Skip to content

[DO NOT MERGE] Debug T4 runner performance - #982

Closed
leofang with Copilot wants to merge 2 commits into
mainfrom
copilot/fix-bac5ba47-0510-4706-ba04-b7a571b52bd1
Closed

leofang with Copilot wants to merge 2 commits into
mainfrom
copilot/fix-bac5ba47-0510-4706-ba04-b7a571b52bd1

Conversation

Copilot AI commented Sep 18, 2025 •

Copy link
Copy Markdown
Contributor

This PR switches all Linux AMD64 GPU runners from L4 to T4 in the CI test matrix to debug performance characteristics of T4 runners.

Changes Made

  • Modified ci/test-matrix.json to update all Linux AMD64 entries from "GPU": "l4" to "GPU": "t4"
  • Updated 35 total entries across both pull-request (10 entries) and nightly (25 entries) test matrices
  • Updated the description comment to reflect the new T4 GPU assignment for Linux AMD64

What's Preserved

  • ARM64 entries continue using A100 GPUs unchanged
  • Windows entries remain with their original L4/T4 mix unchanged
  • Special runners (H100) remain unchanged
  • All other configuration parameters (Python versions, CUDA versions, drivers, etc.) are preserved

Purpose

This change allows us to evaluate T4 runner performance characteristics compared to the previous L4 configuration for Linux AMD64 workloads. The change is marked as "DO NOT MERGE" as this is intended for performance testing and debugging purposes only.

Validation

  • JSON structure validation passes
  • All Linux AMD64 entries confirmed to use T4 GPUs
  • No unintended modifications to other platform configurations

✨ Let Copilot coding agent set things up for you — coding agent works faster and does higher quality work when set up for your repo.

@copy-pr-bot

copy-pr-bot Bot commented Sep 18, 2025

Copy link
Copy Markdown
Contributor

This pull request requires additional validation before any workflows can run on NVIDIA's runners.

Pull request vetters can view their responsibilities here.

Contributors can view more details about this message here.

Co-authored-by: leofang <5534781+leofang@users.noreply.github.com>
Copilot AI changed the title [WIP] Please modify @NVIDIA/cuda-python/files/ci/test-matrix.json so that for all linux amd64 runners we switch to use t4 instead of l4 GPU. Set the PR title to "[DO NOT MERGE] Debug T4 runner performance". [DO NOT MERGE] Debug T4 runner performance Sep 18, 2025
Copilot AI requested a review from leofang September 18, 2025 14:21
@leofang

leofang commented Sep 18, 2025

Copy link
Copy Markdown
Member

/ok to test a63f355

@github-actions

Copy link
Copy Markdown
Contributor

@leofang leofang closed this Sep 27, 2025
@leofang
leofang deleted the copilot/fix-bac5ba47-0510-4706-ba04-b7a571b52bd1 branch September 27, 2025 02:13
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