Strengthening the HDF5 Ecosystem: 2.1.0 Release & SHINES Updates
Hello HDF Community,

We are excited to share several major updates as we move further into 2026. From the release of HDF5 2.1.0 to new initiatives in safety and security, our team and community are working hard to modernize the HDF5 experience.
 

Contents

 HDF5 2.1.0 is now available 

Building on the milestone of 2.0.0, HDF5 2.1.0 introduces targeted enhancements for modern AI/ML and High-Performance Computing (HPC) workflows. This first minor release in the 2.x series prioritizes emerging hardware support and a more robust developer experience.

Key Highlights:

  • AI/ML Datatypes: Added predefined 4-bit and 6-bit floating-point formats (FP4 and FP6) for immediate compatibility with specialized hardware buffers.

  • Fortran SWMR: New wrappers (including h5fstart_swmr_write_f) bring direct Single Writer Multiple Reader functionality to Fortran users.

  • Performance & Security: Includes significant I/O gains for specific chunked dataset patterns and critical security patches for three newly identified CVEs.

  • Platform Stability: Improved libaec detection and resolved complex number support issues for MSVC.

As always, 2.1.0 is fully backward compatible with 2.0.0 and previous versions. Note that 2.0.0 and later builds are CMake-only, and Autotools are no longer supported.

Read more about this release or download HDF5 2.1.0 today!
 HDF5 SHINES initiatives

Ensuring the longevity and reliability of the HDF5 ecosystem is the driving force behind the HDF5 SHINES (Sustainable HDF5 Infrastructure for Next-generation Engineering and Science) project. A core pillar of this work is our focus on Safety, Security, and Privacy (SSP)—critical areas for modern data management in an increasingly complex digital landscape. Through community-led meetings, new technical resources, and the development of a unified vocabulary, we are working to make HDF5 more robust and resilient for the next generation of scientific discovery.

A Shared Vocabulary for SSP
Understanding "Safety, Security, and Privacy" in the context of high-performance data can be complex. We recently published a blog post, Safety, Security, and Privacy in HDF5: A Shared Vocabulary, to establish common ground for these critical discussions.

Essential Resource: HDF5 2.0.0 Migration Guide
If you are moving your projects to the 2.x series, our HDF5 2.0 Migration Guide is a must-read. It covers the major changes from 1.x and ensures a smooth transition for your build scripts and applications.

Support the SHINES Project: Survey Reminder
We need your help! Please take a few minutes to complete our community survey to support the SHINES project. Your feedback directly impacts how we develop HDF5 to meet the needs of modern scientific research.
 HDFView 3.4.1 Release

The latest version of HDFView 3.4.1, is now available for download. This maintenance release includes important bug fixes and usability improvements to help you explore your HDF5 files more efficiently. Learn more about this release in the release announcement or download today!
 Call the Doctor

Matt Larson headshotCall the Doctor: Matt Larson on Agentic HDF5

Join us for our next Call the Doctor session, where HDF Group software engineer Matt Larson will provide an update on the latest experimental tools for the ecosystem. Matt will go over the skills and tools integrated into Agentic HDF5, a project aimed at simplifying complex data tasks. He will demonstrate how these tools can assist users in implementing more advanced HDF5 features and showcase the project's unique visualization capabilities. Whether you are a power user or just getting started with automation, this session will offer a fascinating look at the "agentic" future of data management.
Click To Join - Tuesdays At 12:20 Central (US/Canada)
You can always check out the Call the Doctor playlist, or subscribe to our YouTube channel to get notified of new content. Subscribing not only notifies you of new content, but helps others discover our content as well. 
 HDF5 NISAR Data Now Available

Image from NASA
Image courtesy of NASA/JPL-Caltech
In collaboration with NASA, we are excited to announce that NISAR (NASA-ISRO Synthetic Aperture Radar) sample data products are now available. These datasets represent a massive step forward in Earth observation and showcase the power of cloud optimized HDF5 to enable seamless, high-performance access to massive Earth science datasets both in the cloud and on-prem. Need help getting your data cloud optimized? Contact us.
 Sponsor Development on GitHub

You can now donate on Github to help fund the work of The HDF Group on HDF5, HDF4, HDFView and others

The HDF Group develops and maintains the feature-packed HDF technology suite. Our engineers put in thousands of hours of work each year refining code, creating new features, answering questions on our free help desk and forum, and more. 

Your sponsorship helps fund the work that The HDF Group does every day to maintain and extend the HDF file format and library. Through a GitHub sponsorship you can donate one time or make a recurring donation. Please let us know if you or your organization need more information before or after your donation. The HDF Group is a 501(c)3 non-profit in the United States, which means your contribution may be tax deductible--please check with a tax professional.
Become a Sponsor on GitHub
Twitter
Website