mirror of
https://github.com/ROCm/composable_kernel.git
synced 2026-05-13 01:36:06 +00:00
* properly split conv_nd_bwd_data instances * split conv2d_fwd instance data types * split the gemm, conv2d_fwd and batched_gemm_softamx_gemm * split the tests by data types where possible * filter examples by DTYPES * split few remaining examples by DTYPES * filter most instances by DTYPES * add new lines at end of headers, fix grouped_gemm profiler * fix syntax * split the ckprofiler instances by DTYPES * split the conv2d and quantization DL and XDL instances * fix the splitting of conv2d DL instances * split softmax and pool_fwd tests for fp16 and fp32 types * fix syntax * fix the dl_int8 quantization instances isolation
29 lines
1.3 KiB
CMake
29 lines
1.3 KiB
CMake
list(APPEND gpu_list gfx908 gfx90a gfx940 gfx941 gfx942)
|
|
set(target 0)
|
|
foreach(gpu IN LISTS GPU_TARGETS)
|
|
if(gpu IN_LIST gpu_list AND target EQUAL 0)
|
|
add_custom_target(example_splitK_gemm_xdl)
|
|
if(DTYPES MATCHES "int8" OR NOT DEFINED DTYPES)
|
|
add_example_executable(example_splitK_gemm_xdl_fp32 splitK_gemm_xdl_fp32.cpp)
|
|
add_dependencies(example_splitK_gemm_xdl example_splitK_gemm_xdl_fp32)
|
|
endif()
|
|
if(DTYPES MATCHES "int8" OR NOT DEFINED DTYPES)
|
|
add_example_executable(example_splitK_gemm_xdl_fp16 splitK_gemm_xdl_fp16.cpp)
|
|
add_dependencies(example_splitK_gemm_xdl example_splitK_gemm_xdl_fp16)
|
|
endif()
|
|
if(DTYPES MATCHES "int8" OR NOT DEFINED DTYPES)
|
|
add_example_executable(example_splitK_gemm_xdl_bfp16 splitK_gemm_xdl_bfp16.cpp)
|
|
add_dependencies(example_splitK_gemm_xdl example_splitK_gemm_xdl_bfp16)
|
|
endif()
|
|
if(DTYPES MATCHES "int8" OR NOT DEFINED DTYPES)
|
|
add_example_executable(example_splitK_gemm_xdl_int8 splitK_gemm_xdl_int8.cpp)
|
|
add_dependencies(example_splitK_gemm_xdl example_splitK_gemm_xdl_int8)
|
|
endif()
|
|
if(USE_BITINT_EXTENSION_INT4)
|
|
add_example_executable(example_splitK_gemm_xdl_int4 splitK_gemm_xdl_int4.cpp)
|
|
add_dependencies(example_splitK_gemm_xdl example_splitK_gemm_xdl_int4)
|
|
endif()
|
|
set(target 1)
|
|
endif()
|
|
endforeach()
|