| id | et-71ae | |
|---|---|---|
| status | closed | |
| deps | ||
| links |
|
|
| created | 2026-01-31 14:47:37 UTC | |
| type | task | |
| priority | 2 | |
| assignee | no-materials |
We have an initial Criterion-based wind-tunnel crate (execution_tape_wind_tunnel) with a few core scenarios.
Add additional benchmarks that better match real workloads and reduce measurement bias:
- “Many ops per run” variants (e.g. 1_000 host calls in one run; 1_000 calls in one run) to avoid over-measuring per-run setup.
- Allocation/heap pressure scenarios (bytes/str concat, aggregate construction/access) to surface allocation regressions.
- Optional parameterization for sizes (N, payload sizes) to get scaling curves.
Acceptance:
- New benches compile and run under
cargo bench -p execution_tape_wind_tunnel --bench vm. - Each bench has a short comment describing what it measures.
2026-02-01T05:45:28Z
Bench updates likely land under et-eaf4 (PR7: Cleanup + wind-tunnel perf confirmation).
2026-02-05T10:55:07Z
Implemented new wind-tunnel benches (bytes/str concat chains + array_alloc_loop) and expanded loop iterations to include 1_000. Ran cargo bench -p execution_tape_wind_tunnel --bench vm successfully.