mirror of
https://github.com/ROCm/composable_kernel.git
synced 2026-07-18 09:38:17 +00:00
2a73eb3bc0
* tf32:bf16x3:use bf16x3 emulate tf32 gemm * change blockwiseGemm to demo bf16x3 * temp push * self review * self review * fix multi-device compile error * bug fix * code refactor * limit to gfx950 * enhance gemm gfx942 threshold * lower change from blockwise to warpwise * refact codes * refact codes * error fix * change threshold * bug fix * fix threshold error * change host reference implement to same as device * bug fix * bug fix * code refact * fix clang-format fail * code refine
32 lines
1.5 KiB
CMake
32 lines
1.5 KiB
CMake
add_example_executable(example_convnd_fwd_xdl_fp32 convnd_fwd_xdl_fp32.cpp)
|
|
add_example_executable(example_convnd_fwd_xdl_fp16 convnd_fwd_xdl_fp16.cpp)
|
|
add_example_executable(example_convnd_fwd_xdl_bf16 convnd_fwd_xdl_bf16.cpp)
|
|
add_example_executable(example_convnd_fwd_xdl_int8 convnd_fwd_xdl_int8.cpp)
|
|
add_example_executable(example_convnd_fwd_xdl_fp8 convnd_fwd_xdl_fp8.cpp)
|
|
add_example_executable(example_convnd_fwd_xdl_bf8 convnd_fwd_xdl_bf8.cpp)
|
|
add_example_executable(example_convnd_fwd_xdl_fp16_comp_fp8 convnd_fwd_xdl_fp16_comp_fp8.cpp)
|
|
add_example_executable(example_convnd_fwd_xdl_fp8_bf8 convnd_fwd_xdl_fp8_bf8.cpp)
|
|
add_example_executable(example_convnd_fwd_xdl_bf8_fp8 convnd_fwd_xdl_bf8_fp8.cpp)
|
|
add_example_executable(example_convnd_fwd_dl_fp16 convnd_fwd_dl_fp16.cpp)
|
|
add_example_executable(example_convnd_fwd_dl_fp32 convnd_fwd_dl_fp32.cpp)
|
|
add_example_executable(example_convnd_fwd_dl_int8 convnd_fwd_dl_int8.cpp)
|
|
|
|
# only build fp64 example for the following targets
|
|
list(APPEND gpu_list gfx90a gfx942 gfx950)
|
|
set(target 0)
|
|
foreach(gpu IN LISTS GPU_TARGETS)
|
|
if(gpu IN_LIST gpu_list AND target EQUAL 0)
|
|
add_example_executable(example_convnd_fwd_xdl_fp64 convnd_fwd_xdl_fp64.cpp)
|
|
set(target 1)
|
|
endif()
|
|
endforeach()
|
|
|
|
list(APPEND gpu_list_tf32 gfx942 gfx950)
|
|
set(target 0)
|
|
foreach(gpu IN LISTS GPU_TARGETS)
|
|
if(gpu IN_LIST gpu_list_tf32 AND target EQUAL 0)
|
|
add_example_executable(example_convnd_fwd_xdl_fp32_tf32 convnd_fwd_xdl_fp32_tf32.cpp)
|
|
set(target 1)
|
|
endif()
|
|
endforeach()
|