cmdperf

Interactive command-line benchmarking tool

TLDR

Benchmark a single command with default 10 runs

$ cmdperf "sleep 0.1"

Compare two commands side-by-side

$ cmdperf "grep 'ERROR' log.txt" "rg 'ERROR' log.txt"

Run 100 iterations of a command

$ cmdperf -n 100 "ls -la"

Simulate 10 concurrent requests

$ cmdperf -c 10 "curl -s https://example.com > /dev/null"

Run for a fixed duration of 30 seconds

$ cmdperf -d 30s "redis-cli PING"

Export results to a markdown file

$ cmdperf --markdown [results.md] "sleep 0.1" "sleep 0.2"

SYNOPSIS

cmdperf [options] command [command...]

cmdperf runs one or more shell commands repeatedly and gathers timing statistics, displaying them in a live terminal UI with a progress bar, ETA, and key metrics including mean execution time, standard deviation, and min/max ranges. It supports single command benchmarking, multi-command comparison, concurrent execution to simulate load, duration-based runs, and rate limiting for HTTP endpoint testing.

PARAMETERS

-n, --runs=N

Number of iterations to run (default: 10).

-c, --concurrency=N

Number of parallel executions (default: 1).

-d, --duration=DURATION

Run for a fixed duration instead of a fixed number of runs.

-r, --rate=RATE

Target request rate in requests per second.

-t, --timeout=DURATION

Per-command timeout (default: 1m).

-N, --no-shell

Execute command directly without wrapping in a shell.

--csv=FILE

Export results to CSV file.

--markdown=FILE

Export results to Markdown file.

--color-scheme=SCHEME

Set color scheme (default, catppuccin, tokyonight, nord, monokai, solarized, gruvbox, monochrome).

CAVEATS

Designed for quick interactive benchmarking, not a replacement for rigorous statistical profiling tools. Default shell wrapping adds overhead; use -N/--no-shell for tighter measurements when shell features are not needed.

HISTORY

cmdperf was created by Miklos Niedermayer and is written in Go under the MIT license. The latest release is v0.1.4 (January 2026). It fills a niche as a lightweight, interactive alternative to tools like hyperfine, with a focus on live TUI experience and concurrent execution simulation.