From 7db1684a8b6e90956dc0996b9687dbd539c8c300 Mon Sep 17 00:00:00 2001 From: ben0i0d <ben0i0d@foxmail.com> Date: Mon, 20 Jan 2025 03:01:55 +0800 Subject: [PATCH] switch cuda & rocm to base & disable rocm(not support debian13) --- .gitea/workflows/Flow.yaml | 31 +++++++++++++++---------------- README.md | 6 +++--- README_CN.md | 6 +++--- base/cuda/base/Dockerfile | 7 ++----- base/rocm/Dockerfile | 28 ++++++++-------------------- 5 files changed, 31 insertions(+), 47 deletions(-) diff --git a/.gitea/workflows/Flow.yaml b/.gitea/workflows/Flow.yaml index e5ff6ab..eee74d9 100644 --- a/.gitea/workflows/Flow.yaml +++ b/.gitea/workflows/Flow.yaml @@ -16,6 +16,7 @@ jobs: CUDA_Base: runs-on: runner + needs: Base steps: - name: checkout code uses: https://eoelab.org:1027/actions/checkout@v4 @@ -23,8 +24,6 @@ jobs: uses: https://eoelab.org:1027/actions/build-push-action@v6 with: context: base/cuda/base - build-args: | - DEBIAN_MIRROR=mirrors.ustc.edu.cn tags: eoelab.org:1027/${{ gitea.repository }}:cuda_base CUDA_Runtime: @@ -51,20 +50,20 @@ jobs: context: base/cuda/devel tags: eoelab.org:1027/${{ gitea.repository }}:cuda_devel - ROCM: - runs-on: runner - steps: - - name: checkout code - uses: https://eoelab.org:1027/actions/checkout@v4 - - name: ROCM build - uses: https://eoelab.org:1027/actions/build-push-action@v6 - with: - context: base/rocm - build-args: | - UBUNTU_MIRROR=mirrors.ustc.edu.cn - http_proxy=http://192.168.2.2:7890 - https_proxy=http://192.168.2.2:7890 - tags: eoelab.org:1027/${{ gitea.repository }}:rocm + # ROCM: + # runs-on: runner + # needs: Base + # steps: + # - name: checkout code + # uses: https://eoelab.org:1027/actions/checkout@v4 + # - name: ROCM build + # uses: https://eoelab.org:1027/actions/build-push-action@v6 + # with: + # context: base/rocm + # build-args: | + # http_proxy=http://192.168.2.2:7890 + # https_proxy=http://192.168.2.2:7890 + # tags: eoelab.org:1027/${{ gitea.repository }}:rocm C: runs-on: runner diff --git a/README.md b/README.md index 7685c76..a09f4ae 100644 --- a/README.md +++ b/README.md @@ -13,7 +13,7 @@ English | [中文](README_CN.md) * CUDA : nvidia-container-toolkit/nvidia-docker2 ## Platform -* OS : debian 13 / ubuntu 24.04(ROCM only) +* OS : debian 13 * ARCH : x86_64 ## Version @@ -35,8 +35,8 @@ English | [中文](README_CN.md) graph LR CPU-->B{Base} -CUDA-->B -ROCM-->B +B-->CUDA +B-->ROCM B-->C{Renv} C-->CA(Hexo) diff --git a/README_CN.md b/README_CN.md index 0ea9c07..023ecf2 100644 --- a/README_CN.md +++ b/README_CN.md @@ -13,7 +13,7 @@ English | [中文](README_CN.md) * CUDA : nvidia-container-toolkit/nvidia-docker2 ## 目标平台 -* OS : debian 13 / ubuntu 24.04(只有ROCM) +* OS : debian 13 * ARCH : x86_64 ## 版本 @@ -35,8 +35,8 @@ English | [中文](README_CN.md) graph LR CPU-->B{Base} -CUDA-->B -ROCM-->B +B-->CUDA +B-->ROCM B-->C{Renv} C-->CA(Hexo) diff --git a/base/cuda/base/Dockerfile b/base/cuda/base/Dockerfile index 64a5d3d..f89ad88 100644 --- a/base/cuda/base/Dockerfile +++ b/base/cuda/base/Dockerfile @@ -1,6 +1,4 @@ -FROM debian:trixie-slim - -ARG DEBIAN_MIRROR +FROM eoelab.org:1027/eoeair/cenv:base ARG NV_CUDA_CUDART_VERSION=12.4.127-1 \ NV_CUDA_COMPAT_PACKAGE=cuda-compat-12-4 \ @@ -12,8 +10,7 @@ ENV NVIDIA_VISIBLE_DEVICES=all \ PATH=/usr/local/nvidia/bin:/usr/local/cuda/bin:${PATH} \ LD_LIBRARY_PATH=/usr/local/nvidia/lib:/usr/local/nvidia/lib64:${LD_LIBRARY_PATH} -RUN sed -i "s/deb.debian.org/${DEBIAN_MIRROR}/g" /etc/apt/sources.list.d/debian.sources && \ - apt-get update && apt-get install -y --no-install-recommends gnupg2 curl ca-certificates && \ +RUN apt-get update && apt-get install -y --no-install-recommends gnupg2 curl ca-certificates && \ curl -fsSLO https://developer.download.nvidia.com/compute/cuda/repos/ubuntu2204/${NVARCH}/cuda-keyring_1.1-1_all.deb && \ dpkg -i cuda-keyring_1.1-1_all.deb && rm cuda-keyring_1.1-1_all.deb &&\ # # For libraries in the cuda-compat-* package: https://docs.nvidia.com/cuda/eula/index.html#attachment-a diff --git a/base/rocm/Dockerfile b/base/rocm/Dockerfile index 05d3aaf..25896e5 100644 --- a/base/rocm/Dockerfile +++ b/base/rocm/Dockerfile @@ -1,25 +1,13 @@ -FROM ubuntu:24.04 - -ARG UBUNTU_MIRROR +FROM eoelab.org:1027/eoeair/cenv:base ARG VERSION=6.3.1 -ENV PATH="${PATH}:/opt/rocm/bin" \ - DEBIAN_FRONTEND=noninteractive +ENV PATH="${PATH}:/opt/rocm/bin" - # switch to mirror -RUN sed -i "s@//.*archive.ubuntu.com@//${UBUNTU_MIRROR}@g" /etc/apt/sources.list.d/ubuntu.sources && \ - apt-get update && apt-get install -y --no-install-recommends ca-certificates wget gnupg2 && \ - # add ROCm GPG key - mkdir --parents --mode=0755 /etc/apt/keyrings && \ - wget https://repo.radeon.com/rocm/rocm.gpg.key -O - | gpg --dearmor | tee /etc/apt/keyrings/rocm.gpg > /dev/null && \ - # add ROCm repository - echo "deb [arch=amd64 signed-by=/etc/apt/keyrings/rocm.gpg] http://repo.radeon.com/amdgpu/${VERSION}/ubuntu noble main" | tee /etc/apt/sources.list.d/amdgpu.list && \ - echo "deb [arch=amd64 signed-by=/etc/apt/keyrings/rocm.gpg] http://repo.radeon.com/rocm/apt/${VERSION} noble main" | tee --append /etc/apt/sources.list.d/rocm.list && \ - echo "Package: *\nPin: release o=repo.radeon.com\nPin-Priority: 600" | tee /etc/apt/preferences.d/rocm-pin-600 && \ +RUN apt-get update && apt-get install -y --no-install-recommends ca-certificates wget gnupg2 && \ + wget https://repo.radeon.com/amdgpu-install/6.3.1/ubuntu/jammy/amdgpu-install_6.3.60301-1_all.deb && \ + # add ROCm + apt-get install -y ./amdgpu-install_6.3.60301-1_all.deb && rm ./amdgpu-install_6.3.60301-1_all.deb && \ # Install ROCm-ml-libraries - apt-get update && apt-get install -y rocm-ml-libraries && \ - apt-get clean && rm -rf /var/lib/apt/lists/* && \ - # Configure the system linker by indicating where to find the shared objects - echo "/opt/rocm/lib\n/opt/rocm/lib64" > /etc/ld.so.conf.d/rocm.conf - + apt-get update && apt-get update && apt-get install -y rocm-ml-libraries && \ + apt-get clean && rm -rf /var/lib/apt/lists/*