diff --git a/Dockerfile_trtllm b/Dockerfile_trtllm index da507e3d..cbe4b3d3 100644 --- a/Dockerfile_trtllm +++ b/Dockerfile_trtllm @@ -1,14 +1,14 @@ -ARG cuda_arch_list="75-real;80-real;86-real;89-real;90-real" +ARG cuda_arch_list="75-real;80-real;86-real;89-real;90-real;100-real" +ARG cuda_base=12.8.0 ARG build_type=release ARG ompi_version=4.1.7 ARG sccache_gha_enabled=off ARG actions_cache_url="" ARG actions_runtime_token="" -ENV CUDA_BASE=12.8.0 # CUDA dependent dependencies resolver stage -FROM nvidia/cuda:${CUDA_BASE}-cudnn-devel-ubuntu24.04 AS cuda-builder +FROM nvidia/cuda:${cuda_base}-cudnn-devel-ubuntu24.04 AS cuda-builder RUN apt-get update && DEBIAN_FRONTEND=noninteractive apt-get install -y \ build-essential \ @@ -107,7 +107,7 @@ RUN export CMAKE_C_COMPILER_LAUNCHER=sccache && \ cargo build --profile ${build_type} --package text-generation-backends-trtllm --bin text-generation-backends-trtllm && \ sccache --show-stats -FROM nvidia/cuda:${CUDA_BASE}-cudnn-runtime-ubuntu24.04 AS runtime +FROM nvidia/cuda:${cuda_base}-cudnn-runtime-ubuntu24.04 AS runtime RUN apt update && apt install -y libucx0 pipx python3-minimal python3-dev python3-pip python3-venv && \ rm -rf /var/lib/{apt,dpkg,cache,log}/ && \ pipx ensurepath && \ @@ -126,7 +126,7 @@ COPY --from=tgi-builder /usr/local/tgi /usr/local/tgi COPY --from=tgi-builder /usr/src/text-generation-inference/target/release/text-generation-backends-trtllm /usr/local/tgi/bin/text-generation-launcher # This is used only for the CI/CD -FROM nvidia/cuda:${CUDA_BASE}-cudnn-runtime-ubuntu24.04 AS ci-runtime +FROM nvidia/cuda:${cuda_base}-cudnn-runtime-ubuntu24.04 AS ci-runtime RUN apt update && apt install -y libasan8 libubsan1 libucx0 pipx python3-minimal python3-dev python3-pip python3-venv && \ rm -rf /var/lib/{apt,dpkg,cache,log}/ && \ pipx ensurepath && \