fio

Overview

fio is a command-line tool for defining and running storage I/O workloads, especially for performance testing and reproducing bugs. It suits engineers who need to shape tests around particular storage behavior rather than rely on a fixed benchmark pattern. Its breadth of controls and free GPL license are compelling, though it expects users to work from the command line and understand the workload they want to generate.

Key features

Workload control

Users can define jobs in files or pass parameters directly, then set I/O patterns, block and I/O sizes, engines, queue depth, targets, and whether fio uses threads or processes. This makes fio useful for targeted random or sequential tests, IOPS and latency measurements, and repeatable configurations. The flexibility is a strength for technical users, but it also means fio is not a ready-made, simplified benchmark interface.

Results and remote jobs

Output is available in normal, terse, JSON, and JSON+ formats, giving users options for reading results or feeding them into other workflows. In client/server mode, fio can accept jobs from clients on other hosts, useful when workloads need to run remotely. Its HTTP engine also supports access to HTTP and S3 storage.

Safety and permissions

The --readonly option enables checks intended to prevent writes and trims, and fio aborts destructive jobs on devices that appear mounted unless allow_mounted_write is set. These controls help reduce the risk of an unintended destructive run, but they do not remove the need to choose targets and options carefully. Root access is not generally required unless the selected files or devices demand it; some options may also have permission restrictions.

Pricing

fio has one free plan: free. It is free software under version 2 of the GNU GPL, so there is no paid tier to unlock its workload controls or reporting formats. The project mailing list handles discussion, questions, bug reports, and development; bug reports can also go to GitHub.

Platforms

fio supports Linux, Solaris, AIX, HP-UX, OSX, NetBSD, OpenBSD, Windows, FreeBSD, and DragonFly. For Windows, the documented minimums are Windows 7 and Windows Server 2008 R2, and building fio requires Cygwin. Tagged-release Windows installers are available through GitHub Releases, while Linux distribution packages are also available.

Who it's for

fio is best for storage and systems practitioners who need configurable, repeatable I/O workloads, measurements of IOPS or latency, or remote job execution. Its job files and multiple output formats also suit workflows that need to preserve test definitions or process results. Readers looking for a graphical, one-click disk benchmark may be better served by a tool designed around that simpler use case.

Pros and cons

Pros: Fine-grained control over I/O patterns, sizes, engines, queue depth, targets, and execution model helps match a test to a specific question.

Pros: Normal, terse, JSON, and JSON+ outputs and remote client/server execution support varied reporting and deployment needs.

Pros: Free GPLv2 software spans a broad range of operating systems, with Windows installers and Linux packages available.

Cons: A command-line interface and extensive workload options suit technical users better than people seeking a guided benchmark.

Cons: Windows users must account for the minimum supported versions, and building from source requires Cygwin.

Alternatives

For a broader browse, see the Disk Benchmark Software directory. Choose Parkdale if you want a free disk utility with Android support as well as macOS and Windows. simple-disk-benchmark is another free command-line option, for Linux, macOS, and Windows. Linux users seeking an open-source disk benchmark can consider KDiskMark; Windows users can consider the freeware AS SSD Benchmark or D-Bench, which supports Windows 8/8.1/10/11 or Vista/7. Filebench is a free Linux option. DaVinci Resolve and Buddi are also listed as alternatives.

Verdict

Choose fio if you need free, cross-platform control over storage workloads, detailed measurements, or remote execution. Its main advantage is how precisely a test can be configured; look elsewhere if you want a graphical benchmark that hides workload setup behind a simple interface.