Latest PyTorch (2.11.0) is not compatible with current GCP GPU benchmark environment

[R-Palazzo]

Resolve #589
CU-86b9atcr2

The boot image used by the benchmark stopped being supported by Google on April 13, which caused this error:

NotFound: 404 POST https://compute.googleapis.com/compute/v1/projects/sdgym-337614/zones/us-central1-a/instances: The resource 'projects/deeplearning-platform-release/global/images/family/common-cu128-ubuntu-2204-nvidia-570' was not found

I’m sharing below a screenshot of Google’s machine image lifecycle table. Based on it, I updated the boot image to common-cu129-ubuntu-2204-nvidia-580, which is the closest supported replacement to the one we were using. This image is supported until August 4, 2028.

I tested the change by running RealTabFormer with the new image, and it worked. The results are available here.

I also added a comment to the original issue to mention the boot image update and note that it resolves the original PyTorch/CUDA version conflict simultaneously.

[sdv-team]

Task linked: CU-86b9atcr2 SDGym - Latest PyTorch (2.11.0) is not compatible with current GCP GPU benchmark environment #589

[codecov]

Codecov Report

✅ All modified and coverable lines are covered by tests.
✅ Project coverage is 85.12%. Comparing base (f1679ec) to head (c9097cd).
⚠️ Report is 1 commits behind head on main.

Additional details and impacted files

@@           Coverage Diff           @@
##             main     #595   +/-   ##
=======================================
  Coverage   85.12%   85.12%           
=======================================
  Files          40       40           
  Lines        3544     3544           
=======================================
  Hits         3017     3017           
  Misses        527      527           

Flag	Coverage Δ
integration	45.11% <ø> (ø)
unit	80.78% <ø> (ø)

Flags with carried forward coverage won't be shown. Click here to find out more.

☔ View full report in Codecov by Sentry.
📢 Have feedback on the report? Share it here.

🚀 New features to boost your workflow:

• ❄️ Test Analytics: Detect flaky tests, report on failures, and find test suite problems.