Dockerfile

# 25.10 ships with Arc drivers
FROM ubuntu:25.10

ENV DEBIAN_FRONTEND=noninteractive

# Change the default shell to bash so we can source scripts natively
SHELL ["/bin/bash", "-c"]

# Install Compute Runtimes and Basekit 
# https://dgpu-docs.intel.com/driver/client/overview.html
# https://www.intel.com/content/www/us/en/docs/oneapi/installation-guide-linux/2025-2/cpp-apt.html
RUN apt-get update && apt-get install -y software-properties-common wget gnupg2 \
    && add-apt-repository -y ppa:kobuk-team/intel-graphics \
    && wget -O- https://apt.repos.intel.com/intel-gpg-keys/GPG-PUB-KEY-INTEL-SW-PRODUCTS.PUB | gpg --dearmor > /usr/share/keyrings/oneapi-archive-keyring.gpg \
    && echo "deb [signed-by=/usr/share/keyrings/oneapi-archive-keyring.gpg] https://apt.repos.intel.com/oneapi all main" | tee /etc/apt/sources.list.d/oneAPI.list \
    && rm -rf /var/lib/apt/lists/*

RUN apt-get update && apt-get install -y \
    libze-intel-gpu1 libze1 intel-metrics-discovery intel-opencl-icd clinfo intel-gsc libze-dev intel-ocloc \
    intel-cpp-essentials cmake pkg-config build-essential libssl-dev git \
    && rm -rf /var/lib/apt/lists/*

WORKDIR /app

ARG LLAMA_CPP_VERSION=master

# Clone and Compile 
# https://github.com/ggml-org/llama.cpp/blob/master/docs/backend/SYCL.md
RUN git clone --depth 1 --branch ${LLAMA_CPP_VERSION} https://github.com/ggml-org/llama.cpp . && \
    source /opt/intel/oneapi/setvars.sh && \
    cmake -B build -DGGML_SYCL=ON -DGGML_SYCL_TARGET=INTEL -DCMAKE_C_COMPILER=icx -DCMAKE_CXX_COMPILER=icpx -DGGML_SYCL_F16=ON -DLLAMA_OPENSSL=ON && \
    cmake --build build --config Release -j $(nproc)

# Intel Arc stability flag
ENV ZES_ENABLE_SYSMAN=1

EXPOSE 8080

# Execution
ENTRYPOINT ["/bin/bash", "-c", "source /opt/intel/oneapi/setvars.sh && exec /app/build/bin/llama-server \"$@\"", "--"]