Files
TheRock/compiler/amd-llvm/amd/comgr/test-lit/hotswap/rewriter/hotswap-wmma-split-vgpr-msb-table-coverage.s
T
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
Flatten subprojects into a single base monorepo
2026-09-02 16:34:47 -07:00

114 lines
3.5 KiB
ArmAsm

// Table coverage: exercise every entry in the WMMA split table through the
// VGPR-MSB-aware split + bracket path. Each kernel holds one splittable opcode
// whose upper source half crosses v255, so the split runs and brackets the
// crossing half. The per-SplitKind bracket bytes are checked in detail by
// hotswap-wmma-split-vgpr-msb.s (Split128to64FP8BF8) and
// hotswap-wmma-split-vgpr-msb-m-roles.s (Split32x16to16x16F4); this test
// guarantees no table entry is left unsplit.
// RUN: %clang -target amdgcn-amd-amdhsa -mcpu=gfx1250 -nostdlib %s -o %t.elf
// RUN: env AMD_COMGR_EMIT_VERBOSE_LOGS=1 hotswap-rewrite %t.elf \
// RUN: amdgcn-amd-amdhsa--gfx1250 amdgcn-amd-amdhsa--gfx1250 \
// RUN: --output %t.out.elf 2>&1 | %FileCheck --check-prefix=API %s
// API-NOT: error:
// API: WMMA split: patched v_wmma_f16_16x16x128_fp8_fp8
// API: WMMA split: patched v_wmma_f16_16x16x128_fp8_bf8
// API: WMMA split: patched v_wmma_f16_16x16x128_bf8_fp8
// API: WMMA split: patched v_wmma_f16_16x16x128_bf8_bf8
// API: WMMA split: patched v_wmma_f32_16x16x128_fp8_fp8
// API: WMMA split: patched v_wmma_f32_16x16x128_fp8_bf8
// API: WMMA split: patched v_wmma_f32_16x16x128_bf8_fp8
// API: WMMA split: patched v_wmma_f32_16x16x128_bf8_bf8
// API: WMMA split: patched v_wmma_f32_32x16x128_f4
// API: RESULT: SUCCESS
// A second rewrite must be byte-identical.
// RUN: hotswap-rewrite %t.out.elf \
// RUN: amdgcn-amd-amdhsa--gfx1250 amdgcn-amd-amdhsa--gfx1250 \
// RUN: --check-idempotent | %FileCheck --check-prefix=IDEM %s
// IDEM: IDEMPOTENT: YES
.amdgcn_target "amdgcn-amd-amdhsa--gfx1250"
.text
.globl k_f16_fp8_fp8
.p2align 8
.type k_f16_fp8_fp8,@function
k_f16_fp8_fp8:
s_set_vgpr_msb 0
v_wmma_f16_16x16x128_fp8_fp8 v[16:19], v[250:265], v[100:115], v[16:19]
s_endpgm
.size k_f16_fp8_fp8, .-k_f16_fp8_fp8
.globl k_f16_fp8_bf8
.p2align 8
.type k_f16_fp8_bf8,@function
k_f16_fp8_bf8:
s_set_vgpr_msb 0
v_wmma_f16_16x16x128_fp8_bf8 v[16:19], v[250:265], v[100:115], v[16:19]
s_endpgm
.size k_f16_fp8_bf8, .-k_f16_fp8_bf8
.globl k_f16_bf8_fp8
.p2align 8
.type k_f16_bf8_fp8,@function
k_f16_bf8_fp8:
s_set_vgpr_msb 0
v_wmma_f16_16x16x128_bf8_fp8 v[16:19], v[250:265], v[100:115], v[16:19]
s_endpgm
.size k_f16_bf8_fp8, .-k_f16_bf8_fp8
.globl k_f16_bf8_bf8
.p2align 8
.type k_f16_bf8_bf8,@function
k_f16_bf8_bf8:
s_set_vgpr_msb 0
v_wmma_f16_16x16x128_bf8_bf8 v[16:19], v[250:265], v[100:115], v[16:19]
s_endpgm
.size k_f16_bf8_bf8, .-k_f16_bf8_bf8
.globl k_f32_fp8_fp8
.p2align 8
.type k_f32_fp8_fp8,@function
k_f32_fp8_fp8:
s_set_vgpr_msb 0
v_wmma_f32_16x16x128_fp8_fp8 v[16:23], v[250:265], v[100:115], v[16:23]
s_endpgm
.size k_f32_fp8_fp8, .-k_f32_fp8_fp8
.globl k_f32_fp8_bf8
.p2align 8
.type k_f32_fp8_bf8,@function
k_f32_fp8_bf8:
s_set_vgpr_msb 0
v_wmma_f32_16x16x128_fp8_bf8 v[16:23], v[250:265], v[100:115], v[16:23]
s_endpgm
.size k_f32_fp8_bf8, .-k_f32_fp8_bf8
.globl k_f32_bf8_fp8
.p2align 8
.type k_f32_bf8_fp8,@function
k_f32_bf8_fp8:
s_set_vgpr_msb 0
v_wmma_f32_16x16x128_bf8_fp8 v[16:23], v[250:265], v[100:115], v[16:23]
s_endpgm
.size k_f32_bf8_fp8, .-k_f32_bf8_fp8
.globl k_f32_bf8_bf8
.p2align 8
.type k_f32_bf8_bf8,@function
k_f32_bf8_bf8:
s_set_vgpr_msb 0
v_wmma_f32_16x16x128_bf8_bf8 v[16:23], v[250:265], v[100:115], v[16:23]
s_endpgm
.size k_f32_bf8_bf8, .-k_f32_bf8_bf8
.globl k_f32_32x16x128_f4
.p2align 8
.type k_f32_32x16x128_f4,@function
k_f32_32x16x128_f4:
s_set_vgpr_msb 0
v_wmma_f32_32x16x128_f4 v[250:265], v[248:263], v[100:107], v[252:267]
s_endpgm
.size k_f32_32x16x128_f4, .-k_f32_32x16x128_f4