Dear Colleagues,
We are pleased to announce the launch of the HPC I/O Archive website https://hpcioanalysis.zdv.uni-mainz.de/ and would like to invite you and your team to contribute traces and explore its online analysis capabilities.
Core Purpose
===========
At the intersection of AI/ML and HPC, the analysis of I/O patterns is critical for both application and storage optimization. Therefore, we are building a repository of Darshan-based I/O traces to foster interaction among system and storage researchers, developers, and application specialists from diverse scientific fields.
Key Highlights
============
Analytics Hub: The website goes beyond storing traces and provides interactive graphs for comprehensive I/O analysis.
Contextual Analysis: Our platform captures application I/O data and metadata, enabling comparisons of I/O traces across diverse HPC environments.
Open Collaboration: Embracing the principles of open science, our framework simplifies trace submission and encourages cross-disciplinary interactions. Each trace gets its own DOI, so that traces can be individually referenced to recognize your efforts in collecting traces.
Your active participation is critical to the success of the initiative. For details on how to contribute, please visit the “Getting Involved” section of our website.
We would like to thank early contributors, such as those from the EuroHPC projects ADMIRE and IOSEA. We are also happy to thank the Darshan team for an exceptional collaboration, which is central to this project.
We hope that you will benefit from using the HPC I/O Trace Archive and are looking forward to your feedback.
Thank you and best regards,
André Brinkmann (for the developer team)
Darshan users,
Darshan version 3.4.4 is now officially available for download here: https://www.mcs.anl.gov/research/projects/darshan/download/. This point release includes a few minor bug fixes:
* Fixed bug leading to inconsistent heatmap record shapes when Darshan shared file reductions are disabled
* Also added a darshan-util library fix to resolve this inconsistency on already impacted logs (any generated with 3.4.0+ versions of Darshan)
* Added workaround for potential undefined symbol errors for 'H5FD_mpio_init' when LD_PRELOADing an HDF5-enabled runtime library
* Bug triggered by 1.13+ versions of HDF5
We have also released PyDarshan 3.4.4.0 on PyPI, though this is just to track the 3.4.4 darshan-util library. There are no modifications to PyDarshan functionality.
Documentation for Darshan and PyDarshan is available here: https://www.mcs.anl.gov/research/projects/darshan/documentation/.
Please report any questions, issues, or concerns with this release using this mailing list or by opening an issue on our GitHub: https://github.com/darshan-hpc/darshan.
Thanks!
--Darshan team
