Repository navigation
Make failure dumps downloadable from CI - #1077
Merged
Merged
Conversation
When a back-end or SPIR-V validation/optimization fails, the module is written to a temporary file to attach to a bug report. On CI that file disappears with the runner, so a failure that doesn't reproduce locally can't be debugged. Write these dumps through one helper that, on GitHub Actions, puts them in `$RUNNER_TEMP/gpucompiler-dumps` (with a notice) for an `actions/upload-artifact` step to collect, and on Buildkite uploads them as artifacts. The input and output of a failed SPIR-V step now share a base name. Unlike the dump removed in #861, this only runs once a compilation has failed, so it costs nothing otherwise.
Codecov Report❌ Patch coverage is
Additional details and impacted files@@ Coverage Diff @@
## main #1077 +/- ##
==========================================
+ Coverage 92.65% 92.74% +0.09%
==========================================
Files 31 31
Lines 7501 7514 +13
==========================================
+ Hits 6950 6969 +19
+ Misses 551 545 -6 ☔ View full report in Codecov by Harness. 🚀 New features to boost your workflow:
|
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
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Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be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.
When a back-end compilation or a SPIR-V validation/optimization step fails, GPUCompiler writes the module to a temporary file and asks the user to attach it to a bug report. On CI that file disappears with the runner. That's what happened with a one-off
Failed to validate generated SPIR-V: ID '297[%297]' has not been definedin AcceleratedKernels.jl's CI. Locally, the SPIR-V back-end produces identical, valid output for every module of that test, so only the dumped bitcode could explain the failure.This routes those dumps through a single
dump_failurehelper:$RUNNER_TEMP/gpucompiler-dumps, with a::notice, so that anactions/upload-artifactstep can collect them;buildkite-agent artifact upload;tempdir()as before.The
.bcand.spvfiles of a failed SPIR-V step now share a base name. The mechanism follows the one from 0d5bdc4, which #861 removed because it serialized every module before optimization. This version only runs once a compilation has failed, so it costs nothing otherwise.