MAELSTROM

Machine learning has a significant impact on high-performance computing in general, and on weather and climate predictions in particular. To strengthen HPC and weather and climate (W&C) predictions in Europe, MAELSTROM will:

• Open W&C predictions as a new usage domain for machine learning applications that can exploit Exascale levels of performance
• Develop a near-optimal software environment to develop Exascale-ready machine learning tools that can be used across the workflow of W&C predictions
• Optimise compute system designs for machine learning applications for W&C predictions at the node and system level

Status / Highlights

• The MAELSTROM co-design cycle was closed already once
  (Applications → Software → Hardware → Application)
• MAELSTROM datasets already comprise 16 TB of data which are documented, published, and available for download
• MAELSTROM applications are already used by industry as reference benchmarks for HPC performance and machine learning application quality
• The first MAELSTROM Dissemination Workshop attracted 208 registered participants.
• The first MAELSTROM hackathon/bootcamp had more than 30 participants and 16 scientific advisors
• MAELSTROM scientists have already provided more than 30 presentations and published 13 papers