Add distributed Jacobi preconditioning#68
Open
sanjanag wants to merge 3 commits intolinkedin:masterfrom
Open
Conversation
dist_jacobi_precondition reconstructs full row L2-norms from a column-partitioned matrix via all_reduce(SUM) over per-rank partial sums of squares, then scales each rank's local slice in place. Rank 0 optionally scales b and persists the row norms. Includes a no-dist guard test plus two torchrun-gated tests covering correctness of the global norms / A scaling / b scaling and the norm-save path. Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
Co-Authored-By: Claude Opus 4.7 (1M context) <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Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be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.
Summary
Adds
dist_jacobi_preconditionfor column-partitioned constraint matrices,enabling Jacobi preconditioning in multi-node / multi-GPU settings. Each rank
holds a column slice; full row L2-norms are reconstructed via an
all_reduce(SUM)over per-rank partial sums of squares, then each rank scalesits local slice in place. Rank 0 optionally scales
band persists the rownorms for later inversion.
Test plan
torchrun-gated tests cover correctness of global norms, A scaling,b scaling, and the norm-save path (skipped without torchrun + 2 GPUs)