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

step3 add fp16, slower 31->28

Browse source
This commit is contained in:
jianyuzh 2023-12-31 12:38:57 +08:00 • committed by Meng, Hengyu
parent 02dffb68b8
commit 43f2c35859
2 changed files with 108 additions and 707 deletions
Download patch file Download diff file Expand all files Collapse all files

809
ggml-sycl.cpp
View file

File diff suppressed because it is too large Load diff

2
setup.sh
View file

@ -3,5 +3,7 @@ cd build
source /opt/intel/oneapi/setvars.sh
#cmake .. -DLLAMA_CLBLAST=ON -DCMAKE_C_COMPILER=icx -DCMAKE_CXX_COMPILER=icpx
#cmake .. -DLLAMA_SYCL=ON -DCMAKE_C_COMPILER=icx -DCMAKE_CXX_COMPILER=icpx -DGGML_SYCL_F16=ON
cmake .. -DLLAMA_SYCL=ON -DCMAKE_C_COMPILER=icx -DCMAKE_CXX_COMPILER=icpx
cmake --build . --config Release -v