Update Dockerfile and add workflow to register.

[ci skip]

.github/workflows/Container.yml

@@ -0,0 +1,105 @@
+name: Publish Docker image
+
+on:
+  workflow_dispatch:
+    inputs:
+      tag:
+        description: 'Tag to build instead'
+        required: false
+        default: ''
+      mark_as_latest:
+        description: 'Mark as latest'
+        type: boolean
+        required: false
+        default: false
+  push:
+    tags:
+      - 'v*'
+    branches:
+      - master
+
+jobs:
+  push_to_registry:
+    name: Build container - Julia ${{ matrix.julia }} - CUDA ${{ matrix.cuda }}
+    runs-on: ubuntu-latest
+    permissions:
+      contents: read
+      packages: write
+
+    strategy:
+      matrix:
+        julia: ["1.10", "1.11"]
+        cuda: ["11.8", "12.6"]
+        include:
+          - julia: "1.11"
+            cuda: "12.6"
+            default: true
+
+    steps:
+      - name: Check out the repo
+        uses: actions/checkout@v4
+
+      - name: Check out the package
+        uses: actions/checkout@v4
+        with:
+          ref: ${{ inputs.ref || github.ref_name }}
+          path: package
+
+      - name: Get package spec
+        id: pkg
+        run: |
+          if [[ -n "${{ inputs.tag }}" ]]; then
+            echo "ref=${{ inputs.tag }}" >> $GITHUB_OUTPUT
+            echo "name=${{ inputs.tag }}" >> $GITHUB_OUTPUT
+          elif [[ "${{ github.ref_type }}" == "tag" ]]; then
+            echo "ref=${{ github.ref_name }}" >> $GITHUB_OUTPUT
+            echo "name=${{ github.ref_name }}" >> $GITHUB_OUTPUT
+          else
+            echo "ref=${{ github.sha }}" >> $GITHUB_OUTPUT
+            echo "name=dev" >> $GITHUB_OUTPUT
+          fi
+
+          VERSION=$(grep "^version = " package/Project.toml | cut -d'"' -f2)
+          echo "version=$VERSION" >> $GITHUB_OUTPUT
+
+      - name: Get CUDA major version
+        id: cuda
+        run: |
+          CUDA_MAJOR=$(echo ${{ matrix.cuda }} | cut -d'.' -f1)
+          echo "major=${CUDA_MAJOR}" >> $GITHUB_OUTPUT
+
+      - name: Log in to registry
+        uses: docker/login-action@v3
+        with:
+          registry: ghcr.io
+          username: ${{ github.actor }}
+          password: ${{ secrets.GITHUB_TOKEN }}
+
+      - name: Extract metadata
+        id: meta
+        uses: docker/metadata-action@v5
+        with:
+          images: ghcr.io/${{ github.repository }}
+          tags: |
+              type=raw,value=${{ steps.pkg.outputs.name }}-julia${{ matrix.julia }}-julia${{ steps.cuda.outputs.major }}
+              type=raw,value=${{ steps.pkg.outputs.name }},enable=${{ matrix.default == true && (github.ref_type == 'tag' || inputs.tag != '') }}
+              type=raw,value=latest,enable=${{ matrix.default == true && (github.ref_type == 'tag' || (inputs.tag != '' && inputs.mark_as_latest)) }}
+              type=raw,value=dev,enable=${{ matrix.default == true && github.ref_type == 'branch' && inputs.tag == '' }}
+          labels: |
+              org.opencontainers.image.version=${{ steps.pkg.outputs.version }}
+
+      - name: Set up Docker Buildx
+        uses: docker/setup-buildx-action@v3
+
+      - name: Build and push image
+        uses: docker/build-push-action@v6
+        with:
+          context: .
+          push: true
+          provenance: false # the build fetches the repo again, so provenance tracking is not useful
+          tags: ${{ steps.meta.outputs.tags }}
+          labels: ${{ steps.meta.outputs.labels }}
+          build-args: |
+            JULIA_VERSION=${{ matrix.julia }}
+            CUDA_VERSION=${{ matrix.cuda }}
+            PACKAGE_SPEC=CUDA#${{ steps.pkg.outputs.ref }}

Dockerfile

@@ -1,21 +1,59 @@
 # example of a Docker container for CUDA.jl with a specific toolkit embedded at run time.
+#
+# supports selecting a Julia and CUDA toolkit version, as well as baking in specific
+# versions of the CUDA.jl package (identified by a spec recorgnized by the Pkg REPL).
+#
+# CUDA.jl and other packages are shipped in a system depot, with the user depot mounted
+# at `/depot`. persistency is possible by mounting a volume at this location.
+# running with reduced privileges (by using `--user`) is also supported.
 
-FROM julia:1.8-bullseye
+ARG JULIA_VERSION=1
+FROM julia:${JULIA_VERSION}
+
+ARG CUDA_VERSION=12.6
+
+ARG PACKAGE_SPEC=CUDA
+
+LABEL org.opencontainers.image.authors="Tim Besard <[email protected]>" \
+      org.opencontainers.image.description="A CUDA.jl container with CUDA ${CUDA_VERSION} and Julia ${JULIA_VERSION}" \
+      org.opencontainers.image.title="CUDA.jl" \
+      org.opencontainers.image.url="https://juliagpu.org/cuda/" \
+      org.opencontainers.image.source="https://github.com/JuliaGPU/CUDA.jl" \
+      org.opencontainers.image.licenses="MIT"
 
 
 # system-wide packages
 
+# no trailing ':' as to ensure we don't touch anything outside this directory. without it,
+# Julia touches the compilecache timestamps in its shipped depot (for some reason; a bug?)
 ENV JULIA_DEPOT_PATH=/usr/local/share/julia
 
-RUN julia -e 'using Pkg; Pkg.add("CUDA")'
+# pre-install the CUDA toolkit from an artifact. we do this separately from CUDA.jl so that
+# this layer can be cached independently. it also avoids double precompilation of CUDA.jl in
+# order to call `CUDA.set_runtime_version!`.
+RUN julia -e '#= configure the preference =# \
+              env = "/usr/local/share/julia/environments/v$(VERSION.major).$(VERSION.minor)"; \
+              mkpath(env); \
+              write("$env/LocalPreferences.toml", \
+                    "[CUDA_Runtime_jll]\nversion = \"'${CUDA_VERSION}'\""); \
+              \
+              #= install the JLL =# \
+              using Pkg; \
+              Pkg.add("CUDA_Runtime_jll")' && \
+    #= demote the JLL to an [extras] dep =# \
+    find /usr/local/share/julia/environments -name Project.toml -exec sed -i 's/deps/extras/' {} + && \
+    #= remove nondeterminisms =# \
+    cd /usr/local/share/julia && \
+    rm -rf compiled registries scratchspaces logs && \
+    find -exec touch -h -d "@0" {} + && \
+    touch -h -d "@0" /usr/local/share
 
-# hard-code a CUDA toolkit version
-RUN julia -e 'using CUDA; CUDA.set_runtime_version!(v"12.2")'
-# re-importing CUDA.jl below will trigger a download of the relevant artifacts
-
-# generate the device runtime library for all known and supported devices.
-# this is to avoid having to do this over and over at run time.
-RUN julia -e 'using CUDA; CUDA.precompile_runtime()'
+# install CUDA.jl itself
+RUN julia -e 'using Pkg; pkg"add '${PACKAGE_SPEC}'"; \
+              using CUDA; CUDA.precompile_runtime()' && \
+    #= remove useless stuff =# \
+    cd /usr/local/share/julia && \
+    rm -rf registries scratchspaces logs
 
 
 # user environment
@@ -25,6 +63,17 @@ RUN julia -e 'using CUDA; CUDA.precompile_runtime()'
 # case there might not be a (writable) home directory.
 
 RUN mkdir -m 0777 /depot
-ENV JULIA_DEPOT_PATH=/depot:/usr/local/share/julia
+
+# we add the user environment from a start-up script
+# so that the user can mount `/depot` for persistency
+ENV JULIA_DEPOT_PATH=/usr/local/share/julia:
+COPY <<EOF /usr/local/share/julia/config/startup.jl
+if !isdir("/depot/environments/v$(VERSION.major).$(VERSION.minor)")
+    mkpath("/depot/environments")
+    cp("/usr/local/share/julia/environments/v$(VERSION.major).$(VERSION.minor)",
+       "/depot/environments/v$(VERSION.major).$(VERSION.minor)")
+end
+pushfirst!(DEPOT_PATH, "/depot")
+EOF
 
 WORKDIR "/workspace"

README.md

@@ -59,6 +59,15 @@ This may take a while, as it will precompile the package and download a suitable
 the CUDA toolkit. If your GPU is not fully supported, the above command (or any other
 command that initializes the toolkit) will issue a warning.
 
+For quick testing, you can also use [the `juliagpu/cuda.jl` container
+image](https://github.com/JuliaGPU/CUDA.jl/pkgs/container/cuda.jl/versions) from the GitHub
+Container Registry, which provides Julia, a precompiled version of CUDA.jl, and a matching
+CUDA toolkit:
+
+```sh
+docker run -it --rm --gpus=all ghcr.io/juliagpu/cuda.jl:latest  # other tags available too
+```
+
 For more usage instructions and other information, please refer to [the
 documentation](https://juliagpu.github.io/CUDA.jl/stable/).