lazymark

A simple, cross-platform CLI tool for timing command-line benchmarks.

Measure how long any command takes to execute — once, multiple times with statistics, or in batch from a file. Works on Linux, macOS, and Windows.

Installation

cargo install --path .

Or build from source:

cargo build --release

Usage

Direct mode

Benchmark a single command:

lazymark echo hello
# ⏱  echo hello ··· 12.345ms

lazymark -n 10 echo hello
# ⏱  echo hello (10 runs)
#     Min: 11.200ms  Max: 15.800ms  Mean: 12.500ms  StdDev: 1.200ms

Note: Place lazymark flags (-n, --show-output, etc.) before the command to benchmark.

File mode

Read commands from a file:

lazymark -f benchmarks.json
lazymark -f commands.csv -n 5 --log results.json --log-format json

Interactive mode

Launch by running lazymark with no arguments:

$ lazymark
lazymark interactive mode. Type a command to benchmark it.
Use :help for available commands, :quit to exit.

lazymark> cargo build
⏱  cargo build ··· 2.345s

lazymark> :n 5 cargo test
Iterations set to 5
⏱  cargo test (5 runs)
    Min: 1.200s  Max: 1.800s  Mean: 1.450s  StdDev: 0.230s

lazymark> :quit

Interactive commands

Command	Description
:quit, :q	Exit
:iterations N, :n N	Set iteration count
:n N COMMAND	Set iterations and run command
:show-output	Show command stdout/stderr
:hide-output	Hide command output (default)
:log PATH	Log results to file
:log-format FORMAT	Set log format (csv, json, text)
:log-stop	Stop logging
:history	Show session benchmark history
:help, :h	Show help

CLI Options

Usage: lazymark [OPTIONS] [COMMAND]...

Arguments:
  [COMMAND]...  Command and arguments to benchmark

Options:
  -f, --file <PATH>          Read commands from file (csv, newline, json, yaml)
  -n, --iterations <N>       Number of iterations per command [default: 1]
      --show-output          Show stdout/stderr of benchmarked commands
  -l, --log <PATH>           Log results to a file
      --log-format <FORMAT>  Log format: csv, json, text [default: csv]
  -h, --help                 Print help
  -V, --version              Print version

File Formats

CSV (.csv)

Comma-separated commands, one or more per line:

echo hello,echo world
cargo build

Newline-delimited (.txt)

One command per line. Empty lines and # comments are ignored:

echo hello
# This is a comment
cargo build
cargo test

JSON (.json)

{
  "benchmarks": [
    {
      "name": "build-and-test",
      "commands": ["cargo build", "cargo test"]
    },
    {
      "commands": ["echo hello"]
    }
  ]
}

When multiple commands are in the same object, their execution times are summed and reported as a group.

YAML (.yml / .yaml)

benchmarks:
  - name: build-and-test
    commands:
      - cargo build
      - cargo test
  - commands:
      - echo hello

Same grouping behavior as JSON — commands in the same entry have times summed.

Logging

Log benchmark results to a file in any mode:

# Direct mode
lazymark -n 5 --log results.csv echo hello

# File mode with JSON log
lazymark -f benchmarks.json --log results.json --log-format json

In interactive mode, use :log and :log-format commands.

Log formats

CSV (default) — appends rows, writes header on new file:

name,command,iteration,duration_secs
,echo hello,1,0.012345

JSON — valid JSON array, rewritten on each append:

[
  {
    "name": null,
    "command": "echo hello",
    "iteration": 1,
    "duration_secs": 0.012345
  }
]

Text — one result per line, appended:

echo hello: 0.012345s

Multi-iteration Statistics

With -n N where N > 1, lazymark reports:

• Min / Max — fastest and slowest runs
• Mean — arithmetic average
• StdDev — sample standard deviation (n−1)

Duration display adapts to magnitude:

Range	Format
< 1 second	milliseconds (123.456ms)
< 1 minute	seconds (12.345s)
≥ 1 minute	minutes + seconds (1m 30.500s)

License

MIT