haraldwolff/llama.cpp
1
0
Fork 0
Code Issues Pull requests Packages Projects Releases Wiki Activity

CUDA: Fixed OpenLLaMA 3b mmq, reduced compile time (#2590)

Browse source
This commit is contained in:
Johannes Gäßler 2023-08-13 00:24:45 +02:00 committed by GitHub
parent b19edd54d5
commit f64d44a9b9
No known key found for this signature in database
GPG key ID: 4AEE18F83AFDEB23
2 changed files with 587 additions and 391 deletions
Show stats Download patch file Download diff file Expand all files Collapse all files

2
CMakeLists.txt
View file

@ -69,7 +69,6 @@ option(LLAMA_BLAS "llama: use BLAS"
set(LLAMA_BLAS_VENDOR "Generic" CACHE STRING "llama: BLAS library vendor")
option(LLAMA_CUBLAS "llama: use CUDA" OFF)
#option(LLAMA_CUDA_CUBLAS "llama: use cuBLAS for prompt processing" OFF)
set(LLAMA_CUDA_MMQ_Y "64" CACHE STRING "llama: y tile size for mmq CUDA kernels")
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_MMV_Y "1" CACHE STRING "llama: y block size for mmv CUDA kernels")
@ -256,7 +255,6 @@ if (LLAMA_CUBLAS)
# if (LLAMA_CUDA_CUBLAS)
# add_compile_definitions(GGML_CUDA_CUBLAS)
# endif()
add_compile_definitions(GGML_CUDA_MMQ_Y=${LLAMA_CUDA_MMQ_Y})
if (LLAMA_CUDA_FORCE_DMMV)
add_compile_definitions(GGML_CUDA_FORCE_DMMV)
endif()

976
ggml-cuda.cu
View file

File diff suppressed because it is too large Load diff