kanshan/llama.cpp
1
0
Fork 0
mirror of https://github.com/ggml-org/llama.cpp.git synced 2026-07-10 13:10:22 +00:00
Code Issues Packages Projects Releases Wiki Activity

ggml-hip: enable -funsafe-math-optimizations (#24668)

Browse Source 
CUDA is compiled with fast math and AMD/HIP is not — this flag lets AMD use fast math too.

We can't use -ffast-math: it implies -ffinite-math-only, which won't compile (ggml uses INFINITY for masking) and produces NaNs. -funsafe-math-optimizations gives the speedup without the NaN problems.

Co-authored-by: Mark Caldwell <mark@cloudhands.ai>
This commit is contained in:
RapidMark
2026-07-09 01:02:26 -07:00
committed by GitHub
parent 2021515a1a
commit ccb0c34223
1 changed files with 3 additions and 0 deletions
Download Patch File Download Diff File Expand all files Collapse all files
ggml/src/ggml-hip/CMakeLists.txt
+3
View File
@@ -130,6 +130,9 @@ if (GGML_HIP_EXPORT_METRICS)
set(CMAKE_HIP_FLAGS "${CMAKE_HIP_FLAGS} -Rpass-analysis=kernel-resource-usage --save-temps")
endif()
# Fast math for HIP, like CUDA's -use_fast_math. Not -ffast-math: that implies -ffinite-math-only, which breaks ggml's INFINITY masking and produces NaNs.
set(CMAKE_HIP_FLAGS "${CMAKE_HIP_FLAGS} -funsafe-math-optimizations")
if (NOT GGML_CUDA_FA)
add_compile_definitions(GGML_CUDA_NO_FA)
endif()