fix: align GPU CI sharding tags with detection tags#11179
Merged
pinin4fjords merged 1 commit intomasterfrom Apr 14, 2026
Merged
fix: align GPU CI sharding tags with detection tags#11179pinin4fjords merged 1 commit intomasterfrom
pinin4fjords merged 1 commit intomasterfrom
Conversation
The detection step (line 67) uses `tags: "gpu,gpu_highmem"` to find changed GPU test files, but the sharding step (line 84) only uses `tags: "gpu"` for its nf-test dry run. This means tests tagged only with `gpu_highmem` (like parabricks) are detected but produce 0 shards, causing the GPU jobs to be skipped with "No tests to execute." Currently this is masked because parabricks tests are the only GPU tests, and the `include` matrix always generates the `gpu_highmem` runner. But any test tagged `gpu` without `gpu_highmem` would pass sharding and then silently not execute on the `gpu_highmem` runner. Aligning the sharding tags with the detection tags fixes both cases. Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
6 tasks
mashehu
approved these changes
Apr 14, 2026
oliverdrechsel
pushed a commit
to oliverdrechsel/nf-core.modules
that referenced
this pull request
Apr 14, 2026
fix: include gpu_highmem tag in GPU CI sharding step The detection step (line 67) uses `tags: "gpu,gpu_highmem"` to find changed GPU test files, but the sharding step (line 84) only uses `tags: "gpu"` for its nf-test dry run. This means tests tagged only with `gpu_highmem` (like parabricks) are detected but produce 0 shards, causing the GPU jobs to be skipped with "No tests to execute." Currently this is masked because parabricks tests are the only GPU tests, and the `include` matrix always generates the `gpu_highmem` runner. But any test tagged `gpu` without `gpu_highmem` would pass sharding and then silently not execute on the `gpu_highmem` runner. Aligning the sharding tags with the detection tags fixes both cases. Co-authored-by: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
pinin4fjords
added a commit
that referenced
this pull request
Apr 14, 2026
With #11179 merged, the sharding step uses both gpu and gpu_highmem tags. The gpu tag alone is sufficient and runs on g4dn.xlarge which is adequate for ribodetector. Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
This was referenced Apr 14, 2026
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
2 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.
Summary
One-line fix: the GPU CI sharding step uses
tags: "gpu"but the detection step usestags: "gpu,gpu_highmem". This means the shard dry-run only countsgpu-tagged tests, missinggpu_highmem-tagged tests (like parabricks).The bug
In
.github/workflows/nf-test-gpu.yml:tags: "gpu,gpu_highmem"- finds test files with either tagtags: "gpu"- only countsgpu-tagged tests in its dry runWhen only
gpu_highmem-tagged tests change (e.g. parabricks-only PRs), the sharding step finds 0 tests andtotal_shards=0, which would skip all GPU jobs. This is currently masked because the matrixincludegeneratesgpu_highmemrunner entries regardless, but the nf-test execution still runs with--tag gpu_highmemand finds the tests. However, a test tagged onlygpu(notgpu_highmem) would pass sharding but silently not execute on thegpu_highmemrunner.Fix
Change line 84 from
tags: "gpu"totags: "gpu,gpu_highmem"to match line 67.Discovered while adding GPU tests for ribodetector (see #11177, #11178).
🤖 Generated with Claude Code