cmake : do not hide GGML options + rename option (llama/9465)

* cmake : do not hide GGML options

ggml-ci

* build : rename flag GGML_CUDA_USE_GRAPHS -> GGML_CUDA_GRAPHS

for consistency

ggml-ci
This commit is contained in:
Georgi Gerganov 2024-09-16 10:27:50 +03:00
parent 374e9e0c5e
commit 1fd78999e8
2 changed files with 12 additions and 3 deletions

View File

@ -56,6 +56,15 @@ else()
set(GGML_NATIVE_DEFAULT ON)
endif()
# defaults
if (NOT GGML_LLAMAFILE_DEFAULT)
set(GGML_LLAMAFILE_DEFAULT OFF)
endif()
if (NOT GGML_CUDA_GRAPHS_DEFAULT)
set(GGML_CUDA_GRAPHS_DEFAULT OFF)
endif()
# general
option(GGML_STATIC "ggml: static link libraries" OFF)
option(GGML_NATIVE "ggml: enable -march=native flag" ${GGML_NATIVE_DEFAULT})
@ -110,7 +119,7 @@ option(GGML_ACCELERATE "ggml: enable Accelerate framework"
option(GGML_BLAS "ggml: use BLAS" ${GGML_BLAS_DEFAULT})
set(GGML_BLAS_VENDOR ${GGML_BLAS_VENDOR_DEFAULT} CACHE STRING
"ggml: BLAS library vendor")
option(GGML_LLAMAFILE "ggml: use LLAMAFILE" OFF)
option(GGML_LLAMAFILE "ggml: use LLAMAFILE" ${GGML_LLAMAFILE_DEFAULT})
option(GGML_CUDA "ggml: use CUDA" OFF)
option(GGML_MUSA "ggml: use MUSA" OFF)
@ -127,7 +136,7 @@ set (GGML_CUDA_PEER_MAX_BATCH_SIZE "128" CACHE STRING
option(GGML_CUDA_NO_PEER_COPY "ggml: do not use peer to peer copies" OFF)
option(GGML_CUDA_NO_VMM "ggml: do not try to use CUDA VMM" OFF)
option(GGML_CUDA_FA_ALL_QUANTS "ggml: compile all quants for FlashAttention" OFF)
option(GGML_CUDA_USE_GRAPHS "ggml: use CUDA graphs (llama.cpp only)" OFF)
option(GGML_CUDA_GRAPHS "ggml: use CUDA graphs (llama.cpp only)" ${GGML_CUDA_GRAPHS_DEFAULT})
option(GGML_HIPBLAS "ggml: use hipBLAS" OFF)
option(GGML_HIP_UMA "ggml: use HIP unified memory architecture" OFF)

View File

@ -329,7 +329,7 @@ if (GGML_CUDA)
add_compile_definitions(K_QUANTS_PER_ITERATION=${GGML_CUDA_KQUANTS_ITER})
add_compile_definitions(GGML_CUDA_PEER_MAX_BATCH_SIZE=${GGML_CUDA_PEER_MAX_BATCH_SIZE})
if (GGML_CUDA_USE_GRAPHS)
if (GGML_CUDA_GRAPHS)
add_compile_definitions(GGML_CUDA_USE_GRAPHS)
endif()