Skip to content

[CI] Split TPU and GPU nightly builds to run E2E tests in parallel#4400

Closed
RexBearIU wants to merge 1 commit into
mainfrom
jackyf/ci-test-parallel
Closed

[CI] Split TPU and GPU nightly builds to run E2E tests in parallel#4400
RexBearIU wants to merge 1 commit into
mainfrom
jackyf/ci-test-parallel

Conversation

@RexBearIU

Copy link
Copy Markdown
Collaborator

Re-submitting fork changes to trigger CI.

Original description:

  • Split the single build_and_push_docker_image matrix job into parallel build_and_push_tpu_docker_images and build_and_push_gpu_docker_images jobs.
  • Update run_e2e_tests to depend only on the TPU builds, allowing the Airflow E2E tests to trigger immediately after TPU pre/post-training images are pushed.
  • Ensure GPU CI tests run in parallel with the Airflow E2E tests, drastically reducing overall pipeline runtime.
  • Correct the maxtext_sha reference to use the correct job ID build_maxtext_package.
  • Update notify_failure to monitor both build jobs and the E2E tests.

- Split the single `build_and_push_docker_image` matrix job into parallel `build_and_push_tpu_docker_images` and `build_and_push_gpu_docker_images` jobs.
- Update `run_e2e_tests` to depend only on the TPU builds, allowing the Airflow E2E tests to trigger immediately after TPU
pre/post-training images are pushed.
- Ensure GPU CI tests run in parallel with the Airflow E2E tests, drastically reducing overall pipeline runtime.
- Correct the `maxtext_sha` reference to use the correct job ID `build_maxtext_package`.
- Update `notify_failure` to monitor both build jobs and the E2E tests.
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

2 participants