Command-line reference¶
benchbro run [target] [options]
benchbro list [target] [--verbose]
benchbro compare BASELINE.json CURRENT.json
benchbro baseline update [target] --baseline NAME
benchbro baseline list
benchbro history list
benchbro history show SELECTOR
benchbro history compare OLDER NEWER
benchbro init [path]
benchbro completion bash|zsh|fish
The original benchbro TARGET [options] form remains supported.
Common run options¶
| Option | Purpose |
|---|---|
--config PATH |
Read a specific standalone configuration file |
--case NAME |
Select a case; repeat to select more than one |
--tag TAG |
Select cases by tag; repeat to select more than one |
--repeats N |
Set the maximum repeat count |
--warmup N |
Set warmup invocations |
--min-iterations N |
Set invocations per repeat |
--adaptive / --fixed |
Enable or disable adaptive stopping |
--min-repeats N |
Require an adaptive sample floor |
--min-time SECONDS |
Require a minimum measurement duration |
--max-time SECONDS |
Cap measurement duration |
--target-margin PCT |
Set the target 95% CI relative margin |
--noise-threshold PCT |
Set the noisy-result CV threshold |
--isolation process |
Run each benchmark in a worker process |
--cpu-affinity 2,3 |
Pin measurement to Linux CPU IDs |
--timeout SECONDS |
Set a benchmark timeout |
--profile NAME |
Attach a registered profiler |
--histogram |
Draw terminal histograms for time benchmarks |
Outputs and comparison¶
| Option | Purpose |
|---|---|
--output-json PATH |
Write a schema-versioned JSON result |
--output-csv PATH |
Write long-form metric rows |
--output-md PATH |
Write a Markdown summary table |
--baseline NAME |
Select a named baseline |
--new-baseline |
Replace the selected baseline |
--save-history |
Retain the run in local history |
--no-compare |
Skip comparison |
--ci |
Use the strict committed CI baseline |
--allow-environment-mismatch |
Permit an intentional cross-environment comparison |
Run benchbro COMMAND --help for the complete options accepted by one command.
Exit codes¶
| Code | Meaning |
|---|---|
0 |
Successful or non-regressing run |
1 |
Invalid input, discovery, configuration, or comparison setup |
2 |
Statistically supported regression |
3 |
Benchmark execution failure |
LIKELY and INCONCLUSIVE threshold crossings do not return exit code 2.
Shell completion¶
Generate a script for Bash, Zsh, or Fish:
Load it using the completion mechanism for your shell.