perf: release failed high-density route searches by ShiboSoftwareDev · Pull Request #2894 · tscircuit/tscircuit-autorouter

ShiboSoftwareDev · 2026-10-06T20:57:38Z

What changed

Production routing disables search-debug capture, but a failed intra-node grid search was still retained by its portfolio candidate. Dataset 18 sample 002 accumulates 62 failed candidates before another candidate solves the node.

This preserves the failure message and retains the failed solver only when search-debug capture is enabled.

Results

Seven alternating fresh-process A/B pairs using the real Dataset 18 sample-002 large node with the production configuration:

Metric	Main	PR	Change
Peak heap	287.5 MiB	198.2 MiB	-31.0%
Retained heap	117.6 MiB	91.7 MiB	-22.0%
Process peak	601.0 MiB	485.4 MiB	-19.2%
Runtime	3.386 s	3.362 s	-0.7%

Every run solved in 97,190 iterations with the same output hash. A forced-GC heap snapshot also showed 126,401 fewer retained objects and 17.0 MiB less collection backing storage.

The repository's same-machine Dataset 18 benchmark also reported Memory P50 -15.2%, P80 -8.8%, and P90 -0.7%, with no outcome, DRC, timeout, via, or trace changes. Per-sample process peaks remained noisy (8 lower and 8 higher), so the repeated large-node A/B result above is the direct proof of the lifecycle improvement.

Validation

bun run build
bunx tsc --noEmit
19 focused tests, including the Dataset 18 large-node repro and failed-search retention assertion

vercel · 2026-10-06T20:57:45Z

The latest updates on your projects. Learn more about Vercel for GitHub.

Project	Deployment	Actions	Updated
capacity-node-autorouter	Ready	Preview	Oct 6, 2026 11:49pm UTC

tscircuitbot · 2026-10-06T20:57:48Z

Benchmark This PR

Run benchmarks by commenting on this PR:

/benchmark [benchmark.sh args...]
/benchmark-long [benchmark.sh args...]
/benchmark-effort
/benchmark-all [benchmark.sh args except --dataset]
/benchmark --same-machine
/benchmark-all --same-machine
/profile --dataset 18

Comment /benchmark to compare the PR base and head sequentially on the same 8-vCPU Blacksmith runner using the default dataset, or append any arguments accepted by ./benchmark.sh. Same-machine comparison is enabled by default, except for --pipeline 9net, which uses its dedicated cold/hot benchmark workflow.
Comment /benchmark-effort to compare 1x, 1.5x, and 2x effort on dataset18, including matched-sample via counts and scaled timeouts.
Comment /benchmark-long for an 8-vCPU run that defaults to 8 workers and has an eight-hour timeout.
Every benchmark runs each sample in a fresh process and reports Memory P50, P80, and P90.
Comment /benchmark-all to start separate workflow runs and result comments for dataset01 and dataset18. You may append benchmark arguments such as --pipeline 9; --dataset is not accepted because the command selects both datasets.
Append --same-machine to /benchmark-long or /benchmark-all to compare the PR base and head sequentially on the same 8-vCPU Blacksmith runner.
Comment /profile --dataset 18 to compare direct Pipeline 7 stage-time percentages between current main and the PR head sequentially on one Blacksmith runner.

Everything after /benchmark, /benchmark-long, or /benchmark-all is safely forwarded to ./benchmark.sh, except --profile-solvers, which enables profile comparison tables, and --same-machine, which enables the paired runner.
Examples: /benchmark --dataset 18 --sample-timeout 2000s, /benchmark --pipeline 10 --dataset 29, /benchmark --pipeline 10 --dataset 29 --sample 1, /benchmark --pipeline 7 --limit 20, /benchmark all 20 --concurrency 8, /benchmark --same-machine, /benchmark-all --pipeline 9 --same-machine, /benchmark-long --dataset 18, and /benchmark-long --same-machine --pipeline 9 --dataset 24 --sample-numbers 21 --sample-timeout 240s --concurrency 1.

Use /update-snapshots (or /us) to run BUN_UPDATE_SNAPSHOTS=1 bun test --timeout 120_000 on the PR branch and auto-commit snapshot updates.
Use /usf to read recent failed test files, update and verify their exact CI test shards, and auto-commit only their snapshots. It uses the configured fast benchmark runner by default; use /usf --ubuntu-latest for GitHub-hosted x64 CI parity.

Any PR whose title contains [BENCHMARK TEST] will automatically run one default-dataset benchmark on PR updates; it does not post a PR result comment.

ShiboSoftwareDev · 2026-10-06T21:13:29Z

/benchmark --same-machine

ShiboSoftwareDev · 2026-10-06T21:13:33Z

/benchmark --same-machine --dataset 18

ShiboSoftwareDev · 2026-10-06T21:13:40Z

/benchmark --same-machine --dataset 24

tscircuitbot · 2026-10-06T21:14:12Z

Same Machine Benchmark Results

Both revisions ran sequentially in one Blacksmith job on blacksmith-8vcpu-ubuntu-2404-arm-w2gkzeqqcp56zrc4.

Dataset: dataset01 · Scenarios: 85
Base: 374f3b5 · PR: b1d54c9

Solver	Metric	Base	PR	Delta
Pipeline9	Completion	100.0%	100.0%	0.0 pp
Pipeline9	Relaxed DRC pass	100.0%	100.0%	0.0 pp
Pipeline9	DRC issues	0	0	0
Pipeline9	Timeouts	0	0	0
Pipeline9	P50 time	4.7s	4.9s	+4.0%
Pipeline9	P60 time	6.3s	6.3s	-0.2%
Pipeline9	P70 time	7.6s	7.4s	-2.8%
Pipeline9	P80 time	9.4s	9.1s	-3.2%
Pipeline9	P90 time	13.1s	13.5s	+2.8%
Pipeline9	P95 time	15.6s	15.7s	+1.0%
Pipeline9	Memory P50	367.9 MiB	367.1 MiB	-0.2%
Pipeline9	Memory P80	468.0 MiB	483.1 MiB	+3.2%
Pipeline9	Memory P90	570.9 MiB	549.6 MiB	-3.7%
Pipeline9	Average vias	38.99	38.99	0.0%
Pipeline9	Avg Angled Traces	7.44	7.44	0.0%

Outcome changes: 0 improved, 0 regressed. DRC issues are totaled across solved samples. Timing percentiles include all samples, with failed and timed-out samples counted at their configured timeout; negative timing deltas are faster. Historical failures without timeout metadata make timing percentiles unavailable.
Style errors are averaged per completed sample with recorded lint counts for that type; historical/unlinted results are n/a. Angled traces counts violating segments.

Base pipeline stage timings

AutoroutingPipelineSolver9_PreloadedTraceGraph

Recorded timings: 85/85 samples (0 partial, including failed or timed-out samples).
Times are summed across samples; percentages use the total recorded stage time for this solver.

Stage	Total time	% of stage time
preprocessSimpleRouteJsonSolver	0.630s	0.1%
componentDetectionSolver	0.442s	0.1%
escapeViaLocationSolver	0.273s	0.1%
netToPointPairsSolver	0.311s	0.1%
topologyPlanningSolver	21.460s	4.8%
topologyMergingSolver	5.616s	1.3%
nodeDimensionSubdivisionSolver	0.237s	0.1%
edgeSolver	3.367s	0.8%
availableSegmentPointSolver	1.532s	0.3%
necessaryCrampedPortPointSolver	4.108s	0.9%
preloadedTraceGraphSolver	0.306s	0.1%
portPointPathingSolver	62.363s	14.0%
uniformPortDistributionSolver	1.320s	0.3%
highDensityRouteSolver	228.656s	51.5%
highDensityForceImproveSolver	35.819s	8.1%
highDensityRepairSolver	13.967s	3.1%
highDensityStitchSolver	2.663s	0.6%
traceSimplificationSolver	49.653s	11.2%
mutatedPreloadedTraceSimplificationSolver	0.701s	0.2%
traceWidthSolver	1.333s	0.3%
globalDrcForceImproveSolver	4.193s	0.9%
pipeline9JointDrcRepairSolver	3.650s	0.8%
effortCleanupSolver	0.008s	0.0%
lengthMatchingPostProcessingSolver	1.307s	0.3%
powerTraceExpansionSolver	0.004s	0.0%
Total	443.919s	100.0%

PR pipeline stage timings

AutoroutingPipelineSolver9_PreloadedTraceGraph

Recorded timings: 85/85 samples (0 partial, including failed or timed-out samples).
Times are summed across samples; percentages use the total recorded stage time for this solver.

Stage	Total time	% of stage time
preprocessSimpleRouteJsonSolver	0.605s	0.1%
componentDetectionSolver	0.504s	0.1%
escapeViaLocationSolver	0.309s	0.1%
netToPointPairsSolver	0.313s	0.1%
topologyPlanningSolver	21.477s	4.9%
topologyMergingSolver	5.512s	1.3%
nodeDimensionSubdivisionSolver	0.293s	0.1%
edgeSolver	3.101s	0.7%
availableSegmentPointSolver	1.196s	0.3%
necessaryCrampedPortPointSolver	4.100s	0.9%
preloadedTraceGraphSolver	0.259s	0.1%
portPointPathingSolver	59.594s	13.6%
uniformPortDistributionSolver	1.113s	0.3%
highDensityRouteSolver	226.083s	51.5%
highDensityForceImproveSolver	35.266s	8.0%
highDensityRepairSolver	15.340s	3.5%
highDensityStitchSolver	2.358s	0.5%
traceSimplificationSolver	50.631s	11.5%
mutatedPreloadedTraceSimplificationSolver	0.723s	0.2%
traceWidthSolver	1.372s	0.3%
globalDrcForceImproveSolver	4.516s	1.0%
pipeline9JointDrcRepairSolver	2.952s	0.7%
effortCleanupSolver	0.003s	0.0%
lengthMatchingPostProcessingSolver	1.473s	0.3%
powerTraceExpansionSolver	0.004s	0.0%
Total	439.095s	100.0%

Workflow: View run
Artifact: Download raw reports

tscircuitbot · 2026-10-06T21:14:15Z

Same Machine Benchmark Results

Both revisions ran sequentially in one Blacksmith job on blacksmith-8vcpu-ubuntu-2404-arm-q3vb3oaqckb25gg3.

Dataset: srj18 · Scenarios: 16
Base: 374f3b5 · PR: b1d54c9

Solver	Metric	Base	PR	Delta
Pipeline9	Completion	87.5% (🕒12.5%)	87.5% (🕒12.5%)	0.0 pp
Pipeline9	Relaxed DRC pass	87.5% (🕒12.5%)	87.5% (🕒12.5%)	0.0 pp
Pipeline9	DRC issues	0	0	0
Pipeline9	Timeouts	2	2	0
Pipeline9	P50 time	90.1s	80.9s	-10.2%
Pipeline9	P60 time	135.1s	127.8s	-5.4%
Pipeline9	P70 time	157.0s	149.0s	-5.1%
Pipeline9	P80 time	279.3s	258.1s	-7.6%
Pipeline9	P90 time	340.2s	340.3s	+0.0%
Pipeline9	P95 time	360.0s	360.0s	0.0%
Pipeline9	Memory P50	1553.5 MiB	1611.8 MiB	+3.8%
Pipeline9	Memory P80	2945.3 MiB	3025.5 MiB	+2.7%
Pipeline9	Memory P90	4567.1 MiB	3406.9 MiB	-25.4%
Pipeline9	Average vias	217.57	217.57	0.0%
Pipeline9	Avg Angled Traces	529.86	529.86	0.0%

Outcome changes: 0 improved, 0 regressed. DRC issues are totaled across solved samples. Timing percentiles include all samples, with failed and timed-out samples counted at their configured timeout; negative timing deltas are faster. Historical failures without timeout metadata make timing percentiles unavailable.
Style errors are averaged per completed sample with recorded lint counts for that type; historical/unlinted results are n/a. Angled traces counts violating segments.

Base pipeline stage timings

AutoroutingPipelineSolver9_PreloadedTraceGraph

Recorded timings: 16/16 samples (2 partial, including failed or timed-out samples).
Times are summed across samples; percentages use the total recorded stage time for this solver.

Stage	Total time	% of stage time
preprocessSimpleRouteJsonSolver	3.761s	0.2%
componentDetectionSolver	0.233s	0.0%
escapeViaLocationSolver	0.835s	0.0%
netToPointPairsSolver	0.992s	0.0%
topologyPlanningSolver	24.395s	1.1%
topologyMergingSolver	23.392s	1.1%
nodeDimensionSubdivisionSolver	0.427s	0.0%
edgeSolver	5.238s	0.2%
availableSegmentPointSolver	1.591s	0.1%
necessaryCrampedPortPointSolver	5.166s	0.2%
preloadedTraceGraphSolver	0.139s	0.0%
portPointPathingSolver	451.218s	20.9%
uniformPortDistributionSolver	8.804s	0.4%
highDensityRouteSolver	838.650s	38.9%
highDensityForceImproveSolver	48.438s	2.2%
highDensityRepairSolver	122.410s	5.7%
highDensityStitchSolver	2.431s	0.1%
traceSimplificationSolver	165.160s	7.7%
mutatedPreloadedTraceSimplificationSolver	0.461s	0.0%
traceWidthSolver	30.914s	1.4%
globalDrcForceImproveSolver	57.121s	2.7%
pipeline9JointDrcRepairSolver	362.837s	16.8%
effortCleanupSolver	0.001s	0.0%
lengthMatchingPostProcessingSolver	0.769s	0.0%
powerTraceExpansionSolver	0.001s	0.0%
Total	2155.386s	100.0%

PR pipeline stage timings

AutoroutingPipelineSolver9_PreloadedTraceGraph

Recorded timings: 16/16 samples (2 partial, including failed or timed-out samples).
Times are summed across samples; percentages use the total recorded stage time for this solver.

Stage	Total time	% of stage time
preprocessSimpleRouteJsonSolver	3.565s	0.2%
componentDetectionSolver	0.185s	0.0%
escapeViaLocationSolver	0.670s	0.0%
netToPointPairsSolver	0.873s	0.0%
topologyPlanningSolver	21.235s	1.0%
topologyMergingSolver	20.554s	1.0%
nodeDimensionSubdivisionSolver	0.428s	0.0%
edgeSolver	3.982s	0.2%
availableSegmentPointSolver	1.348s	0.1%
necessaryCrampedPortPointSolver	3.586s	0.2%
preloadedTraceGraphSolver	0.110s	0.0%
portPointPathingSolver	437.612s	21.3%
uniformPortDistributionSolver	9.064s	0.4%
highDensityRouteSolver	784.221s	38.1%
highDensityForceImproveSolver	52.530s	2.6%
highDensityRepairSolver	108.653s	5.3%
highDensityStitchSolver	1.864s	0.1%
traceSimplificationSolver	159.330s	7.7%
mutatedPreloadedTraceSimplificationSolver	0.518s	0.0%
traceWidthSolver	30.833s	1.5%
globalDrcForceImproveSolver	55.184s	2.7%
pipeline9JointDrcRepairSolver	358.855s	17.5%
effortCleanupSolver	0.001s	0.0%
lengthMatchingPostProcessingSolver	0.713s	0.0%
powerTraceExpansionSolver	0.001s	0.0%
Total	2055.916s	100.0%

Workflow: View run
Artifact: Download raw reports

tscircuitbot · 2026-10-06T21:14:25Z

Same Machine Benchmark Results

Both revisions ran sequentially in one Blacksmith job on blacksmith-8vcpu-ubuntu-2404-arm-qs6mctfvyvzwgkyb.

Dataset: srj24 · Scenarios: 26
Base: 374f3b5 · PR: b1d54c9

Solver	Metric	Base	PR	Delta
Pipeline9	Completion	57.7% (🕒38.5%)	57.7% (🕒38.5%)	0.0 pp
Pipeline9	Relaxed DRC pass	53.8% (🕒38.5%)	53.8% (🕒38.5%)	0.0 pp
Pipeline9	DRC issues	2	2	0
Pipeline9	Timeouts	10	10	0
Pipeline9	P50 time	152.5s	150.8s	-1.1%
Pipeline9	P60 time	360.0s	360.0s	0.0%
Pipeline9	P70 time	360.0s	360.0s	0.0%
Pipeline9	P80 time	360.0s	360.0s	0.0%
Pipeline9	P90 time	360.0s	360.0s	0.0%
Pipeline9	P95 time	360.0s	360.0s	0.0%
Pipeline9	Memory P50	2046.0 MiB	2246.8 MiB	+9.8%
Pipeline9	Memory P80	4228.0 MiB	4476.1 MiB	+5.9%
Pipeline9	Memory P90	4989.1 MiB	4823.2 MiB	-3.3%
Pipeline9	Average vias	123.40	123.40	0.0%
Pipeline9	Avg Angled Traces	252.07	252.07	0.0%

Outcome changes: 0 improved, 0 regressed. DRC issues are totaled across solved samples. Timing percentiles include all samples, with failed and timed-out samples counted at their configured timeout; negative timing deltas are faster. Historical failures without timeout metadata make timing percentiles unavailable.
Style errors are averaged per completed sample with recorded lint counts for that type; historical/unlinted results are n/a. Angled traces counts violating segments.

Base pipeline stage timings

AutoroutingPipelineSolver9_PreloadedTraceGraph

Recorded timings: 26/26 samples (11 partial, including failed or timed-out samples).
Times are summed across samples; percentages use the total recorded stage time for this solver.

Stage	Total time	% of stage time
preprocessSimpleRouteJsonSolver	48.177s	1.1%
componentDetectionSolver	0.515s	0.0%
escapeViaLocationSolver	2.759s	0.1%
netToPointPairsSolver	3.522s	0.1%
topologyPlanningSolver	90.748s	2.0%
topologyMergingSolver	505.029s	11.2%
nodeDimensionSubdivisionSolver	4.988s	0.1%
edgeSolver	20.243s	0.4%
availableSegmentPointSolver	6.714s	0.1%
necessaryCrampedPortPointSolver	18.136s	0.4%
preloadedTraceGraphSolver	0.372s	0.0%
portPointPathingSolver	916.926s	20.3%
uniformPortDistributionSolver	118.679s	2.6%
highDensityRouteSolver	2136.857s	47.4%
highDensityForceImproveSolver	57.660s	1.3%
highDensityRepairSolver	83.795s	1.9%
highDensityStitchSolver	2.792s	0.1%
traceSimplificationSolver	217.931s	4.8%
mutatedPreloadedTraceSimplificationSolver	1.193s	0.0%
traceWidthSolver	55.581s	1.2%
globalDrcForceImproveSolver	53.902s	1.2%
pipeline9JointDrcRepairSolver	165.193s	3.7%
effortCleanupSolver	0.001s	0.0%
lengthMatchingPostProcessingSolver	0.658s	0.0%
powerTraceExpansionSolver	0.001s	0.0%
Total	4512.372s	100.0%

PR pipeline stage timings

AutoroutingPipelineSolver9_PreloadedTraceGraph

Recorded timings: 26/26 samples (11 partial, including failed or timed-out samples).
Times are summed across samples; percentages use the total recorded stage time for this solver.

Stage	Total time	% of stage time
preprocessSimpleRouteJsonSolver	50.728s	1.1%
componentDetectionSolver	0.482s	0.0%
escapeViaLocationSolver	2.491s	0.1%
netToPointPairsSolver	3.086s	0.1%
topologyPlanningSolver	92.138s	2.0%
topologyMergingSolver	479.806s	10.5%
nodeDimensionSubdivisionSolver	6.575s	0.1%
edgeSolver	20.677s	0.5%
availableSegmentPointSolver	5.485s	0.1%
necessaryCrampedPortPointSolver	14.660s	0.3%
preloadedTraceGraphSolver	0.289s	0.0%
portPointPathingSolver	951.737s	20.9%
uniformPortDistributionSolver	119.134s	2.6%
highDensityRouteSolver	2098.017s	46.1%
highDensityForceImproveSolver	64.068s	1.4%
highDensityRepairSolver	85.174s	1.9%
highDensityStitchSolver	2.882s	0.1%
traceSimplificationSolver	235.010s	5.2%
mutatedPreloadedTraceSimplificationSolver	0.507s	0.0%
traceWidthSolver	54.728s	1.2%
globalDrcForceImproveSolver	53.481s	1.2%
pipeline9JointDrcRepairSolver	213.790s	4.7%
effortCleanupSolver	0.001s	0.0%
lengthMatchingPostProcessingSolver	0.537s	0.0%
powerTraceExpansionSolver	0.001s	0.0%
Total	4555.482s	100.0%

Workflow: View run
Artifact: Download raw reports

ShiboSoftwareDev · 2026-10-06T23:47:22Z

/benchmark --same-machine --dataset 18

tscircuitbot · 2026-10-06T23:48:04Z

Same Machine Benchmark Results

The paired benchmark ended with cancelled before both reports were produced.

Workflow: View run
Artifact: Download raw reports

ShiboSoftwareDev · 2026-10-06T23:48:43Z

/benchmark --same-machine --dataset 18

tscircuitbot · 2026-10-06T23:49:22Z

Same Machine Benchmark Results

Both revisions ran sequentially in one Blacksmith job on blacksmith-8vcpu-ubuntu-2404-arm-kiuui3sswsti7da7.

Dataset: srj18 · Scenarios: 16
Base: 374f3b5 · PR: 01ce1d2

Solver	Metric	Base	PR	Delta
Pipeline9	Completion	87.5% (🕒12.5%)	87.5% (🕒12.5%)	0.0 pp
Pipeline9	Relaxed DRC pass	87.5% (🕒12.5%)	87.5% (🕒12.5%)	0.0 pp
Pipeline9	DRC issues	0	0	0
Pipeline9	Timeouts	2	2	0
Pipeline9	P50 time	85.7s	87.6s	+2.1%
Pipeline9	P60 time	128.2s	133.8s	+4.3%
Pipeline9	P70 time	153.2s	154.9s	+1.1%
Pipeline9	P80 time	276.3s	259.6s	-6.0%
Pipeline9	P90 time	339.1s	339.7s	+0.2%
Pipeline9	P95 time	360.0s	360.0s	0.0%
Pipeline9	Memory P50	1958.5 MiB	1660.2 MiB	-15.2%
Pipeline9	Memory P80	3004.4 MiB	2741.1 MiB	-8.8%
Pipeline9	Memory P90	3385.5 MiB	3360.3 MiB	-0.7%
Pipeline9	Average vias	217.57	217.57	0.0%
Pipeline9	Avg Angled Traces	529.86	529.86	0.0%

Outcome changes: 0 improved, 0 regressed. DRC issues are totaled across solved samples. Timing percentiles include all samples, with failed and timed-out samples counted at their configured timeout; negative timing deltas are faster. Historical failures without timeout metadata make timing percentiles unavailable.
Style errors are averaged per completed sample with recorded lint counts for that type; historical/unlinted results are n/a. Angled traces counts violating segments.

Base pipeline stage timings

AutoroutingPipelineSolver9_PreloadedTraceGraph

Recorded timings: 16/16 samples (2 partial, including failed or timed-out samples).
Times are summed across samples; percentages use the total recorded stage time for this solver.

Stage	Total time	% of stage time
preprocessSimpleRouteJsonSolver	4.380s	0.2%
componentDetectionSolver	0.227s	0.0%
escapeViaLocationSolver	0.631s	0.0%
netToPointPairsSolver	0.823s	0.0%
topologyPlanningSolver	20.686s	1.0%
topologyMergingSolver	21.536s	1.0%
nodeDimensionSubdivisionSolver	0.428s	0.0%
edgeSolver	4.195s	0.2%
availableSegmentPointSolver	1.146s	0.1%
necessaryCrampedPortPointSolver	4.727s	0.2%
preloadedTraceGraphSolver	0.087s	0.0%
portPointPathingSolver	401.403s	19.1%
uniformPortDistributionSolver	8.348s	0.4%
highDensityRouteSolver	791.036s	37.6%
highDensityForceImproveSolver	67.942s	3.2%
highDensityRepairSolver	128.085s	6.1%
highDensityStitchSolver	2.402s	0.1%
traceSimplificationSolver	176.727s	8.4%
mutatedPreloadedTraceSimplificationSolver	0.406s	0.0%
traceWidthSolver	36.479s	1.7%
globalDrcForceImproveSolver	63.173s	3.0%
pipeline9JointDrcRepairSolver	367.056s	17.5%
effortCleanupSolver	0.000s	0.0%
lengthMatchingPostProcessingSolver	0.598s	0.0%
powerTraceExpansionSolver	0.000s	0.0%
Total	2102.524s	100.0%

PR pipeline stage timings

AutoroutingPipelineSolver9_PreloadedTraceGraph

Recorded timings: 16/16 samples (2 partial, including failed or timed-out samples).
Times are summed across samples; percentages use the total recorded stage time for this solver.

Stage	Total time	% of stage time
preprocessSimpleRouteJsonSolver	4.442s	0.2%
componentDetectionSolver	0.213s	0.0%
escapeViaLocationSolver	0.580s	0.0%
netToPointPairsSolver	0.902s	0.0%
topologyPlanningSolver	20.569s	1.0%
topologyMergingSolver	20.964s	1.0%
nodeDimensionSubdivisionSolver	0.383s	0.0%
edgeSolver	3.949s	0.2%
availableSegmentPointSolver	1.205s	0.1%
necessaryCrampedPortPointSolver	4.026s	0.2%
preloadedTraceGraphSolver	0.109s	0.0%
portPointPathingSolver	408.921s	19.4%
uniformPortDistributionSolver	8.533s	0.4%
highDensityRouteSolver	804.557s	38.2%
highDensityForceImproveSolver	64.467s	3.1%
highDensityRepairSolver	118.594s	5.6%
highDensityStitchSolver	2.345s	0.1%
traceSimplificationSolver	178.767s	8.5%
mutatedPreloadedTraceSimplificationSolver	0.415s	0.0%
traceWidthSolver	36.416s	1.7%
globalDrcForceImproveSolver	62.732s	3.0%
pipeline9JointDrcRepairSolver	361.888s	17.2%
effortCleanupSolver	0.000s	0.0%
lengthMatchingPostProcessingSolver	0.582s	0.0%
powerTraceExpansionSolver	0.000s	0.0%
Total	2105.561s	100.0%

Workflow: View run
Artifact: Download raw reports

perf: reduce high-density route node memory

b1d54c9

vercel Bot deployed to Preview October 6, 2026 20:58 View deployment

perf: release failed high-density search state

dc77393

ShiboSoftwareDev changed the title ~~perf: reduce high-density route node memory~~ perf: release failed high-density route searches Oct 6, 2026

vercel Bot deployed to Preview October 6, 2026 23:48 View deployment

test: verify failed search state is released

01ce1d2

vercel Bot deployed to Preview October 6, 2026 23:49 View deployment

Navigation Menu

Repository navigation

perf: release failed high-density route searches - #2894

perf: release failed high-density route searches#2894
ShiboSoftwareDev wants to merge 3 commits into
mainfrom
perf/reduce-high-density-node-memory

ShiboSoftwareDev commented Oct 6, 2026 •

edited

Loading

Uh oh!

vercel Bot commented Oct 6, 2026 •

edited

Loading

Uh oh!

tscircuitbot commented Oct 6, 2026

Uh oh!

ShiboSoftwareDev commented Oct 6, 2026

Uh oh!

ShiboSoftwareDev commented Oct 6, 2026

Uh oh!

ShiboSoftwareDev commented Oct 6, 2026

Uh oh!

tscircuitbot commented Oct 6, 2026 •

edited

Loading

Uh oh!

tscircuitbot commented Oct 6, 2026 •

edited

Loading

Uh oh!

tscircuitbot commented Oct 6, 2026 •

edited

Loading

Uh oh!

ShiboSoftwareDev commented Oct 6, 2026

Uh oh!

tscircuitbot commented Oct 6, 2026 •

edited

Loading

Uh oh!

ShiboSoftwareDev commented Oct 6, 2026

Uh oh!

tscircuitbot commented Oct 6, 2026 •

edited

Loading

Uh oh!

This branch was successfully deployed

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

2 participants

Conversation

ShiboSoftwareDev commented Oct 6, 2026 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

What changed

Results

Validation

Uh oh!

vercel Bot commented Oct 6, 2026 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

tscircuitbot commented Oct 6, 2026

Benchmark This PR

Uh oh!

ShiboSoftwareDev commented Oct 6, 2026

Uh oh!

ShiboSoftwareDev commented Oct 6, 2026

Uh oh!

ShiboSoftwareDev commented Oct 6, 2026

Uh oh!

tscircuitbot commented Oct 6, 2026 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Same Machine Benchmark Results

Uh oh!

tscircuitbot commented Oct 6, 2026 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Same Machine Benchmark Results

Uh oh!

tscircuitbot commented Oct 6, 2026 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Same Machine Benchmark Results

Uh oh!

ShiboSoftwareDev commented Oct 6, 2026

Uh oh!

tscircuitbot commented Oct 6, 2026 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Same Machine Benchmark Results

Uh oh!

ShiboSoftwareDev commented Oct 6, 2026

Uh oh!

tscircuitbot commented Oct 6, 2026 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Same Machine Benchmark Results

Uh oh!

This branch was successfully deployed

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

2 participants

ShiboSoftwareDev commented Oct 6, 2026 •

edited

Loading

vercel Bot commented Oct 6, 2026 •

edited

Loading

tscircuitbot commented Oct 6, 2026 •

edited

Loading

tscircuitbot commented Oct 6, 2026 •

edited

Loading

tscircuitbot commented Oct 6, 2026 •

edited

Loading

tscircuitbot commented Oct 6, 2026 •

edited

Loading

tscircuitbot commented Oct 6, 2026 •

edited

Loading