Integrate fixed-topology portal-layer refinement - #1928
Conversation
|
The latest updates on your projects. Learn more about Vercel for GitHub.
|
Benchmark This PRRun benchmarks by commenting on this PR: Comment Everything after Use Any PR whose title contains |
|
/benchmark-all |
Default Benchmark ResultsPrevious main runDataset: dataset01
Previous main run details
Dataset: dataset01
PR run details
Workflow: View run |
srj18 ResultsPrevious main runDataset: srj18
Previous main run details
Dataset: srj18
PR run details
Workflow: View run |
srj19 ResultsPrevious main runDataset: srj19
Previous main run details
Dataset: srj19
PR run details
Workflow: View run |
srj20 ResultsPrevious main runDataset: srj20
Previous main run details
Dataset: srj20
PR run details
Workflow: View run |
srj21 ResultsPrevious main runDataset: srj21
Previous main run details
Dataset: srj21
PR run details
Workflow: View run |
srj23 ResultsPrevious main runDataset: srj23
Previous main run details
Dataset: srj23
PR run details
Workflow: View run |
|
/benchmark-all |
srj18 ResultsPrevious main runDataset: srj18
Previous main run details
Dataset: srj18
PR run details
Workflow: View run |
Default Benchmark ResultsPrevious main runDataset: dataset01
Previous main run details
Dataset: dataset01
PR run details
Workflow: View run |
srj19 ResultsPrevious main runDataset: srj19
Previous main run details
Dataset: srj19
PR run details
Workflow: View run |
srj20 ResultsPrevious main runDataset: srj20
Previous main run details
Dataset: srj20
PR run details
Workflow: View run |
srj21 ResultsPrevious main runDataset: srj21
Previous main run details
Dataset: srj21
PR run details
Workflow: View run |
srj23 ResultsPrevious main runDataset: srj23
Previous main run details
Dataset: srj23
PR run details
Workflow: View run |
|
/benchmark-all |
srj18 ResultsPrevious main runDataset: srj18
Previous main run details
Dataset: srj18
PR run details
Workflow: View run |
Default Benchmark ResultsPrevious main runDataset: dataset01
Previous main run details
Dataset: dataset01
PR run details
Workflow: View run |
srj19 ResultsPrevious main runDataset: srj19
Previous main run details
Dataset: srj19
PR run details
Workflow: View run |
srj20 ResultsPrevious main runDataset: srj20
Previous main run details
Dataset: srj20
PR run details
Workflow: View run |
srj21 ResultsPrevious main runDataset: srj21
Previous main run details
Dataset: srj21
PR run details
Workflow: View run |
srj23 ResultsPrevious main runDataset: srj23
Previous main run details
Dataset: srj23
PR run details
Workflow: View run |
|
/benchmark --dataset srj18 --sample-numbers 6 --sample-timeout 600s --concurrency 1 |
srj18 FailedBenchmark workflow ended with cancelled before completion. This can happen during setup, dependency installation, or benchmark execution. Previous main runDataset: srj18
Previous main run details
Workflow: View run |
|
/benchmark --dataset srj18 --sample-numbers 6 --sample-timeout 600s --concurrency 1 |
srj18 FailedBenchmark workflow ended with cancelled before completion. This can happen during setup, dependency installation, or benchmark execution. Previous main runDataset: srj18
Previous main run details
Workflow: View run |
srj18 ResultsPrevious main runDataset: srj18
Previous main run details
Paired comparison
PR refinement acceptedCandidateCount: 0.00 Portal-layer refinement diagnostics
Dataset: srj18
PR run details
Workflow: View run |
srj18 ResultsPrevious main runDataset: srj18
Previous main run details
Paired comparison
PR refinement acceptedCandidateCount: 3.00 Portal-layer refinement diagnostics
Dataset: srj18
PR run details
Workflow: View run |
|
This PR has been automatically marked as stale because it has had no recent activity. It will be closed if no further activity occurs. |
|
This PR was closed because it has been inactive for 1 day since being marked as stale. |
Summary
physicalPortGroupIdand serialize it into tiny-hypergraph.tiny-hypergraphto9cc3a682a8abf8ca6fd20144140716273f2f11ae(companion draft: Add fixed-topology portal-layer refinement tiny-hypergraph#152).Current validation
bun test tests/benchmark-portal-refinement-stats.test.ts tests/features/pipeline9-minimal-preloaded-trace-graph.test.ts: 2 passed, 0 failed, 77 expectationsbunx tsc --noEmitnode --checkdataset01/sample1benchmark worker output includes the exact diagnostics in both the result and snapshot JSONgit diff --checkbugreport80 diagnosis
The failed CI rendering was not an accepted refiner change:
acceptedCandidateCountis 0 and both outputs contain 76 vias. The CI-produced SVG is byte-for-byte the originalmainsnapshot, while the prior expected snapshot was generated locally under Bun 1.3.9 and CI runs Bun 1.3.8. The branch now restores the original snapshot instead of blessing an unrelated rendering difference.Benchmark evidence and current gate
The previous
/benchmark-allartifacts were mostly paired no-ops. The only apparent final-via change (srj20/sample43) also regressed relaxed DRC and is not reproducible locally: current branch and baseline both finish with 78 vias and zero relaxed-DRC errors while the refiner accepts zero candidates.A targeted Blacksmith run for the real accepted-candidate case
srj18/sample6is running on commit00bcac3644b52db1b55cdcff754eb36691dfc83b: https://github.com/tscircuit/tscircuit-autorouter/actions/runs/31058429519The expensive full
/benchmark-allgate will be rerun only if targeted evidence shows a genuine final improvement without completion or DRC regression.