Repository navigation
Upload GPUCompiler's failure dumps from the SPIR-V CI jobs - #207
Merged
Merged
Conversation
When a kernel fails to compile or its SPIR-V fails to validate, GPUCompiler writes the module to a file to attach to a bug report. On CI that file was lost with the runner, which left a one-off SPIR-V validation failure in the host-kernel job impossible to debug. Upload these files as an artifact when the OpenCL or host-kernel jobs fail, both from where JuliaGPU/GPUCompiler.jl#1077 puts them and from where released versions do.
maleadt
force-pushed
the
tb/ci-compilation-dumps
branch
from
November 5, 2026 11:39
bd3bff4 to
d927755
Compare
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Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be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.
The host-kernel job on #197 failed once with
Failed to validate generated SPIR-V: ID '297[%297]' has not been defined, insort: settings(log). GPUCompiler wrote the failing.bcand.spvto/tmpon the runner, and they were lost with it.I couldn't reproduce the failure. With CI's exact package versions, the SPIR-V back-end gives byte-identical, valid output for all 213 modules that
sort.jlcompiles. That holds across thousands of repeated compiles, separate processes, malloc perturbation, and CI's test order on its worker. So whatever went wrong depends on the exact bitcode that CI sent the back-end, and only the dumped file can show it.This uploads those dumps as an artifact when the OpenCL or host-kernel jobs fail. JuliaGPU/GPUCompiler.jl#1077 moves them to
$RUNNER_TEMP/gpucompiler-dumps. Released versions still write/tmp/jl_*.bcand/tmp/jl_*.spv, so the step collects both locations.