Repository navigation
Conversation
|
The latest updates on your projects. Learn more about Vercel for GitHub.
|
This branch was successfully deployed
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Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be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.
Two deterministic reproductions for the AM3352 meshing blocker: edge enumeration on a dense two-layer grid and expansion of a large finite set of boundary seeds. Both stages consume one input item per step, but the inherited size-independent budget rejects them before the queues are exhausted. This PR contains tests, expected visual artifacts and benchmark tooling, with no solver fix.
The inspected expected completed-grid visual is generated with the stacked fix; the reproduction does not claim the unfixed solver completed it.
Benchmark effects: local Bun 1.4.0 / solver-utils 0.0.23 on shared 4-vCPU hardware: the original grid stops at 100,000 iterations with 3,968 edges unprocessed; the fix processes all 103,968 edges and yields exactly 912 exterior segments. Both finite-queue tests pass on the fix. Paired pinned base/head finite-queue and full canonical routing benchmark runs are configured on Blacksmith; hosted results are pending and must be reported before merging. No failure-versus-success speedup is claimed.
The board-sized all-fixed experiment advances past both limits, then exhausts memory during topology merging. This change alone does not establish AM3352 routing completion. Details:
docs/finite-gap-budget/analysis.md.Hosted benchmark (run 40192855321, Bun 1.4.0, Blacksmith 4-vCPU, 4 concurrent canonical workers): both finite cases fail on
46fe334and pass on9f6800cwith zero remaining items. Grid: 0.1290 s failing → 0.1369 s complete; expansion: 0.00866 s failing → 0.00865 s complete. These are correctness results, not failure-versus-success speedups.The repository's Pipeline4 → PortPointPathing benchmark is 202/207 on both revisions, the same five iteration-limit failures, zero timeouts and no lost completions. P50 346 → 329 ms; P95 1.9 → 1.9 s; total wall 103.9 → 100.5 s. One paired trial does not establish a general speedup. This benchmark stops at port-point pathing; it is not a final routed-board DRC result. Hosted unit CI passes for the fix. Raw results are in the workflow artifact
finite-gap-benchmark-comparison.