Google Git
Sign in
opensecura / 3p / openxla / iree / 538891d1a9422b8941c0e7622b2e60544d80f39b / . / build_tools / docker
tree: cb09cf793f281b0b33df6804545cda20a62217ea [path history] [tgz]
  1. base/
  2. bazel/
  3. bazel-python/
  4. bazel-tensorflow/
  5. bazel-tensorflow-nvidia/
  6. bazel-tensorflow-swiftshader/
  7. bazel-tensorflow-vulkan/
  8. cmake/
  9. cmake-android/
  10. cmake-python/
  11. cmake-python-nvidia/
  12. cmake-python-swiftshader/
  13. cmake-python-vulkan/
  14. rbe-toolchain/
  15. swiftshader/
  16. util/
  17. gcloud_setup.sh
  18. manage_images.py
  19. README.md
build_tools/docker/README.md

IREE Docker Configuration

This directory contains the Dockerfiles that specify the container images used for IREE. Images are uploaded to Google Container Registry (GCR).

To build an image, use docker build, e.g.:

docker build build_tools/docker/cmake --tag cmake

To explore an image interactively, use docker run, e.g.

docker run --interactive --tty --rm cmake

You can find more information in the official Docker docs.

Multi-stage Builds

We use multi-stage builds to limit duplication in our Dockerfiles and reduce the final image size. There is still duplication in cases where it's difficult to determine the correct files to copy.

Dependencies Between Images

IREE images follow a consistent structure. The image defined by build_tools/docker/foo-bar/Dockerfile is uploaded to GCR as gcr.io/iree-oss/foo-bar. It may be tagged as latest or prod, e.g. gcr.io/iree-oss/foo-bar:latest. Dockerfile image definitions should list their dependencies based on these image names.

We use a helper python script to manage the Docker image deployment. It lists all images and their dependencies and manages their canonical registry location. When creating a new image, add it to this mapping. To build an image and all images it depends on:

python3 build_tools/docker/manage_images.py --build --image cmake

To build multiple images

python3 build_tools/docker/manage_images.py --build --image cmake --image bazel

There is also the special option --image all to build all registered images.

Pushing images to GCR requires the Storage Admin role in the iree-oss GCP project. To push these images to GCR with the latest tag:

python3 build_tools/docker/manage_images.py --image cmake --push

Kokoro build scripts and RBE configuration refer to images by their repository digest. You can update references to the digest:

python3 build_tools/docker/manage_images.py --images all --tag latest --update_references

This requires that the tagged image have a repository digest, which means it was pushed to or pulled from GCR.

Deploying New Images

  1. Modify the Dockerfiles as desired.

  2. Update manage_images.py to include the new image and its dependencies.

  3. Build and push the new image to GCR and update references to it:

    python3 build_tools/docker/manage_images.py --image "${IMAGE?}" --build --push --update_references

  4. Commit changes and send a PR for review.

  5. Merge your PR after is approved and all builds pass.

  6. Kokoro builds preload images tagged with prod on VM creation, so after changing the images used, you should also update the images tagged as prod in GCR. Update your local reference to the prod tag to point at the new image:

    python3 build_tools/docker/manage_images.py --image "${IMAGE?}" --tag prod --build --update_references

    The build steps here should all be cache hits and no references should actually be changed. If they are, that indicates the images you‘ve just built are different from the ones that are being referenced. Stop and fix this before proceeding. This relies on you keeping your local copy of the Docker images. If you didn’t, you'll have to manually pull the missing images by their digest.

  7. Push the new images with the prod tag to GCR.

    python3 build_tools/docker/manage_images.py --image "${IMAGE?}" --tag prod --push