A command-line benchmarking tool
Go to file
Christian Duerr 4be435436e Allow forwarding benchmark's stdout to hyperfine
This change makes it possible to forward the stdout of the command the
user tries to benchmark to hyperfine's stdout using the `--print-stdout`
flag.

This should be irrelevant for most use-cases, however it makes it
possible to get some crude benchmarks of terminal emulators similar to
what's currently often done with `time cat file`.
2018-05-22 22:54:08 +02:00
ci Activate Travis CI 2018-01-18 19:08:19 +01:00
src Allow forwarding benchmark's stdout to hyperfine 2018-05-22 22:54:08 +02:00
.gitignore Initial commit 2018-01-13 16:54:38 +01:00
.travis.yml Add correct secure key 2018-01-21 18:11:27 +01:00
appveyor.yml Fix quotes 2018-02-02 22:42:53 +01:00
Cargo.lock Bump version to 1.0 2018-03-24 20:44:26 +01:00
Cargo.toml Bump version to 1.0 2018-03-24 20:44:26 +01:00
LICENSE-APACHE Add Apache license 2018-01-13 20:39:55 +01:00
LICENSE-MIT Add MIT license 2018-01-13 20:39:24 +01:00
README.md Update install instructions 2018-03-24 22:06:03 +01:00

hyperfine

Build Status Build status Version info

A command-line benchmarking tool (inspired by bench).

Demo: Benchmarking fd and find:

hyperfine

Features

  • Statistical analysis across multiple runs.
  • Support for arbitrary shell commands.
  • Constant feedback about the benchmark progress and current estimates.
  • Warmup runs can be executed before the actual benchmark.
  • Cache-clearing commands can be set up before each timing run.
  • Statistical outlier detection.
  • Export results to various formats: CSV, JSON, Markdown.
  • Parameterized benchmarks.
  • Cross-platform

Usage

Basic benchmark

To run a benchmark, you can simply call hyperfine <command>.... The argument(s) can be any shell command. For example:

hyperfine 'sleep 0.3'

Hyperfine will automatically determine the number of runs to perform for each command. By default, it will perform at least 10 benchmarking runs. To change this, you can use the -m/--min-runs option:

hyperfine --min-runs 5 'sleep 0.2' 'sleep 3.2'

I/O-heavy programs

If the program execution time is limited by disk I/O, the benchmarking results can be heavily influenced by disk caches and whether they are cold or warm.

If you want to run the benchmark on a warm cache, you can use the -w/--warmup option to perform a certain number of program executions before the actual benchmark:

hyperfine --warmup 3 'grep -R TODO *'

Conversely, if you want to run the benchmark for a cold cache, you can use the -p/--prepare option to run a special command before each timing run. For example, to clear harddisk caches on Linux, you can run

sync; echo 3 | sudo tee /proc/sys/vm/drop_caches

To use this specific command with Hyperfine, call sudo -v to temporarily gain sudo permissions and then call:

hyperfine --prepare 'sync; echo 3 | sudo tee /proc/sys/vm/drop_caches' 'grep -R TODO *'

Parameterized benchmarks

If you want to run a benchmark where only a single parameter is varied (say, the number of threads), you can use the -P/--parameter-scan option and call:

hyperfine --prepare 'make clean' --parameter-scan num_threads 1 12 'make -j {num_threads}'

Export results

Hyperfine has multiple options for exporting benchmark results: CSV, JSON, Markdown (see --help text for details). To export results to Markdown, for example, you can use the --export-markdown option that will create tables like this:

Command Mean [ms] Min…Max [ms]
find . -iregex '.*[0-9]\.jpg$' 506.0 ± 8.1 495.4…518.6
find . -iname '*[0-9].jpg' 304.9 ± 3.1 299.8…309.3
fd -HI '.*[0-9]\.jpg$' 66.2 ± 5.8 62.5…86.3

Installation

On Ubuntu

Download the appropriate .deb package from the Release page and install it via dpkg:

wget https://github.com/sharkdp/hyperfine/releases/download/v1.0.0/hyperfine_1.0.0_amd64.deb
sudo dpkg -i hyperfine_1.0.0_amd64.deb

On Arch Linux

On Arch Linux, hyperfine can be installed from the AUR:

yaourt -S hyperfine

On Void Linux

Hyperfine can be installed via xbps

xbps-install -S hyperfine

With cargo (Linux, macOS, Windows)

Hyperfine can be installed via cargo:

cargo install hyperfine

From binaries (Linux, macOS)

Download the corresponding archive from the Release page.

Origin of the name

The name hyperfine was chosen in reference to the hyperfine levels of caesium 133 which play a crucial role in the definition of our base unit of time — the second.