Repository navigation
Physical Gaussian RHS caller: paid two-component gather and return - #5483
Open
JoyboyBrian wants to merge 2 commits into
Open
JoyboyBrian wants to merge 2 commits into
JoyboyBrian wants to merge 2 commits into
Conversation
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Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be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 physical inverse-block boundary consumer required batch-major RHS input, while the actual Gaussian Core produces separate scalar component streams. This package executes the unchanged
Core.gather/Core.first_cellpair, physically packs real and imaginary RHSs, invokes the unchangedBoundaryPagesconsumer and recovers the original component/row order. Padding, scans, rewinds, protected output consumption, parked input and retained backing are charged.The source-relative proof covers all positive n/B/w for this actual R=2 layout. The P16/w32 five-row fixture executes both interior first passes and distinguishes skipping the consumer. B1 and B4 cases cover separate batches and incomplete rows/columns. Exact LU(round_P(S)) is the supplied reference; old scalar BandedSolve has different rounding. Second-cell reconstruction, normalized Core installation, production coefficient preparation and the fast seven-product kernel remain open. No new κ or selected-frontier change.
Validation:
Canonical claim: #4182 (comment)
Prepared by JoyboyBrian with substantial OpenAI Codex assistance. Credit Jose Manuel/jmbarrancoml's Core/runtime/fixture and block handoff, the physical-boundary-pages predecessor, Vlado Vojdanovski, Cinna Davis, Chafik Boukhalfa and the retained Gaussian/FFT/tape lineage. Apache-2.0; frozen sources and notices unchanged.