mirror of
https://github.com/ROCm/composable_kernel.git
synced 2026-06-11 00:39:02 +00:00
3aee45e115
[CK_TILE] Add CShuffleLds microbenchmark suite (#5383) ## Summary Microbenchmarks isolating LDS store/load operations in CShuffleEpilogue for bank conflict analysis. ## Motivation CShuffleEpilogue performs LDS store (MFMA registers → LDS) and load (LDS → registers for coalesced global writes). This suite isolates each operation to: - Identify which operation causes bank conflicts - Measure pure LDS bandwidth per access pattern - Validate access patterns across MFMA tile sizes and wave layouts ## Components - **Microkernels** (`tile_load_store_microkernels.hpp`): `StoreTile<Setup>`, `LoadTile<Setup>` - **Setup Adapters** (`benchmark_cshuffle_lds.hpp`): Wire CShuffleEpilogue to microkernels - **Template** (`benchmark_template.cpp.in`): Generated benchmarks with timing ## Build ```bash cmake -G Ninja -B build -S . \ -DGPU_TARGETS=gfx950 \ -DBUILD_CK_EXAMPLES=ON \ -DBUILD_CK_TILE_CSHUFFLE_LDS_BENCHMARKS=ON ninja -C build bench_lds_fp8_16x16x128_2x2_fp8 ``` ## New CMake Options | Option | Default | Description | |--------|---------|-------------| | `BUILD_CK_TILE_CSHUFFLE_LDS_BENCHMARKS` | OFF | LDS microbenchmarks | | `BUILD_CK_TILE_FMHA_TESTS` | ON | FMHA tests | | `BUILD_CK_TILE_ENGINE` | ON | Tile engine | | `BUILD_CK_TILE_ENGINE_TESTS` | ON | Tile engine tests | | `BUILD_CK_EXAMPLES` | ON | Examples | | `BUILD_CK_TUTORIALS` | ON | Tutorials | | `BUILD_CK_DEVICE_INSTANCES` | ON | Device instances | | `BUILD_CK_PROFILER` | ON | Profiler | Setting guards to OFF reduces cmake configure from ~150s to ~5s. --------- Made-with: Claude Code, Opus 4.5
107 lines
3.2 KiB
JSON
107 lines
3.2 KiB
JSON
{
|
|
"version": 3,
|
|
"cmakeMinimumRequired": {
|
|
"major": 3,
|
|
"minor": 21,
|
|
"patch": 0
|
|
},
|
|
"configurePresets": [
|
|
{
|
|
"name": "use-gfx908",
|
|
"hidden": true,
|
|
"cacheVariables": {
|
|
"GPU_TARGETS": "gfx908"
|
|
}
|
|
},
|
|
{
|
|
"name": "use-gfx90a",
|
|
"hidden": true,
|
|
"cacheVariables": {
|
|
"GPU_TARGETS": "gfx90a"
|
|
}
|
|
},
|
|
{
|
|
"name": "use-gfx942",
|
|
"hidden": true,
|
|
"cacheVariables": {
|
|
"GPU_TARGETS": "gfx942"
|
|
}
|
|
},
|
|
{
|
|
"name": "use-gfx950",
|
|
"hidden": true,
|
|
"cacheVariables": {
|
|
"GPU_TARGETS": "gfx950"
|
|
}
|
|
},
|
|
{
|
|
"name": "dev",
|
|
"binaryDir": "${sourceDir}/build",
|
|
"displayName": "CK Dev",
|
|
"environment": {},
|
|
"cacheVariables": {
|
|
"CMAKE_PREFIX_PATH": "/opt/rocm/",
|
|
"CMAKE_CXX_COMPILER": "/opt/rocm/llvm/bin/clang++",
|
|
"CMAKE_HIP_COMPILER": "/opt/rocm/llvm/bin/clang++",
|
|
"CMAKE_CXX_FLAGS": "-ftemplate-backtrace-limit=0 -fPIE -Wno-gnu-line-marker -fbracket-depth=1024",
|
|
"CMAKE_BUILD_TYPE": "Release",
|
|
"BUILD_DEV": "ON",
|
|
"CMAKE_VERBOSE_MAKEFILE": "ON",
|
|
"USE_BITINT_EXTENSION_INT4": "OFF",
|
|
"GPU_TARGETS": "gfx908;gfx90a;gfx942"
|
|
}
|
|
},
|
|
{
|
|
"name": "dev-minimal",
|
|
"binaryDir": "${sourceDir}/build",
|
|
"displayName": "CK Dev - Minimal Build",
|
|
"description": "Fast iteration build with minimal components (configure ~5s vs ~150s)",
|
|
"inherits": ["dev"],
|
|
"cacheVariables": {
|
|
"BUILD_CK_DEVICE_INSTANCES": "OFF",
|
|
"BUILD_CK_PROFILER": "OFF",
|
|
"BUILD_CK_EXAMPLES": "OFF",
|
|
"BUILD_CK_TUTORIALS": "OFF",
|
|
"BUILD_CK_TILE_ENGINE": "OFF",
|
|
"BUILD_CK_TILE_ENGINE_TESTS": "OFF",
|
|
"BUILD_CK_TILE_FMHA_TESTS": "OFF"
|
|
}
|
|
},
|
|
{
|
|
"name": "dev-gfx908",
|
|
"displayName": "CK Dev - gfx908",
|
|
"description": "Development build for AMD GPU gfx908",
|
|
"inherits": [
|
|
"use-gfx908",
|
|
"dev"
|
|
]
|
|
},
|
|
{
|
|
"name": "dev-gfx90a",
|
|
"displayName": "CK Dev - gfx90a",
|
|
"description": "Development build for AMD GPU gfx90a",
|
|
"inherits": [
|
|
"use-gfx90a",
|
|
"dev"
|
|
]
|
|
},
|
|
{
|
|
"name": "dev-gfx942",
|
|
"displayName": "CK Dev - gfx942",
|
|
"description": "Development build for AMD GPU gfx942",
|
|
"inherits": [
|
|
"use-gfx942",
|
|
"dev"
|
|
]
|
|
},
|
|
{
|
|
"name": "dev-gfx950",
|
|
"displayName": "CK Dev - gfx950",
|
|
"description": "Development build for AMD GPU gfx950",
|
|
"inherits": [
|
|
"use-gfx950",
|
|
"dev"
|
|
]
|
|
}
|
|
]
|
|
} |