Install hdf5 with Homebrew, apk, apt, dnf, MacPorts, Nix, pacman, zypper

Project links

• Homepage
• Repository
• Upstream docs
• Package manager page

Commands and aliases

• h5c++
• h5cc
• h5clear
• h5copy
• h5debug
• h5delete
• h5diff
• h5dump
• h5fc
• h5format_convert
• h5import
• h5jam
• h5ls
• h5mkgrp
• h5perf_serial
• h5repack
• h5repart
• h5stat
• h5unjam
• h5watch

Project history

The HDF lineage began in 1987 at the National Center for Supercomputing Applications at the University of Illinois, where the Graphics Foundations Task Force set out to create an architecture-independent library and file format for sharing scientific data across different computing platforms.

By 1996, the original HDF design was no longer enough for growing data volumes and computing systems. HDF5 was implemented in 1998 as a more scalable successor to HDF4, with core funding and requirements from Department of Energy national laboratories and other scientific-data stakeholders.

The HDF Group separated from the University/NCSA setting in the 2005-2006 period to steward HDF technologies, support users, and keep the software aligned with advancing data requirements. The GitHub repository was created in 2020 as the official public source repository for the HDF5 library.

Adoption history

HDF5's adoption is unusually broad for a file-format library. The HDF Group describes it as used across academic research, aerospace, automotive, electronic instruments, finance, government, medicine, oil and gas, silicon manufacturing, astronomy, computational fluid dynamics, earth sciences, genomics, physics, and other data-heavy fields.

The format spread because it combines portability, self-description, hierarchical groups, multidimensional datasets, metadata with data, compression and filtering support, and interfaces for C, C++, Fortran, Java, and higher-level ecosystems. It became infrastructure for tools rather than just an application format.

Package-manager adoption reflects that infrastructure role: the batch input records Alpine, Homebrew, Debian/Ubuntu helpers, Fedora, MacPorts, Nix, Arch, and openSUSE packaging. Downstream software often depends on libhdf5, h5dump, h5ls, h5repack, h5diff, compiler wrappers, and headers rather than the user intentionally launching a single HDF5 app.

How it is used

Users create HDF5 files to store nested groups, datasets, attributes, datatypes, and metadata in one portable file. Command-line tools such as h5dump, h5ls, h5copy, h5diff, h5repack, h5stat, h5import, and h5watch make the package useful even outside application development.

Developers link against HDF5 through C, C++, Fortran, Java, and bindings in scientific languages. The official README points users to install guides, CMake builds, documentation, release notes, examples, support, and HPC guidance, underscoring that HDF5 is both a library and an operational ecosystem.

Parallel HDF5 extends the same model to MPI programs that need shared access to HDF5 files. The serial hdf5 package remains the ordinary default for many desktop, scripting, and build dependencies, while MPI-enabled builds are split out by package managers when parallel I/O semantics are needed.

Why package nerds care

HDF5 is package-nerd famous because it is a dependency with a large ABI and feature surface: serial versus MPI builds, CMake options, C and Fortran compiler wrappers, compression filters, high-level APIs, Java components, command-line tools, and downstream consumers that may require a particular feature combination.

It is also a file-format package where old data matters. Users may be less interested in novelty than in long-term readability, stable tools such as h5dump and h5repack, and a build that keeps older files accessible while still supporting newer platform and performance work.

For Homebrew specifically, hdf5 is the general-purpose package that many scientific packages find at build or runtime. Its sibling hdf5-mpi exists because the parallel build is important enough, and different enough, to deserve a separate dependency path.

Timeline

• 1987: NCSA began the HDF effort for architecture-independent scientific data sharing.
• 1996: HDF maintainers identified scaling limits in the older HDF line.
• 1998: HDF5 was implemented as the scalable successor to HDF4.
• 2005: The HDF Group was incorporated as a not-for-profit company.
• 2006: The HDF Group began operating separately in the University of Illinois Research Park.
• 2020: HDFGroup/hdf5 official source repository was created on GitHub.
• 2026: HDF5 2.x changelog material documents continuing work on build systems, tools, plugins, and compatibility.

Related projects

• HDF4 is the older HDF line that HDF5 superseded for many workloads.
• Parallel HDF5 is the MPI-oriented build and programming model for parallel I/O.
• HDFView is the graphical viewer in the broader HDF ecosystem.
• HSDS is The HDF Group's highly scalable data service for HDF5-style data access.
• netCDF, h5py, HDF5.jl, and many scientific applications depend on or integrate with HDF5.

Sources

• https://api.github.com/repos/HDFGroup/hdf5
• https://github.com/HDFGroup/hdf5
• https://github.com/HDFGroup/hdf5/blob/develop/release_docs/CHANGELOG.md
• https://support.hdfgroup.org/documentation/hdf5/latest/
• https://support.hdfgroup.org/documentation/hdf5/latest/_intro_par_h_d_f5.html
• https://www.hdfgroup.org/2015/03/04/hdf-group/
• https://www.hdfgroup.org/about-us/
• https://www.hdfgroup.org/solutions/hdf5/