ci: Re-add GNN test hashes - #5997
Merged
benjaminhuth merged 1 commit intoSep 1, 2026
Merged
Conversation
github-actions
Bot
added
Infrastructure
benjaminhuth
force-pushed
the
ci/gnn-add-hash-checks-again
branch
from
September 1, 2026 11:22
6c5a1b1 to
f23456c
Compare
benjaminhuth
added a commit
to benjaminhuth/acts
that referenced
this pull request
Sep 1, 2026
664c82a (acts-project#5903) dropped the GNN GPU reference hashes because the heterogeneous GPU pool (T4 + A4000) produced a different ntuple_finding_gnn.root on the A4000 for test_gnn_module_map[gpu-onnx], blamed on ONNX Runtime enabling TF32 GEMMs on Ampere by default. Confirmed experimentally in acts-project#5997: the T4 reproduces the reference hashes with or without TF32 (no TF32 hardware to begin with), and forcing NVIDIA_TF32_OVERRIDE=0 makes the A4000 reproduce them too, on all four gpu-marked GNN tests. No other source of drift was found, so the existing reference hashes did not need updating. Restore the six removed root_file_hashes.txt entries and the assert_root_hash calls in test_gnn_metric_learning and test_gnn_module_map, and set NVIDIA_TF32_OVERRIDE=0 in the gnn_gpu job env. Co-Authored-By: Claude Opus 5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01X1p4KwFbkMWfy4zt6TCyfT
benjaminhuth
force-pushed
the
ci/gnn-add-hash-checks-again
branch
from
September 1, 2026 12:56
f23456c to
9ef8fad
Compare
benjaminhuth
added a commit
to benjaminhuth/acts
that referenced
this pull request
Sep 1, 2026
664c82a (acts-project#5903) dropped the GNN GPU reference hashes because the heterogeneous GPU pool (T4 + A4000) produced a different ntuple_finding_gnn.root on the A4000 for test_gnn_module_map[gpu-onnx], blamed on ONNX Runtime enabling TF32 GEMMs on Ampere by default. Confirmed experimentally in acts-project#5997: the T4 reproduces the reference hashes with or without TF32 (no TF32 hardware to begin with), and forcing NVIDIA_TF32_OVERRIDE=0 makes the A4000 reproduce them too, on all four gpu-marked GNN tests. No other source of drift was found, so the existing reference hashes did not need updating. Restore the six removed root_file_hashes.txt entries and the assert_root_hash calls in test_gnn_metric_learning and test_gnn_module_map, and set NVIDIA_TF32_OVERRIDE=0 in the gnn_gpu job env. Co-Authored-By: Claude Opus 5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01X1p4KwFbkMWfy4zt6TCyfT
benjaminhuth
force-pushed
the
ci/gnn-add-hash-checks-again
branch
from
September 1, 2026 12:58
9ef8fad to
88bd31a
Compare
664c82a (acts-project#5903) dropped the GNN GPU reference hashes because the heterogeneous GPU pool (T4 + A4000) produced a different ntuple_finding_gnn.root on the A4000 for test_gnn_module_map[gpu-onnx], blamed on ONNX Runtime enabling TF32 GEMMs on Ampere by default. Confirmed experimentally in acts-project#5997: the T4 reproduces the reference hashes with or without TF32 (no TF32 hardware to begin with), and forcing NVIDIA_TF32_OVERRIDE=0 makes the A4000 reproduce them too, on all four gpu-marked GNN tests. No other source of drift was found, so the existing reference hashes did not need updating. Restore the six removed root_file_hashes.txt entries and the assert_root_hash calls in test_gnn_metric_learning and test_gnn_module_map, and set NVIDIA_TF32_OVERRIDE=0 in the gnn_gpu job env. Co-Authored-By: Claude Opus 5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01X1p4KwFbkMWfy4zt6TCyfT
benjaminhuth
force-pushed
the
ci/gnn-add-hash-checks-again
branch
from
September 1, 2026 12:59
88bd31a to
39bd368
Compare
paulgessinger
changed the title
paulgessinger
approved these changes
Sep 1, 2026
benjaminhuth
enabled auto-merge
September 1, 2026 13:06
benjaminhuth
changed the title
Contributor
Public API surface diffNo change to the public API surface. ✅ |
Contributor
|
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
3 participants
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
Restore GNN hashes, which are reproducible again with TF32 disabled