nordisk/facett

facett-bench — crate card

Reverse-engineered (from source — verify, don't trust the name)

• One-line pitch: The detached nornir Bencher harness for facett — render-throughput benchmarks, historized in the warehouse and gated on regression by nornir release / nornir bench run.
• Does: An [[example]] (nornir-bench) implementing nornir::bench::api::Bencher for facett workloads: builds a graph of N nodes+edges and measures nodes/s (GraphRender), plus map / map3d / table renders via facett::harness::headless_render. result(...) packs a BenchResult (<unit>_per_sec headline + vertices + elapsed_ms + n). Ported from the #[ignore]'d perf tests (those are now fast downscaled correctness checks; the full-scale throughput measurement lives here so it lands in bench_history + the README). Headline scale N = 50_000. Optional wgpu feature lights facett-core's L0 GPU SDF upload part-bench.
• Key deps: nornir (the Bencher contract), facett (umbrella), facett-core (direct edge for the GPU part-bench), egui, anyhow, serde_json.
• Shape: bin-ish (an example, not a lib); its OWN [workspace] (package facett-bench, version 0.0.0, publish = false, edition 2024) so nornir's heavy dev-dep never touches the published facett crates' lockfile. Lives in bench/.
• Observed role in the workspace: Run by nornir bench run facett (nornir detects <repo>/bench/examples/nornir-bench.rs and runs cargo inside bench/). Feeds the regression gate + bench history.

Fill in (one line each — only the human knows these)

• Primary consumer(s):
• Why it exists (vs the obvious alternative):
• Hard constraint (airgap / no-C-deps / no-FFI / pure-Rust / etc.):
• Release cadence & independence (rides a parent? own repo someday?):
• Status (production / prototype / experimental / stubbed):
• Must-NOT-change invariant:
• Biggest known gap or TODO:
• Anything a newcomer would get wrong about it: