Benchmarks
@brushy/di-core ranked #1 among DI libraries (tsyringe, InversifyJS, awilix) in 8/8 scenarios. Baseline (new direct) is measured separately as overhead reference, not ranked against DI runtimes.
Environment
Results are indicative. Run benchmarks on an idle machine for stable numbers. They are not a substitute for profiling your own app.
Published numbers come from the full suite (
time=3000ms, runs=5). CI on main refreshes BENCHMARK.md weekly via .github/workflows/benchmark.yml.Speed advantage over best competitor
Throughput (p50) by scenario
request_scope: tsyringe and inversify have no native request scope (N/A). Bars are scaled per row (@brushy/di-core = 10 blocks).
Full results and methodology
- BENCHMARK.md: executive summary, rankings, and detailed metrics
- METHODOLOGY.md: scenario definitions, adapters, and CI refresh policy
- latest.json / latest.csv: machine-readable exports
Run locally
Build@brushy/di-core, then run the benchmark suite from the project root:
partial-* files under packages/di-bench/results/. Only full runs update latest.* and BENCHMARK.md.