io500
Leonardo achieves another important certification, this time on the efficiency of its DDN exascaler-based storage system, and wins first place with a Bandwdith score of 807.12 GB/s, placing second overall on the Production List of IO500.
The IO500 is one of the most important HPC storage performance rankings in the world. It has been widely recognized by research institutions including universities, national laboratories and commercial companies.
Since November 2017, the IO500 list has been published at major HPC conferences (SC in the United States and ISC in Germany). In this context, I/O performance has become an important indicator for measuring the efficiency of supercomputer applications. In particular, it measures storage performance using two key indicators: data bandwidth (GB/s) and read/write speed (IOPS).
Storage systems are evaluated by simulating different I/O patterns from light to heavy workloads. Scores are obtained from the geometric mean of the performance values in all scenarios.
To be eligible for the Production List, a submission must meet the definition of a “Production System”. The spirit of the Production System definition below is to ensure entries on the Production List are systems that are used by scientists, quants, security teams, data scientists over an extended period of time. Often this means the cluster has a batch scheduler and queues up incoming Production Application execution requests. It also further means that the duration of deployment is much longer than a few days, weeks or months…and is typically measured in years.
A “Production System” is an IO500 submission that:
- consists of a Compute System and Storage System that on a ‘regular‘, ‘frequent‘, and ‘ongoing‘ basis executes Production Applications that generate Production Data
- achieves the highest Reproducibility Score
- has no single point of failure in its Storage System
These important results confirm that the EuroHPC JU machine operated by Cineca allows great flexibility to address the most demanding I/O use cases in terms of bandwidth and IOPS.
