mirror of
https://github.com/ggml-org/llama.cpp.git
synced 2025-08-15 20:53:00 -04:00
CUDA: compress mode option and default to size (#12029)
cuda 12.8 added the option to specify stronger compression for binaries, so we now default to "size".
This commit is contained in:
@@ -102,6 +102,15 @@ if (CUDAToolkit_FOUND)
|
||||
|
||||
set(CUDA_FLAGS -use_fast_math)
|
||||
|
||||
if (CUDAToolkit_VERSION VERSION_GREATER_EQUAL "12.8")
|
||||
# Options are:
|
||||
# - none (not recommended)
|
||||
# - speed (nvcc's default)
|
||||
# - balance
|
||||
# - size
|
||||
list(APPEND CUDA_FLAGS -compress-mode=${GGML_CUDA_COMPRESSION_MODE})
|
||||
endif()
|
||||
|
||||
if (GGML_FATAL_WARNINGS)
|
||||
list(APPEND CUDA_FLAGS -Werror all-warnings)
|
||||
endif()
|
||||
|
Reference in New Issue
Block a user