Browse documentation

Technical documentation / Reference

Repeated setup benchmark

Credential-free benchmark for startup settings work, repeated classic-prompt setup, ten identical provider overrides, and provider tool-body construction.

Run

ISSUE_379_BENCH_REPETITIONS=50 \
  cargo test --release --locked --lib profile_issue_379_repeated_setup_paths -- --ignored --nocapture

The harness emits seven samples per path with minimum, median, and maximum nanoseconds per operation. Use release builds for comparisons. The fixture uses a temporary MC_HOME, local no-auth provider configuration, local profiles, and no network.

Compare results

Use the same machine, build profile, repetitions, and system load. Dynamic and filtered tool definitions still require request-local construction; unchanged Chat Completions and Anthropic tool shapes reuse immutable definitions.

The recorded implementation run preserves earlier measurements and their limits. It is not a current performance guarantee.

Edit this page on GitHub ↗