Canonical Benchmark Sweep
Performance claims are tied to canonical scenes and stored artifacts, not a single vanity number.
What you are seeing
Scene
A buyer-facing benchmark lane that pairs moving scene workloads with trend graphs and headline throughput numbers so the proof reads as measurement, not UI decoration.
Why it matters
Put benchmark behavior in motion by showing the named workloads used for performance review and regression tracking.
Pass condition
Throughput should stay within the declared benchmark envelope for the reference configuration.
Proof views
Use Public proof for the clean buyer-facing read, Event focus for adaptive slowdowns around the key moment, and Debug proof for contact cues, overlays, frame numbers, and proof-lane instrumentation.
Metrics strip
Technical drawer
Command / source path
build\bin\benchmarks\upe_canonical_scenes.exe
Systems exercised
- world stepping
- scene scaling
- regression tracking
- artifact publishing
Known limitations
- Public benchmark figures are directional and should be re-run on the target buyer workload.
- Different machines and build flags will change the reported numbers.
Recommended context
Use this when a buyer wants the benchmark story to feel concrete before deeper diligence.
Determinism status
- same build: artifact-backed on the reference lane
- cross-run: tracked through trend history
- cross-platform: evaluate per platform
- snapshot / restore: not the primary claim of this scene
Related proof
Mixed Shape Resting Contact
Mixed primitive contacts settle cleanly without obvious interpenetration drift.
View test