ROCm/composable_kernel
1
0
Fork 0
mirror of https://github.com/ROCm/composable_kernel.git synced 2026-05-14 10:09:41 +00:00
Code Issues Packages Projects Releases Wiki Activity

check for if misched-bottomup flag is valid (#2341)

Browse Source 
[ROCm/composable_kernel commit: a0f4db8d9c]
This commit is contained in:
Illia Silin
2025-06-13 13:34:22 -07:00
committed by GitHub
parent 044a8560f7
commit 201ba3edb6
2 changed files with 20 additions and 7 deletions
Download Patch File Download Diff File Expand all files Collapse all files

8
example/65_gemm_multiply_multiply/CMakeLists.txt
View File

@@ -43,7 +43,13 @@ endforeach()
set(GEMM_OPTIONS)
list(APPEND GEMM_OPTIONS "SHELL: -mllvm -greedy-reverse-local-assignment=1 -mllvm --slp-threshold=-32")
set(BLOCKSCALE_GEMM_OPTIONS)
list(APPEND BLOCKSCALE_GEMM_OPTIONS "SHELL: -mllvm -greedy-reverse-local-assignment=1 -mllvm --slp-threshold=-32 -mllvm --schedmodel=0 -mllvm --misched-bottomup=1")
check_cxx_compiler_flag("-mllvm --misched-bottomup=1" HAS_MISCHED_BOTTOMUP)
check_cxx_compiler_flag("-mllvm --misched-prera-direction=bottomup" HAS_MISCHED_PRERA_DIRECTION)
if(HAS_MISCHED_BOTTOMUP)
list(APPEND BLOCKSCALE_GEMM_OPTIONS "SHELL: -mllvm -greedy-reverse-local-assignment=1 -mllvm --slp-threshold=-32 -mllvm --schedmodel=0 -mllvm --misched-bottomup=1")
elseif(HAS_MISCHED_PRERA_DIRECTION)
list(APPEND BLOCKSCALE_GEMM_OPTIONS "SHELL: -mllvm -greedy-reverse-local-assignment=1 -mllvm --slp-threshold=-32 -mllvm --schedmodel=0 -mllvm --misched-prera-direction=bottomup")
endif()
check_cxx_compiler_flag("-mllvm --amdgpu-sched-strategy=gcn-iterative-max-occupancy-experimental " HAS_MAX_OCCUPANCY_EXPERIMENTAL)
if(HAS_MAX_OCCUPANCY_EXPERIMENTAL)
list(APPEND BLOCKSCALE_GEMM_OPTIONS -mllvm --amdgpu-sched-strategy=gcn-iterative-max-occupancy-experimental)

19
library/src/tensor_operation_instance/gpu/gemm_blockscale_wp/CMakeLists.txt
View File

@@ -7,10 +7,17 @@ list(APPEND GEMM_BLOCKSCALE_WP_INSTANCES
device_gemm_blockscale_wp_xdl_f8_f8_bf16/device_gemm_blockscale_wp_xdl_f8_f8_bf16_mk_nk_mn_128_128_128_mem_v1_default_instance.cpp
device_gemm_blockscale_wp_xdl_f8_f8_bf16/device_gemm_blockscale_wp_xdl_f8_f8_bf16_mk_nk_mn_128_128_128_mem_v1_kpadding_instance.cpp
)
set_source_files_properties(device_gemm_blockscale_wp_xdl_f8_f8_bf16/device_gemm_blockscale_wp_xdl_f8_f8_bf16_mk_nk_mn_128_128_128_comp_default_instance.cpp PROPERTIES COMPILE_OPTIONS ";-mllvm;-greedy-reverse-local-assignment=1;-mllvm;--slp-threshold=-32;-mllvm;--misched-bottomup=1")
set_source_files_properties(device_gemm_blockscale_wp_xdl_f8_f8_bf16/device_gemm_blockscale_wp_xdl_f8_f8_bf16_mk_nk_mn_128_128_128_comp_kpadding_instance.cpp PROPERTIES COMPILE_OPTIONS ";-mllvm;-greedy-reverse-local-assignment=1;-mllvm;--slp-threshold=-32;-mllvm;--misched-bottomup=1")
set_source_files_properties(device_gemm_blockscale_wp_xdl_f8_f8_bf16/device_gemm_blockscale_wp_xdl_f8_f8_bf16_mk_nk_mn_128_128_128_mem_v1_default_instance.cpp PROPERTIES COMPILE_OPTIONS ";-mllvm;-greedy-reverse-local-assignment=1;-mllvm;--slp-threshold=-32;-mllvm;--misched-bottomup=1")
set_source_files_properties(device_gemm_blockscale_wp_xdl_f8_f8_bf16/device_gemm_blockscale_wp_xdl_f8_f8_bf16_mk_nk_mn_128_128_128_mem_v1_kpadding_instance.cpp PROPERTIES COMPILE_OPTIONS ";-mllvm;-greedy-reverse-local-assignment=1;-mllvm;--slp-threshold=-32;-mllvm;--misched-bottomup=1")
check_cxx_compiler_flag("-mllvm --misched-bottomup=1" HAS_MISCHED_BOTTOMUP)
check_cxx_compiler_flag("-mllvm --misched-prera-direction=bottomup" HAS_MISCHED_PRERA_DIRECTION)
if(HAS_MISCHED_BOTTOMUP)
set_source_files_properties(device_gemm_blockscale_wp_xdl_f8_f8_bf16/device_gemm_blockscale_wp_xdl_f8_f8_bf16_mk_nk_mn_128_128_128_comp_default_instance.cpp PROPERTIES COMPILE_OPTIONS ";-mllvm;-greedy-reverse-local-assignment=1;-mllvm;--slp-threshold=-32;-mllvm;--misched-bottomup=1")
set_source_files_properties(device_gemm_blockscale_wp_xdl_f8_f8_bf16/device_gemm_blockscale_wp_xdl_f8_f8_bf16_mk_nk_mn_128_128_128_comp_kpadding_instance.cpp PROPERTIES COMPILE_OPTIONS ";-mllvm;-greedy-reverse-local-assignment=1;-mllvm;--slp-threshold=-32;-mllvm;--misched-bottomup=1")
set_source_files_properties(device_gemm_blockscale_wp_xdl_f8_f8_bf16/device_gemm_blockscale_wp_xdl_f8_f8_bf16_mk_nk_mn_128_128_128_mem_v1_default_instance.cpp PROPERTIES COMPILE_OPTIONS ";-mllvm;-greedy-reverse-local-assignment=1;-mllvm;--slp-threshold=-32;-mllvm;--misched-bottomup=1")
set_source_files_properties(device_gemm_blockscale_wp_xdl_f8_f8_bf16/device_gemm_blockscale_wp_xdl_f8_f8_bf16_mk_nk_mn_128_128_128_mem_v1_kpadding_instance.cpp PROPERTIES COMPILE_OPTIONS ";-mllvm;-greedy-reverse-local-assignment=1;-mllvm;--slp-threshold=-32;-mllvm;--misched-bottomup=1")
elseif(HAS_MISCHED_PRERA_DIRECTION)
set_source_files_properties(device_gemm_blockscale_wp_xdl_f8_f8_bf16/device_gemm_blockscale_wp_xdl_f8_f8_bf16_mk_nk_mn_128_128_128_comp_default_instance.cpp PROPERTIES COMPILE_OPTIONS ";-mllvm;-greedy-reverse-local-assignment=1;-mllvm;--slp-threshold=-32;-mllvm;--misched-prera-direction=bottomup")
set_source_files_properties(device_gemm_blockscale_wp_xdl_f8_f8_bf16/device_gemm_blockscale_wp_xdl_f8_f8_bf16_mk_nk_mn_128_128_128_comp_kpadding_instance.cpp PROPERTIES COMPILE_OPTIONS ";-mllvm;-greedy-reverse-local-assignment=1;-mllvm;--slp-threshold=-32;-mllvm;--misched-prera-direction=bottomup")
set_source_files_properties(device_gemm_blockscale_wp_xdl_f8_f8_bf16/device_gemm_blockscale_wp_xdl_f8_f8_bf16_mk_nk_mn_128_128_128_mem_v1_default_instance.cpp PROPERTIES COMPILE_OPTIONS ";-mllvm;-greedy-reverse-local-assignment=1;-mllvm;--slp-threshold=-32;-mllvm;--misched-prera-direction=bottomup")
set_source_files_properties(device_gemm_blockscale_wp_xdl_f8_f8_bf16/device_gemm_blockscale_wp_xdl_f8_f8_bf16_mk_nk_mn_128_128_128_mem_v1_kpadding_instance.cpp PROPERTIES COMPILE_OPTIONS ";-mllvm;-greedy-reverse-local-assignment=1;-mllvm;--slp-threshold=-32;-mllvm;--misched-prera-direction=bottomup")
endif()
add_instance_library(device_gemm_blockscale_wp_instance ${GEMM_BLOCKSCALE_WP_INSTANCES})