Skip to content

Rebalance CI runner utilization and enable apt proxy cache#496

Merged
bdice merged 2 commits intomainfrom
rebalance-ci
Jan 31, 2026
Merged

Rebalance CI runner utilization and enable apt proxy cache#496
bdice merged 2 commits intomainfrom
rebalance-ci

Conversation

@bdice
Copy link
Contributor

@bdice bdice commented Jan 30, 2026

Summary

  • Rebalance CI runner utilization: Shift PR test jobs to reduce pressure on overloaded runners (ARM+A100, amd64+L4) and better utilize underused capacity (V100, H100)
  • Enable apt proxy caching: Add enable-apt: true to all setup-proxy-cache actions to improve package installation performance

Runner rebalancing changes:

  • Move one ARM+A100 Python test to amd64+A100
  • Move one amd64 CUDA 12 C++ test from L4 to V100
  • Move one amd64 L4 job to H100 for conda C++ and Python tests
  • Add amd64+H100 wheel test job

Nightly jobs unchanged to maintain full coverage.

bdice added 2 commits January 30, 2026 12:36
Shift PR test jobs to reduce pressure on overloaded runners and better
utilize underused capacity:

- Move one ARM+A100 Python test to amd64+A100 (reduce ARM+A100 utilization)
- Move one amd64 CUDA 12 C++ test from L4 to V100 (currently underutilized)
- Move one amd64 L4 job to H100 for conda C++ and Python tests (currently underutilized)
- Add amd64+H100 wheel test job (currently underutilized)

Nightlies unchanged to maintain full coverage. Nightly load balancing currently looks fine.
@bdice bdice requested a review from a team as a code owner January 30, 2026 19:15
@bdice bdice requested review from KyleFromNVIDIA and removed request for a team January 30, 2026 19:15
@bdice bdice added improvement Improves an existing functionality non-breaking Introduces a non-breaking change labels Jan 30, 2026
@bdice bdice self-assigned this Jan 30, 2026
@bdice
Copy link
Contributor Author

bdice commented Jan 30, 2026

In the test PR, rapidsai/cudf#21273, all jobs have passed so far. There are a few jobs still queueing which are waiting on runners of the types that we are reducing in this PR. These changes should have real impact in PR turnaround time!

Copy link
Member

@jameslamb jameslamb left a comment

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Love this, let's put those underused machines to work!!

@bdice bdice merged commit ecda76b into main Jan 31, 2026
2 checks passed
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

improvement Improves an existing functionality non-breaking Introduces a non-breaking change

Projects

None yet

Development

Successfully merging this pull request may close these issues.

2 participants