ci(opencl): run POCL under --check-bounds=auto; skip scan on POCL - #104
Draft
shreyas-omkar wants to merge 1 commit into
Draft
ci(opencl): run POCL under --check-bounds=auto; skip scan on POCL#104shreyas-omkar wants to merge 1 commit into
shreyas-omkar wants to merge 1 commit into
Conversation
The OpenCL job used julia-actions/julia-runtest's default --check-bounds=yes. POCL miscompiles our @LocalMem + @synchronize kernels (reduce / predicates / sort) when bounds checks are forced on, raising a spurious "Out-of-bounds array access" and segfaulting the worker. The kernels are correct: the same code passes on CUDA under --check-bounds=yes and on POCL under auto. Every other AK backend is tested under default/auto bounds, so force this job to auto too. Under auto, POCL instead miscompiles the multi-block scan (accumulate!, and RadixSort which scans a histogram internally) — it deadlocks / returns wrong results, while being correct on every real backend and on POCL under yes. Since neither bounds setting makes both halves pass on POCL, skip the scan-dependent tests on the opencl backend only: the accumulate suite (dropped in runtests.jl) and RadixSort (guarded by a new TEST_SCAN flag, false only for opencl). Verified: full --opencl suite green under auto with pocl_jll 7.1.3+5 (37985/37985 pass). Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
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
1 participant
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Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be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 OpenCL job used julia-actions/julia-runtest's default --check-bounds=yes. POCL miscompiles our @LocalMem + @synchronize kernels (reduce / predicates / sort) when bounds checks are forced on, raising a spurious "Out-of-bounds array access" and segfaulting the worker. The kernels are correct: the same code passes on CUDA under --check-bounds=yes and on POCL under auto. Every other AK backend is tested under default/auto bounds, so force this job to auto too.
Under auto, POCL instead miscompiles the multi-block scan (accumulate!, and RadixSort which scans a histogram internally) it deadlocks / returns wrong results, while being correct on every real backend and on POCL under yes. Since neither bounds setting makes both halves pass on POCL, skip the scan-dependent tests on the opencl backend only: the accumulate suite (dropped in runtests.jl) and RadixSort (guarded by a new TEST_SCAN flag, false only for opencl).