LM/llama.cpp for SYCL

Download source

sudo apt install git cmake
sudo apt install libssl-dev openssl
mkdir -p ~/llama.cpp/src
git clone https://github.com/ggerganov/llama.cpp ~/llama.cpp/src

List available options

grep -rn "option(GGML_SYCL" CMakeLists.txt ggml/CMakeLists.txt

Build

# set options
cmake -B build-sycl \
  -DGGML_SYCL=ON \
  -DGGML_SYCL_TARGET=INTEL \
  -DGGML_SYCL_DNN=ON \
  -DGGML_SYCL_F16=ON \
  -DMKL_DIR=/opt/intel/oneapi/mkl/latest/lib/cmake/mkl \
  -DCMAKE_C_COMPILER=icx \
  -DCMAKE_CXX_COMPILER=icpx \
  -DCMAKE_BUILD_TYPE=Release \
  -DCMAKE_INSTALL_RPATH="\$ORIGIN/../lib"

# build
cmake -B build-sycl --config Release

# Install 
cmake --install build-sycl --prefix ~/llama.cpp/sycl

# clean
cmake -B build-sycl --target clean

# test GPU
cd ~/llama.cpp/sycl/bin
./llama-cli --list-devices

Got It!

Available devices:
  SYCL0: Intel(R) Arc(TM) Pro B70 Graphics (32656 MiB, 32128 MiB free)

References

https://sergiiob.dev/posts/intel-arc-pro-b70-full-recipe-llama-cpp-sycl/