mirror of
https://github.com/ROCm/composable_kernel.git
synced 2026-05-04 13:41:24 +00:00
[CK Tile] enable building examples by default (#3259)
* remove EXCLUDE_FROM_ALL from ck-tile examples -> +15 min build time w/ 64 threads for a single arch * fix cpp17 compile error in the ck-tile examples --------- Co-authored-by: khuagarw <khuagarw@amd.com> Co-authored-by: Ding, Yi <yi.ding@amd.com>
This commit is contained in:
@@ -1,12 +1,14 @@
|
||||
add_executable(tile_example_grouped_gemm EXCLUDE_FROM_ALL grouped_gemm.cpp)
|
||||
add_executable(tile_example_quant_grouped_gemm EXCLUDE_FROM_ALL quant_grouped_gemm.cpp)
|
||||
add_executable(tile_example_grouped_gemm_preshuffle EXCLUDE_FROM_ALL grouped_gemm_preshuffle.cpp)
|
||||
add_executable(tile_example_grouped_gemm_multi_d EXCLUDE_FROM_ALL grouped_gemm_multi_d.cpp)
|
||||
set(EXAMPLE_GEMM_COMPILE_OPTIONS)
|
||||
if(CK_USE_OCP_FP8)
|
||||
list(APPEND EXAMPLE_GEMM_COMPILE_OPTIONS -DCK_TILE_USE_OCP_FP8)
|
||||
if(GPU_TARGETS MATCHES "gfx94|gfx95|gfx90a")
|
||||
add_executable(tile_example_grouped_gemm grouped_gemm.cpp)
|
||||
add_executable(tile_example_quant_grouped_gemm quant_grouped_gemm.cpp)
|
||||
add_executable(tile_example_grouped_gemm_preshuffle grouped_gemm_preshuffle.cpp)
|
||||
add_executable(tile_example_grouped_gemm_multi_d grouped_gemm_multi_d.cpp)
|
||||
set(EXAMPLE_GEMM_COMPILE_OPTIONS)
|
||||
if(CK_USE_OCP_FP8)
|
||||
list(APPEND EXAMPLE_GEMM_COMPILE_OPTIONS -DCK_TILE_USE_OCP_FP8)
|
||||
endif()
|
||||
target_compile_options(tile_example_grouped_gemm PRIVATE ${EXAMPLE_GEMM_COMPILE_OPTIONS})
|
||||
target_compile_options(tile_example_grouped_gemm_preshuffle PRIVATE ${EXAMPLE_GEMM_COMPILE_OPTIONS})
|
||||
target_compile_options(tile_example_grouped_gemm_multi_d PRIVATE ${EXAMPLE_GEMM_COMPILE_OPTIONS})
|
||||
target_compile_options(tile_example_quant_grouped_gemm PRIVATE ${EXAMPLE_GEMM_COMPILE_OPTIONS})
|
||||
endif()
|
||||
target_compile_options(tile_example_grouped_gemm PRIVATE ${EXAMPLE_GEMM_COMPILE_OPTIONS})
|
||||
target_compile_options(tile_example_grouped_gemm_preshuffle PRIVATE ${EXAMPLE_GEMM_COMPILE_OPTIONS})
|
||||
target_compile_options(tile_example_grouped_gemm_multi_d PRIVATE ${EXAMPLE_GEMM_COMPILE_OPTIONS})
|
||||
target_compile_options(tile_example_quant_grouped_gemm PRIVATE ${EXAMPLE_GEMM_COMPILE_OPTIONS})
|
||||
@@ -233,7 +233,8 @@ int run_grouped_gemm_example_with_layouts(int argc,
|
||||
// Perform preshuffle for B tensor
|
||||
if constexpr(GemmConfig::Preshuffle)
|
||||
{
|
||||
ck_tile::HostTensor<BDataType> b_shuffle_host = shuffle_b<GemmConfig>(b_k_n_tensors[i]);
|
||||
ck_tile::HostTensor<BDataType> b_shuffle_host =
|
||||
ck_tile::shuffle_b<GemmConfig>(b_k_n_tensors[i]);
|
||||
b_k_n_dev_buf.push_back(std::make_unique<ck_tile::DeviceMem>(b_shuffle_host));
|
||||
}
|
||||
else
|
||||
|
||||
Reference in New Issue
Block a user