tqcq/llama.cpp
0
0
Fork 0
mirror of https://github.com/ggml-org/llama.cpp.git synced 2025-06-26 19:55:04 +00:00
Code Issues Packages Projects Releases Wiki Activity

Convert vector to f16 for dequantize mul mat vec (#1913)

Browse Source 
* Convert vector to f16 for dmmv

* compile option

* Added compilation option description to README

* Changed cmake CUDA_ARCHITECTURES from "OFF" to "native"
This commit is contained in:
Johannes Gäßler
2023-06-19 10:23:56 +02:00
committed by GitHub
parent b24c3049d9
commit 16b9cd1939
5 changed files with 158 additions and 68 deletions
Download Patch File Download Diff File Expand all files Collapse all files

10
CMakeLists.txt
View File

@ -70,6 +70,7 @@ set(LLAMA_BLAS_VENDOR "Generic" CACHE STRING "llama: BLAS library vendor")
option(LLAMA_CUBLAS "llama: use cuBLAS" OFF)
set(LLAMA_CUDA_DMMV_X "32" CACHE STRING "llama: x stride for dmmv CUDA kernels")
set(LLAMA_CUDA_DMMV_Y "1" CACHE STRING "llama: y block size for dmmv CUDA kernels")
option(LLAMA_CUDA_DMMV_F16 "llama: use 16 bit floats for dmmv CUDA kernels" OFF)
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_METAL "llama: use Metal" OFF)
@ -238,6 +239,9 @@ if (LLAMA_CUBLAS)
add_compile_definitions(GGML_USE_CUBLAS)
add_compile_definitions(GGML_CUDA_DMMV_X=${LLAMA_CUDA_DMMV_X})
add_compile_definitions(GGML_CUDA_DMMV_Y=${LLAMA_CUDA_DMMV_Y})
if (LLAMA_CUDA_DMMV_F16)
add_compile_definitions(GGML_CUDA_DMMV_F16)
endif()
add_compile_definitions(K_QUANTS_PER_ITERATION=${LLAMA_CUDA_KQUANTS_ITER})
if (LLAMA_STATIC)
@ -490,13 +494,13 @@ endif()
if (GGML_SOURCES_CUDA)
message(STATUS "GGML CUDA sources found, configuring CUDA architecture")
set_property(TARGET ggml PROPERTY CUDA_ARCHITECTURES OFF)
set_property(TARGET ggml PROPERTY CUDA_ARCHITECTURES "native")
set_property(TARGET ggml PROPERTY CUDA_SELECT_NVCC_ARCH_FLAGS "Auto")
set_property(TARGET ggml_static PROPERTY CUDA_ARCHITECTURES OFF)
set_property(TARGET ggml_static PROPERTY CUDA_ARCHITECTURES "native")
set_property(TARGET ggml_static PROPERTY CUDA_SELECT_NVCC_ARCH_FLAGS "Auto")
set_property(TARGET llama PROPERTY CUDA_ARCHITECTURES OFF)
set_property(TARGET llama PROPERTY CUDA_ARCHITECTURES "native")
endif()