root/llama.cpp
1
0
Fork 0
mirror of https://github.com/ggerganov/llama.cpp.git synced 2025-01-12 03:31:46 +00:00
Code Issues Actions 5 Packages Projects Releases Wiki Activity

CUDA: fixed cmake F16 option (#2471)

Browse Source
This commit is contained in:
Johannes Gäßler 2023-07-31 19:52:22 +02:00 committed by GitHub
parent 0728c5a8b9
commit b772bba42e
No known key found for this signature in database
GPG Key ID: 4AEE18F83AFDEB23
1 changed files with 3 additions and 3 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

6
CMakeLists.txt
View File

@ -73,7 +73,7 @@ set(LLAMA_CUDA_MMQ_Y "64" CACHE STRING "llama: y tile size for mmq CUDA ke
option(LLAMA_CUDA_FORCE_DMMV "llama: use dmmv instead of mmvq CUDA kernels" OFF) option(LLAMA_CUDA_FORCE_DMMV "llama: use dmmv instead of mmvq CUDA kernels" OFF)
set(LLAMA_CUDA_DMMV_X "32" CACHE STRING "llama: x stride for dmmv CUDA kernels") set(LLAMA_CUDA_DMMV_X "32" CACHE STRING "llama: x stride for dmmv CUDA kernels")
set(LLAMA_CUDA_MMV_Y "1" CACHE STRING "llama: y block size for mmv CUDA kernels") set(LLAMA_CUDA_MMV_Y "1" CACHE STRING "llama: y block size for mmv CUDA kernels")
option(LLAMA_CUDA_DMMV_F16 "llama: use 16 bit floats for dmmv CUDA kernels" OFF) option(LLAMA_CUDA_F16 "llama: use 16 bit floats for some calculations" OFF)
set(LLAMA_CUDA_KQUANTS_ITER "2" CACHE STRING "llama: iters./thread per block for Q2_K/Q6_K") set(LLAMA_CUDA_KQUANTS_ITER "2" CACHE STRING "llama: iters./thread per block for Q2_K/Q6_K")
option(LLAMA_CLBLAST "llama: use CLBlast" OFF) option(LLAMA_CLBLAST "llama: use CLBlast" OFF)
option(LLAMA_METAL "llama: use Metal" OFF) option(LLAMA_METAL "llama: use Metal" OFF)
@ -265,8 +265,8 @@ if (LLAMA_CUBLAS)
if (DEFINED LLAMA_CUDA_DMMV_Y) if (DEFINED LLAMA_CUDA_DMMV_Y)
add_compile_definitions(GGML_CUDA_MMV_Y=${LLAMA_CUDA_DMMV_Y}) # for backwards compatibility add_compile_definitions(GGML_CUDA_MMV_Y=${LLAMA_CUDA_DMMV_Y}) # for backwards compatibility
endif() endif()
if (LLAMA_CUDA_DMMV_F16) if (LLAMA_CUDA_F16 OR LLAMA_CUDA_DMMV_F16)
add_compile_definitions(GGML_CUDA_DMMV_F16) add_compile_definitions(GGML_CUDA_F16)
endif() endif()
add_compile_definitions(K_QUANTS_PER_ITERATION=${LLAMA_CUDA_KQUANTS_ITER}) add_compile_definitions(K_QUANTS_PER_ITERATION=${LLAMA_CUDA_KQUANTS_ITER})