thek0tyara/llama-cpp-turboquant
1
0
Fork 0
Code Issues Pull requests Projects Releases Packages Wiki Activity Actions 4

HIP: add cmake option to enable compiler output of kernel resource usage metrics (#15103)

Browse source
This commit is contained in:
uvos 2025-08-07 16:44:14 +02:00 • committed by GitHub
parent 9a96389544
commit 7ad67ba9fe
No known key found for this signature in database
GPG key ID: B5690EEEBB952194
2 changed files with 5 additions and 0 deletions
Download patch file Download diff file Expand all files Collapse all files

4
ggml/src/ggml-hip/CMakeLists.txt
View file

@ -121,6 +121,10 @@ if (GGML_HIP_FORCE_ROCWMMA_FATTN_GFX12 OR ${hip_VERSION} VERSION_GREATER_EQUAL 7
add_compile_definitions(GGML_HIP_ROCWMMA_FATTN_GFX12)
endif()
if (GGML_HIP_EXPORT_METRICS)
set(CMAKE_HIP_FLAGS "${CMAKE_HIP_FLAGS} -Rpass-analysis=kernel-resource-usage --save-temps")
endif()
if (NOT GGML_CUDA_FA)
add_compile_definitions(GGML_CUDA_NO_FA)
endif()