Punch Promoter Engine Performance

Static benchmark report comparing the TypeScript simulation path with the Rust headless engine used by terminal-native long-run simulations.

Benchmark generated 2026-05-14 02:49:57
219.20x
Best speedup
Fastest Rust-over-TS result across workloads
95.40x
Average speedup
Mean of reported workload speedups
3
Fight workloads
Pure scheduled-fight batch tests
3
Weekly workloads
Headless lifecycle windows
1
Runs per sample
Measured runs after warmup
Benchmark

Fight Batch Compute

Pure scheduled-fight workloads using identical deterministic fixtures for TypeScript and Rust.

25 scheduled fights
25.23x
JS 18.97 s / Rust 751.7 ms per fight workload
100 scheduled fights
219.20x
JS 81.03 s / Rust 369.7 ms per fight workload
400 scheduled fights
217.25x
JS 326.40 s / Rust 1.50 s per fight workload
25 scheduled fights JS
18.97 s
25 scheduled fights Rust
751.7 ms
100 scheduled fights JS
81.03 s
100 scheduled fights Rust
369.7 ms
400 scheduled fights JS
326.40 s
400 scheduled fights Rust
1.50 s
WorkloadTypeScript MeanRust MeanSpeedupParity / Output
25 scheduled fights18.97 s751.7 ms25.23xsame fight count
100 scheduled fights81.03 s369.7 ms219.20xsame fight count
400 scheduled fights326.40 s1.50 s217.25xsame fight count
Benchmark

Weekly Lifecycle

Headless game weeks that include simulation orchestration, database writes, scheduling, and persistence.

12 simulated weeks
50.70x
JS 152.54 s / Rust 3.01 s per weekly workload
48 simulated weeks
32.44x
JS 347.46 s / Rust 10.71 s per weekly workload
192 simulated weeks
27.56x
JS 2156.38 s / Rust 78.23 s per weekly workload
12 simulated weeks JS
152.54 s
12 simulated weeks Rust
3.01 s
48 simulated weeks JS
347.46 s
48 simulated weeks Rust
10.71 s
192 simulated weeks JS
2156.38 s
192 simulated weeks Rust
78.23 s
WorkloadTypeScript MeanRust MeanSpeedupParity / Output
12 simulated weeks152.54 s3.01 s50.70xJS fights 179; Rust fights 179
48 simulated weeks347.46 s10.71 s32.44xJS fights 414; Rust fights 412
192 simulated weeks2156.38 s78.23 s27.56xJS fights 3,153; Rust fights 3,228
Methodology

Environment And Run Config

Benchmark metadata is recorded with the report so results can be interpreted as local, reproducible measurements.

PlatformDarwin 25.3.0
Architecturearm64
CPUApple M4
CPU cores10
Memory24 GB
Nodev22.17.1
Rustrustc 1.92.0 (ded5c06cf 2025-12-08)
Cargocargo 1.92.0 (344c4567c 2025-10-21)
Seed42
Warmups0
Notes

How To Read This

These notes describe the benchmark boundaries and caveats.

  • Fight batch benchmarks run pure week-tick fight fixtures without app UI rendering.
  • Weekly lifecycle benchmarks include headless database setup, world initialization, weekly orchestration, simulation, and persistence for each measured sample.
  • Results are machine-specific and should be read as portfolio-scale evidence, not universal hardware-independent constants.