Skip to content

[PowerPC][NFC] autogen mma tc checks via update_cc_test_checks #108584

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Merged
merged 2 commits into from
Sep 13, 2024
Merged

[PowerPC][NFC] autogen mma tc checks via update_cc_test_checks #108584

Show file tree
Hide file tree
Changes from all commits
Commits
Show all changes
2 commits
Select commit Hold shift + click to select a range
70077d3
[PowerPC][NFC] autogen mma tc checks via llvm/utils/update_cc_test_ch…
lei137 Sep 13, 2024
1856d90
add non-opt checks for mma build pair test
lei137 Sep 13, 2024
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
68 changes: 68 additions & 0 deletions clang/test/CodeGen/PowerPC/builtins-ppc-build-pair-mma.c
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -3,6 +3,8 @@
// RUN: -emit-llvm %s -o - | FileCheck %s --check-prefix=CHECK-LE
// RUN: %clang_cc1 -O3 -triple powerpc64-unknown-unknown -target-cpu pwr10 \
// RUN: -emit-llvm %s -o - | FileCheck %s --check-prefix=CHECK-BE
// RUN: %clang_cc1 -O0 -triple powerpc64le-unknown-unknown -target-cpu pwr10 \
// RUN: -emit-llvm %s -o - | FileCheck %s --check-prefix=CHECK-LE-NOOPT

// CHECK-LE-LABEL: @test1(
// CHECK-LE-NEXT: entry:
Expand All @@ -16,6 +18,42 @@
// CHECK-BE-NEXT: store <512 x i1> [[TMP0]], ptr [[RESP:%.*]], align 64, !tbaa [[TBAA2:![0-9]+]]
// CHECK-BE-NEXT: ret void
//
// CHECK-LE-NOOPT-LABEL: @test1(
// CHECK-LE-NOOPT-NEXT: entry:
// CHECK-LE-NOOPT-NEXT: [[VQP_ADDR:%.*]] = alloca ptr, align 8
// CHECK-LE-NOOPT-NEXT: [[VPP_ADDR:%.*]] = alloca ptr, align 8
// CHECK-LE-NOOPT-NEXT: [[VC1_ADDR:%.*]] = alloca <16 x i8>, align 16
// CHECK-LE-NOOPT-NEXT: [[VC2_ADDR:%.*]] = alloca <16 x i8>, align 16
// CHECK-LE-NOOPT-NEXT: [[VC3_ADDR:%.*]] = alloca <16 x i8>, align 16
// CHECK-LE-NOOPT-NEXT: [[VC4_ADDR:%.*]] = alloca <16 x i8>, align 16
// CHECK-LE-NOOPT-NEXT: [[RESP_ADDR:%.*]] = alloca ptr, align 8
// CHECK-LE-NOOPT-NEXT: [[VQ:%.*]] = alloca <512 x i1>, align 64
// CHECK-LE-NOOPT-NEXT: [[VP:%.*]] = alloca <256 x i1>, align 32
// CHECK-LE-NOOPT-NEXT: [[RES:%.*]] = alloca <512 x i1>, align 64
// CHECK-LE-NOOPT-NEXT: store ptr [[VQP:%.*]], ptr [[VQP_ADDR]], align 8
// CHECK-LE-NOOPT-NEXT: store ptr [[VPP:%.*]], ptr [[VPP_ADDR]], align 8
// CHECK-LE-NOOPT-NEXT: store <16 x i8> [[VC1:%.*]], ptr [[VC1_ADDR]], align 16
// CHECK-LE-NOOPT-NEXT: store <16 x i8> [[VC2:%.*]], ptr [[VC2_ADDR]], align 16
// CHECK-LE-NOOPT-NEXT: store <16 x i8> [[VC3:%.*]], ptr [[VC3_ADDR]], align 16
// CHECK-LE-NOOPT-NEXT: store <16 x i8> [[VC4:%.*]], ptr [[VC4_ADDR]], align 16
// CHECK-LE-NOOPT-NEXT: store ptr [[RESP:%.*]], ptr [[RESP_ADDR]], align 8
// CHECK-LE-NOOPT-NEXT: [[TMP0:%.*]] = load ptr, ptr [[VQP_ADDR]], align 8
// CHECK-LE-NOOPT-NEXT: [[TMP1:%.*]] = load <512 x i1>, ptr [[TMP0]], align 64
// CHECK-LE-NOOPT-NEXT: store <512 x i1> [[TMP1]], ptr [[VQ]], align 64
// CHECK-LE-NOOPT-NEXT: [[TMP2:%.*]] = load ptr, ptr [[VPP_ADDR]], align 8
// CHECK-LE-NOOPT-NEXT: [[TMP3:%.*]] = load <256 x i1>, ptr [[TMP2]], align 32
// CHECK-LE-NOOPT-NEXT: store <256 x i1> [[TMP3]], ptr [[VP]], align 32
// CHECK-LE-NOOPT-NEXT: [[TMP4:%.*]] = load <16 x i8>, ptr [[VC1_ADDR]], align 16
// CHECK-LE-NOOPT-NEXT: [[TMP5:%.*]] = load <16 x i8>, ptr [[VC2_ADDR]], align 16
// CHECK-LE-NOOPT-NEXT: [[TMP6:%.*]] = load <16 x i8>, ptr [[VC3_ADDR]], align 16
// CHECK-LE-NOOPT-NEXT: [[TMP7:%.*]] = load <16 x i8>, ptr [[VC4_ADDR]], align 16
// CHECK-LE-NOOPT-NEXT: [[TMP8:%.*]] = call <512 x i1> @llvm.ppc.mma.assemble.acc(<16 x i8> [[TMP7]], <16 x i8> [[TMP6]], <16 x i8> [[TMP5]], <16 x i8> [[TMP4]])
// CHECK-LE-NOOPT-NEXT: store <512 x i1> [[TMP8]], ptr [[RES]], align 64
// CHECK-LE-NOOPT-NEXT: [[TMP9:%.*]] = load <512 x i1>, ptr [[RES]], align 64
// CHECK-LE-NOOPT-NEXT: [[TMP10:%.*]] = load ptr, ptr [[RESP_ADDR]], align 8
// CHECK-LE-NOOPT-NEXT: store <512 x i1> [[TMP9]], ptr [[TMP10]], align 64
// CHECK-LE-NOOPT-NEXT: ret void
//
void test1(unsigned char *vqp, unsigned char *vpp, vector unsigned char vc1, vector unsigned char vc2,
vector unsigned char vc3, vector unsigned char vc4, unsigned char *resp) {
__vector_quad vq = *((__vector_quad *)vqp);
Expand All @@ -37,6 +75,36 @@ void test1(unsigned char *vqp, unsigned char *vpp, vector unsigned char vc1, vec
// CHECK-BE-NEXT: store <256 x i1> [[TMP0]], ptr [[RESP:%.*]], align 32, !tbaa [[TBAA6:![0-9]+]]
// CHECK-BE-NEXT: ret void
//
// CHECK-LE-NOOPT-LABEL: @test2(
// CHECK-LE-NOOPT-NEXT: entry:
// CHECK-LE-NOOPT-NEXT: [[VQP_ADDR:%.*]] = alloca ptr, align 8
// CHECK-LE-NOOPT-NEXT: [[VPP_ADDR:%.*]] = alloca ptr, align 8
// CHECK-LE-NOOPT-NEXT: [[VC1_ADDR:%.*]] = alloca <16 x i8>, align 16
// CHECK-LE-NOOPT-NEXT: [[VC2_ADDR:%.*]] = alloca <16 x i8>, align 16
// CHECK-LE-NOOPT-NEXT: [[RESP_ADDR:%.*]] = alloca ptr, align 8
// CHECK-LE-NOOPT-NEXT: [[VQ:%.*]] = alloca <512 x i1>, align 64
// CHECK-LE-NOOPT-NEXT: [[VP:%.*]] = alloca <256 x i1>, align 32
// CHECK-LE-NOOPT-NEXT: [[RES:%.*]] = alloca <256 x i1>, align 32
// CHECK-LE-NOOPT-NEXT: store ptr [[VQP:%.*]], ptr [[VQP_ADDR]], align 8
// CHECK-LE-NOOPT-NEXT: store ptr [[VPP:%.*]], ptr [[VPP_ADDR]], align 8
// CHECK-LE-NOOPT-NEXT: store <16 x i8> [[VC1:%.*]], ptr [[VC1_ADDR]], align 16
// CHECK-LE-NOOPT-NEXT: store <16 x i8> [[VC2:%.*]], ptr [[VC2_ADDR]], align 16
// CHECK-LE-NOOPT-NEXT: store ptr [[RESP:%.*]], ptr [[RESP_ADDR]], align 8
// CHECK-LE-NOOPT-NEXT: [[TMP0:%.*]] = load ptr, ptr [[VQP_ADDR]], align 8
// CHECK-LE-NOOPT-NEXT: [[TMP1:%.*]] = load <512 x i1>, ptr [[TMP0]], align 64
// CHECK-LE-NOOPT-NEXT: store <512 x i1> [[TMP1]], ptr [[VQ]], align 64
// CHECK-LE-NOOPT-NEXT: [[TMP2:%.*]] = load ptr, ptr [[VPP_ADDR]], align 8
// CHECK-LE-NOOPT-NEXT: [[TMP3:%.*]] = load <256 x i1>, ptr [[TMP2]], align 32
// CHECK-LE-NOOPT-NEXT: store <256 x i1> [[TMP3]], ptr [[VP]], align 32
// CHECK-LE-NOOPT-NEXT: [[TMP4:%.*]] = load <16 x i8>, ptr [[VC1_ADDR]], align 16
// CHECK-LE-NOOPT-NEXT: [[TMP5:%.*]] = load <16 x i8>, ptr [[VC2_ADDR]], align 16
// CHECK-LE-NOOPT-NEXT: [[TMP6:%.*]] = call <256 x i1> @llvm.ppc.vsx.assemble.pair(<16 x i8> [[TMP5]], <16 x i8> [[TMP4]])
// CHECK-LE-NOOPT-NEXT: store <256 x i1> [[TMP6]], ptr [[RES]], align 64
// CHECK-LE-NOOPT-NEXT: [[TMP7:%.*]] = load <256 x i1>, ptr [[RES]], align 32
// CHECK-LE-NOOPT-NEXT: [[TMP8:%.*]] = load ptr, ptr [[RESP_ADDR]], align 8
// CHECK-LE-NOOPT-NEXT: store <256 x i1> [[TMP7]], ptr [[TMP8]], align 32
// CHECK-LE-NOOPT-NEXT: ret void
//
void test2(unsigned char *vqp, unsigned char *vpp, vector unsigned char vc1,
vector unsigned char vc2, unsigned char *resp) {
__vector_quad vq = *((__vector_quad *)vqp);
Expand Down
Loading