Benchmarking
Cuprate has 2 types of benchmarks:
- Criterion benchmarks
cuprate-benchmarkbenchmarks
Criterion is used for micro benchmarks; they time single functions, groups of functions, and generally are small in scope.
cuprate-benchmark and cuprate-benchmark-lib are custom in-house crates Cuprate uses for macro benchmarks; these test sub-systems, sections of a sub-system, or otherwise larger or more complicated code that isn't well-suited for micro benchmarks.
File layout and purpose
All benchmarking related files are in the benches/ folder.
This directory is organized like such:
| Directory | Purpose |
|---|---|
benches/criterion/ | Criterion (micro) benchmarks |
benches/criterion/cuprate-* | Criterion benchmarks for the crate with the same name |
benches/benchmark/ | Cuprate's custom benchmarking files |
benches/benchmark/bin | The cuprate-benchmark crate; the actual binary run that links all benchmarks |
benches/benchmark/lib | The cuprate-benchmark-lib crate; the benchmarking framework all benchmarks plug into |
benches/benchmark/cuprate-* | cuprate-benchmark benchmarks for the crate with the same name |