Compute Pipeline9 node probabilities from one pathing output

[seveibar]

Pipeline9 calculates each high-density node's failure probability by rebuilding the complete pathing output for every input node. Large boards repeatedly pay for identical route conversion and lookup work before high-density routing starts.

Build the solved-node lookup once and calculate every probability from that output. This distills one performance improvement from #2447 into three files. Probability formulas, pipeline stages, dependency versions, clearances, and search budgets are unchanged.

The same-machine SRJ18 comparison runs /benchmark --same-machine --dataset 18 on main 207c4191 and this PR's c4d639eb sequentially on one runner:

Measurement	Main	PR
DRC passing	9/16 (56.3%)	10/16 (62.5%)
Completion	11/16 (68.8%)	12/16 (75.0%)
Timeouts	3	2
Total sample time	3034.1s	2732.0s (-10.0%)
P50 time	143.3s	128.8s (-10.1%)
P80 time	348.8s	300.6s (-13.8%)
P95 time	360.0s	360.0s

Sample 12 changes from a 360-second timeout to DRC passing in 346.5s. There are no regressed outcomes. All 11 samples that complete on both revisions retain identical DRC messages and via counts. These rates describe this paired run; the quality gain comes from completing an existing route within the unchanged deadline.

Validation: all CI checks pass, including the full test suite, TypeScript, and production build. The generic regression verifies identical node probabilities with one output conversion. A separate full native sample 6 comparison preserves every route, reference DRC result, and non-timing solver statistic while reducing runtime from 300.9s to 265.0s (-11.9%).

[vercel]

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

Project	Deployment	Actions	Updated
capacity-node-autorouter	Ready	Preview	Sep 9, 2026 3:51am UTC

[tscircuitbot]

Benchmark This PR

Run benchmarks by commenting on this PR:

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

Comment /benchmark to run the default dataset, or append any arguments accepted by ./benchmark.sh.
Comment /benchmark-long for an 8-vCPU run that defaults to 8 workers and has an eight-hour timeout.
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 or /benchmark-all to compare current main and the PR 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, and /benchmark-long --dataset 18.

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.

[seveibar]

/benchmark --same-machine --dataset 18

[tscircuitbot]

Same Machine Benchmark Results

Both revisions ran sequentially in one Blacksmith job on blacksmith-01m224maz4rm453y3ffjh1z73g-8vcpu.

Dataset: srj18 · Scenarios: 16
Main: 207c419 · PR: c4d639e

Solver	Metric	Main	PR	Delta
Pipeline9	Completion	68.8% (🕒18.8%)	75.0% (🕒12.5%)	+6.2 pp
Pipeline9	Relaxed DRC pass	56.3% (🕒18.8%)	62.5% (🕒12.5%)	+6.2 pp
Pipeline9	DRC issues	44	44	0
Pipeline9	Timeouts	3	2	-1
Pipeline9	P50 time	143.3s	128.8s	-10.1%
Pipeline9	P60 time	184.2s	169.8s	-7.8%
Pipeline9	P70 time	297.2s	265.2s	-10.8%
Pipeline9	P80 time	348.8s	300.6s	-13.8%
Pipeline9	P90 time	360.0s	356.0s	-1.1%
Pipeline9	P95 time	360.0s	360.0s	0.0%
Pipeline9	Average vias	176.91	187.17	+5.8%

Outcome changes: 1 improved, 0 regressed. DRC issues are totaled across solved samples. Timing percentiles include solved and timed-out samples; negative timing deltas are faster.

Changed outcomes (1)

Solver	Sample	Main	PR	Main time	PR time	Delta
Pipeline9	12	Timeout	DRC passed	360.0s	346.5s	Improved

Workflow: View run
Artifact: Download raw reports

[seveibar]

/benchmark --same-machine --dataset 18

[tscircuitbot]

Same Machine Benchmark Results

Both revisions ran sequentially in one Blacksmith job on blacksmith-01m225zwmty50ey68rnr97xwax-8vcpu.

Dataset: srj18 · Scenarios: 16
Main: 207c419 · PR: c4d639e

Solver	Metric	Main	PR	Delta
Pipeline9	Completion	81.3% (🕒6.3%)	81.3% (🕒6.3%)	0.0 pp
Pipeline9	Relaxed DRC pass	62.5% (🕒6.3%)	62.5% (🕒6.3%)	0.0 pp
Pipeline9	DRC issues	51	51	0
Pipeline9	Timeouts	1	1	0
Pipeline9	P50 time	111.3s	100.8s	-9.4%
Pipeline9	P60 time	137.8s	132.7s	-3.7%
Pipeline9	P70 time	247.4s	230.4s	-6.9%
Pipeline9	P80 time	315.2s	283.2s	-10.2%
Pipeline9	P90 time	333.4s	322.0s	-3.4%
Pipeline9	P95 time	343.3s	341.9s	-0.4%
Pipeline9	Average vias	210.92	210.92	0.0%

Outcome changes: 0 improved, 0 regressed. DRC issues are totaled across solved samples. Timing percentiles include solved and timed-out samples; negative timing deltas are faster.

Workflow: View run
Artifact: Download raw reports

[tscircuitbot]

---

Thank you for your contribution! 🎉

PR Rating: ⭐⭐⭐
Impact: Major

Track your contributions and see the leaderboard at: tscircuit Contribution Tracker

---