Argonne Leadership Computing Facility expanded the Aurora supercomputer’s storage system to 800 servers, supporting simulations and AI training.

The Argonne Leadership Computing Facility expanded Aurora’s storage system to 800 servers. The announcement concerns storage infrastructure, not an increase to 800 compute servers, GPUs, or accelerator nodes.

The system serves data-intensive workloads across thousands of Aurora nodes. The supplied HPCwire excerpt identifies massive simulations and AI model training as supported workloads and describes Aurora as an exascale supercomputer.

The reported server count does not establish the storage system’s byte capacity or throughput. Although the announcement describes the storage system as among the world’s fastest, the supplied excerpt gives no petabyte capacity, bandwidth, latency, or before-and-after benchmark.