- Home
- Disk Benchmark Software
- simple-disk-benchmark
simple-disk-benchmark
What it is
simple-disk-benchmark is a free MIT-licensed command-line utility for measuring disk read and write performance on Linux, macOS, and Windows. It can run read tests, write tests, or both, with controls for test-file size, block size, cycle count, and random seeking. Results appear as a bar chart of run timings, and the program can export summary timings and individual runs as JSON. The benchmark creates a file at the selected path, so an existing file there is deleted. Options can retain the test file, skip creating it, and determine whether the file closes between cycles. The project describes cache-bypass approaches using F_NO_CACHE and F_GLOBAL_NOCACHE on macOS and O_DIRECT on Linux. Cargo installation works on all listed platforms, while Homebrew installation is available on macOS. macOS and Linux are tested; Windows support is limited to passing unit tests, with no other testing or validation stated. JSON output can be piped to jq or csvkit for CSV workflows.
For and against
For
- Free MIT-licensed command-line tool
- Runs read, write, and combined tests
- Configurable file, block, and cycle sizes
- Exports timing data as JSON
- Supports random and sequential tests
Against
- Existing target files are deleted
- Windows support has limited validation
- Multithreaded benchmarking remains on the TODO list
- Requires command-line installation
The verdict
simple-disk-benchmark provides focused disk testing with adjustable workloads, charts, and JSON export. Its file-deletion behavior and limited Windows validation deserve attention before running it on important paths.
Who it suits
This tool suits technical users who want configurable command-line disk benchmarks and machine-readable timing output. It is best aligned with macOS and Linux users because Windows validation is limited.
Plans and prices
simple-disk-benchmark
MIT licensed · command-line tool
Read 30 Sep 2026 · github.com
On the record
- Benchmark controls
- Options let users set the test file size, block size, cycle count, and random seeking.github.com · read 30 Sep 2026
- Benchmark types
- It can run read tests, write tests, or both.github.com · read 30 Sep 2026
- Cache handling
- The README says macOS uses F_NO_CACHE and F_GLOBAL_NOCACHE and Linux uses O_DIRECT to bypass the filesystem cache.github.com · read 30 Sep 2026
- Development status
- The README lists multithreaded benchmarking, volume information, and more documentation among its TODOs.github.com · read 30 Sep 2026
- File behavior
- The benchmark creates a file of the specified size, and an existing file at the chosen path will be deleted.github.com · read 30 Sep 2026
- File options
- Options can keep the test file after the run, skip creating it, and control whether it is closed between cycles.github.com · read 30 Sep 2026
- Installation
- It can be installed with Cargo on all platforms or through Homebrew on macOS.github.com · read 30 Sep 2026
- Integrations
- The README suggests piping JSON output to jq or csvkit for CSV use.github.com · read 30 Sep 2026
- License
- The repository identifies the software as MIT licensed.github.com · read 30 Sep 2026
- Platform support
- macOS and Linux are tested; Windows support is limited, with unit tests passing but no other testing or validation stated.github.com · read 30 Sep 2026
- Purpose
- The project describes itself as a simple disk benchmark tool.github.com · read 30 Sep 2026
- Results
- The tool displays a bar chart of run timings and can export timing summaries and individual run timings as JSON.github.com · read 30 Sep 2026
- Support
- The repository shows zero open issues and does not state a support policy.github.com · read 30 Sep 2026
