call Sz_bitcast_i16_to_v16i1
+}
+
+; ERRORS-NOT: ICE translation error
+; DUMP-NOT: SZ