diff --git a/Dockerfile_trtllm b/Dockerfile_trtllm index 2a0636ff..3330abdf 100644 --- a/Dockerfile_trtllm +++ b/Dockerfile_trtllm @@ -105,7 +105,7 @@ RUN export CMAKE_C_COMPILER_LAUNCHER=sccache && \ cargo build --profile ${build_type} --package text-generation-backends-trtllm --bin text-generation-backends-trtllm && \ sccache --show-stats -FROM nvidia/cuda:12.6.3-cudnn-runtime-ubuntu24.04 AS runtime +FROM nvidia/cuda:12.8.0-cudnn-runtime-ubuntu24.04 AS runtime RUN apt update && apt install -y libucx0 pipx python3-minimal python3-dev python3-pip python3-venv && \ rm -rf /var/lib/{apt,dpkg,cache,log}/ && \ pipx ensurepath && \ @@ -124,7 +124,7 @@ COPY --from=tgi-builder /usr/local/tgi /usr/local/tgi COPY --from=tgi-builder /usr/src/text-generation-inference/target/release/text-generation-backends-trtllm /usr/local/tgi/bin/text-generation-launcher # This is used only for the CI/CD -FROM nvidia/cuda:12.6.3-cudnn-runtime-ubuntu24.04 AS ci-runtime +FROM nvidia/cuda:12.8.0-cudnn-runtime-ubuntu24.04 AS ci-runtime RUN apt update && apt install -y libasan8 libubsan1 libucx0 pipx python3-minimal python3-dev python3-pip python3-venv && \ rm -rf /var/lib/{apt,dpkg,cache,log}/ && \ pipx ensurepath && \ diff --git a/backends/trtllm/cmake/trtllm.cmake b/backends/trtllm/cmake/trtllm.cmake index 3e9712c0..95a99e9b 100644 --- a/backends/trtllm/cmake/trtllm.cmake +++ b/backends/trtllm/cmake/trtllm.cmake @@ -28,7 +28,7 @@ find_package(Python3 REQUIRED Interpreter) fetchcontent_declare( trtllm GIT_REPOSITORY https://github.com/nvidia/TensorRT-LLM.git - GIT_TAG v0.16.0 + GIT_TAG v0.17.0 GIT_SHALLOW ON DOWNLOAD_EXTRACT_TIMESTAMP ) diff --git a/backends/trtllm/scripts/install_tensorrt.sh b/backends/trtllm/scripts/install_tensorrt.sh index f3e7270a..e09db6b1 100755 --- a/backends/trtllm/scripts/install_tensorrt.sh +++ b/backends/trtllm/scripts/install_tensorrt.sh @@ -2,13 +2,13 @@ set -ex -TRT_VER_BASE="10.7.0" -TRT_VER_FULL="${TRT_VER_BASE}.23" -CUDA_VER="12.6" -CUDNN_VER="9.5.0.50-1" -NCCL_VER="2.22.3-1+cuda12.6" -CUBLAS_VER="12.6.3.3-1" -NVRTC_VER="12.6.77-1" +TRT_VER_BASE="10.8.0" +TRT_VER_FULL="${TRT_VER_BASE}.43" +CUDA_VER="12.8" +CUDNN_VER="9.7.0.66-1" +NCCL_VER="2.25.1-1+cuda${CUDA_VER}" +CUBLAS_VER="${CUDA_VER}.3.14-1" +NVRTC_VER="${CUDA_VER}.61-1" for i in "$@"; do case $i in @@ -73,7 +73,7 @@ install_centos_requirements() { install_tensorrt() { #PY_VERSION=$(python3 -c 'import sys; print(".".join(map(str, sys.version_info[0:2])))') #PARSED_PY_VERSION=$(echo "${PY_VERSION//./}") - TRT_CUDA_VERSION="12.6" + TRT_CUDA_VERSION="12.8" if [ -z "$RELEASE_URL_TRT" ];then ARCH=${TRT_TARGETARCH}