derek.holloway
7502bf64a4
CodeQL / Analyze (actions) (push) Canceled after 0s
CodeQL / Analyze (python) (push) Canceled after 0s
pre-commit / pre-commit (push) Canceled after 0s
Multi-Arch CI / setup (push) Canceled after 0s
Publish build_manylinux_x86_64 images / publish_build_manylinux_x86_64 (push) Canceled after 0s
Setup (Multi-Arch) / setup (push) Canceled after 0s
Multi-Arch CI / Linux::${{ fromJSON(needs.setup.outputs.linux_build_config || '{}').build_variant_label || 'skip' }} (push) Canceled after 0s
Publish TheRock Docker image / build-and-push-image (push) Canceled after 0s
Publish no_rocm_image_ubi10 image / publish_no_rocm_image_ubi10 (push) Canceled after 0s
Multi-Arch CI / Windows::${{ fromJSON(needs.setup.outputs.windows_build_config || '{}').build_variant_label || 'skip' }} (push) Canceled after 0s
Publish TheRock Docker image / build-and-push-image (push) Canceled after 0s
Multi-Arch CI / Manifest Diff (push) Canceled after 0s
TheRock Manifest Diff Report / Generate Manifest Diff Report (push) Canceled after 0s
Multi-Arch CI / CI Summary (push) Canceled after 0s
Publish no_rocm_image_ubuntu24_04 images / publish_no_rocm_image_ubuntu24_04 (push) Canceled after 0s
Publish no_rocm_image_ubuntu24_04_ocl_rt images / publish_no_rocm_image_ubuntu24_04_ocl_rt (push) Canceled after 0s
Publish TheRock Docker image / build-and-push-image (push) Canceled after 0s
Publish TheRock Docker image / build-and-push-image (push) Canceled after 0s
Publish no_rocm_image_ubuntu24_04 images / publish_no_rocm_image_ubuntu24_04_rocgdb (push) Canceled after 0s
Publish no_rocm_image_ubuntu24_04_rocgdb images / publish_no_rocm_image_ubuntu24_04_rocgdb (push) Canceled after 0s
Publish no_rocm_image_ubuntu24_04 images / publish_no_rocm_image_ubuntu24_04_ocl_rt (push) Canceled after 0s
Publish TheRock Docker image / build-and-push-image (push) Canceled after 0s
Unit Tests / Unit Tests :: ubuntu-24.04 (push) Canceled after 0s
Unit Tests / Unit Tests :: windows-2022 (push) Canceled after 0s
Publish no_rocm_image_ubuntu24_04 images / publish_no_rocm_image_ubuntu24_04_media (push) Canceled after 0s
Unit Tests / Unit Tests Summary (push) Canceled after 0s
Publish no_rocm_image_ubuntu24_04 images / publish_no_rocm_image_ubuntu24_04_openmpi (push) Canceled after 0s
63 lines
4.7 KiB
Plaintext
63 lines
4.7 KiB
Plaintext
// NOTE: Assertions have been autogenerated by utils/update_cc_test_checks.py UTC_ARGS: --version 6
|
|
// REQUIRES: amdgpu-registered-target
|
|
// RUN: %clang_cc1 -triple amdgcn-amd-amdhsa -target-cpu gfx1200 -emit-llvm -fcuda-is-device -o - %s | FileCheck %s --check-prefix=CHECK-GFX1200
|
|
|
|
#define __device__ __attribute__((device))
|
|
|
|
typedef float v8f __attribute__((ext_vector_type(8)));
|
|
typedef _Float16 v8h __attribute__((ext_vector_type(8)));
|
|
|
|
// CHECK-GFX1200-LABEL: define dso_local void @_Z47test_amdgcn_wmma_f32_16x16x16_f16_w32_gfx12_hipPDv8_fDv8_DF16_S1_S_(
|
|
// CHECK-GFX1200-SAME: ptr noundef [[OUT:%.*]], <8 x half> noundef [[A:%.*]], <8 x half> noundef [[B:%.*]], <8 x float> noundef [[C:%.*]]) #[[ATTR0:[0-9]+]] {
|
|
// CHECK-GFX1200-NEXT: [[ENTRY:.*:]]
|
|
// CHECK-GFX1200-NEXT: [[OUT_ADDR:%.*]] = alloca ptr, align 8, addrspace(5)
|
|
// CHECK-GFX1200-NEXT: [[A_ADDR:%.*]] = alloca <8 x half>, align 16, addrspace(5)
|
|
// CHECK-GFX1200-NEXT: [[B_ADDR:%.*]] = alloca <8 x half>, align 16, addrspace(5)
|
|
// CHECK-GFX1200-NEXT: [[C_ADDR:%.*]] = alloca <8 x float>, align 32, addrspace(5)
|
|
// CHECK-GFX1200-NEXT: [[OUT_ADDR_ASCAST:%.*]] = addrspacecast ptr addrspace(5) [[OUT_ADDR]] to ptr
|
|
// CHECK-GFX1200-NEXT: [[A_ADDR_ASCAST:%.*]] = addrspacecast ptr addrspace(5) [[A_ADDR]] to ptr
|
|
// CHECK-GFX1200-NEXT: [[B_ADDR_ASCAST:%.*]] = addrspacecast ptr addrspace(5) [[B_ADDR]] to ptr
|
|
// CHECK-GFX1200-NEXT: [[C_ADDR_ASCAST:%.*]] = addrspacecast ptr addrspace(5) [[C_ADDR]] to ptr
|
|
// CHECK-GFX1200-NEXT: store ptr [[OUT]], ptr [[OUT_ADDR_ASCAST]], align 8
|
|
// CHECK-GFX1200-NEXT: store <8 x half> [[A]], ptr [[A_ADDR_ASCAST]], align 16
|
|
// CHECK-GFX1200-NEXT: store <8 x half> [[B]], ptr [[B_ADDR_ASCAST]], align 16
|
|
// CHECK-GFX1200-NEXT: store <8 x float> [[C]], ptr [[C_ADDR_ASCAST]], align 32
|
|
// CHECK-GFX1200-NEXT: [[TMP0:%.*]] = load <8 x half>, ptr [[A_ADDR_ASCAST]], align 16
|
|
// CHECK-GFX1200-NEXT: [[TMP1:%.*]] = load <8 x half>, ptr [[B_ADDR_ASCAST]], align 16
|
|
// CHECK-GFX1200-NEXT: [[TMP2:%.*]] = load <8 x float>, ptr [[C_ADDR_ASCAST]], align 32
|
|
// CHECK-GFX1200-NEXT: [[TMP3:%.*]] = call contract <8 x float> @llvm.amdgcn.wmma.f32.16x16x16.f16.v8f32.v8f16(<8 x half> [[TMP0]], <8 x half> [[TMP1]], <8 x float> [[TMP2]])
|
|
// CHECK-GFX1200-NEXT: [[TMP4:%.*]] = load ptr, ptr [[OUT_ADDR_ASCAST]], align 8
|
|
// CHECK-GFX1200-NEXT: store <8 x float> [[TMP3]], ptr [[TMP4]], align 32
|
|
// CHECK-GFX1200-NEXT: ret void
|
|
//
|
|
__device__ void test_amdgcn_wmma_f32_16x16x16_f16_w32_gfx12_hip(v8f* out, v8h a, v8h b, v8f c) {
|
|
*out = __builtin_amdgcn_wmma_f32_16x16x16_f16_w32_gfx12(a, b, c);
|
|
}
|
|
|
|
// CHECK-GFX1200-LABEL: define dso_local void @_Z47test_amdgcn_wmma_f16_16x16x16_f16_w32_gfx12_hipPDv8_DF16_S_S_S_(
|
|
// CHECK-GFX1200-SAME: ptr noundef [[OUT:%.*]], <8 x half> noundef [[A:%.*]], <8 x half> noundef [[B:%.*]], <8 x half> noundef [[C:%.*]]) #[[ATTR0]] {
|
|
// CHECK-GFX1200-NEXT: [[ENTRY:.*:]]
|
|
// CHECK-GFX1200-NEXT: [[OUT_ADDR:%.*]] = alloca ptr, align 8, addrspace(5)
|
|
// CHECK-GFX1200-NEXT: [[A_ADDR:%.*]] = alloca <8 x half>, align 16, addrspace(5)
|
|
// CHECK-GFX1200-NEXT: [[B_ADDR:%.*]] = alloca <8 x half>, align 16, addrspace(5)
|
|
// CHECK-GFX1200-NEXT: [[C_ADDR:%.*]] = alloca <8 x half>, align 16, addrspace(5)
|
|
// CHECK-GFX1200-NEXT: [[OUT_ADDR_ASCAST:%.*]] = addrspacecast ptr addrspace(5) [[OUT_ADDR]] to ptr
|
|
// CHECK-GFX1200-NEXT: [[A_ADDR_ASCAST:%.*]] = addrspacecast ptr addrspace(5) [[A_ADDR]] to ptr
|
|
// CHECK-GFX1200-NEXT: [[B_ADDR_ASCAST:%.*]] = addrspacecast ptr addrspace(5) [[B_ADDR]] to ptr
|
|
// CHECK-GFX1200-NEXT: [[C_ADDR_ASCAST:%.*]] = addrspacecast ptr addrspace(5) [[C_ADDR]] to ptr
|
|
// CHECK-GFX1200-NEXT: store ptr [[OUT]], ptr [[OUT_ADDR_ASCAST]], align 8
|
|
// CHECK-GFX1200-NEXT: store <8 x half> [[A]], ptr [[A_ADDR_ASCAST]], align 16
|
|
// CHECK-GFX1200-NEXT: store <8 x half> [[B]], ptr [[B_ADDR_ASCAST]], align 16
|
|
// CHECK-GFX1200-NEXT: store <8 x half> [[C]], ptr [[C_ADDR_ASCAST]], align 16
|
|
// CHECK-GFX1200-NEXT: [[TMP0:%.*]] = load <8 x half>, ptr [[A_ADDR_ASCAST]], align 16
|
|
// CHECK-GFX1200-NEXT: [[TMP1:%.*]] = load <8 x half>, ptr [[B_ADDR_ASCAST]], align 16
|
|
// CHECK-GFX1200-NEXT: [[TMP2:%.*]] = load <8 x half>, ptr [[C_ADDR_ASCAST]], align 16
|
|
// CHECK-GFX1200-NEXT: [[TMP3:%.*]] = call contract <8 x half> @llvm.amdgcn.wmma.f16.16x16x16.f16.v8f16.v8f16(<8 x half> [[TMP0]], <8 x half> [[TMP1]], <8 x half> [[TMP2]], i1 false)
|
|
// CHECK-GFX1200-NEXT: [[TMP4:%.*]] = load ptr, ptr [[OUT_ADDR_ASCAST]], align 8
|
|
// CHECK-GFX1200-NEXT: store <8 x half> [[TMP3]], ptr [[TMP4]], align 16
|
|
// CHECK-GFX1200-NEXT: ret void
|
|
//
|
|
__device__ void test_amdgcn_wmma_f16_16x16x16_f16_w32_gfx12_hip(v8h* out, v8h a, v8h b, v8h c) {
|
|
*out = __builtin_amdgcn_wmma_f16_16x16x16_f16_w32_gfx12(a, b, c);
|
|
}
|