Repository navigation
Revert #5089: agc_driver_waterfall_images fails SPIR-V validation on main - #5824
Merged
Merged
Conversation
…piler-waterfall-descriptors" agc_driver_waterfall_images fails on main since fd438f6: its compute shader runs an LOD query with a local size of 8x1, and the module gets DerivativeGroupQuadsKHR, which SPIR-V validation rejects unless both dimensions are even (VUID-StandaloneSpirv-DerivativeGroupQuadsKHR-10151). This reverts commit fd438f6, reversing changes made to 5103174. Co-Authored-By: Claude <noreply@anthropic.com>
11 tasks done
11 tasks done
11 tasks done
12 tasks done
fdfawef324
added a commit
to fdfawef324/AnyPS5
that referenced
this pull request
Oct 19, 2026
… tables A waterfall loop over a T# that each lane loaded with IDXEN buffer loads (s_ff1/v_readfirstlane, v_readlane of the eight dwords, v_cmpx_eq against the lane's VGPRs, the image instruction) stopped planning with "GetImageResource dword 0 is not a valid runtime value". Bring back the fold of boykopovar#5089 (reverted by boykopovar#5824 for a derivative-group validation error that boykopovar#5829 fixed) on the image table model of boykopovar#757: the T# dwords are the lane's own VGPRs under the loop's EXEC, and the access reads an image table column keyed by the lane's buffer index, with the V#'s stride and NUM_RECORDS addressing of the recompiler's buffer loads, a record count in the table header, NonUniform slots, and whole-table faults for V# layouts and for mixed image modes under implicit derivatives. Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com> Co-Authored-By: Codex <noreply@openai.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.
What
Reverts #5089. Since its merge (fd438f6) every main build fails
agc_driver_waterfall_imageson Linux:The test's compute shader does an LOD query with a local size of 8x1, and
SpirvModuleSetup.cppdeclaresDerivativeGroupQuadsKHRfor every compute shader with derivatives. CI's-DANYPS5_ENABLE_SPIRV_TOOLS=ONvalidation rejects it. Main runs: 38560929488, 38562879025, 38563969869.The PR can come back with the derivative group chosen from the local size (or the test on an even 2D size).
Tested
Full Build dispatched with
linux=trueon this branch; the last green main build (968d326) was right before #5089.Checklist
main; no other open PR implements the same functionsNotImplemented_nid_no_patch); silent stubs are listed in TechnicalDebt.mdfiles or images added to the repository