HPE Cray Supercomputing Storage Systems E2000are the even faster sibling of Cray ClusterStor E1000 Storage Systems that have seen unprecedented adoption in the supercomputing market with more than half of the aggregate compute power of the world’s Top 100 supercomputers relying on Cray ClusterStor E1000 Storage Systems.^[1] The E2000 systems preserve all the value propositions that have made Cray ClusterStor E1000 Storage System the de-facto standard in high-end supercomputing, but more than doubles the performance through the use of new storage controllers. The E2000 systems are the perfect choice for supercomputers and large compute clusters running modeling and simulation, artificial intelligence and high performance data analytics applications in integrated workflows on one machine.

What is new supporting point:
Double the write throughput: The new storage controllers enable up to 140 GB/s write throughput per 2U which translates to up to 2 TB/s in the base rack and 2.2 TB/s for each expansion rack.
Ultra-high IOPS: The new storage controllers provide up to 2.7 million IOPS per 2U which translates to up to 37 million IOPS in the base rack and 43 million IOPS for each expansion rack.
400 Gbps connectivity: Support for HPE Slingshot interconnect 400 for use with HPE Cray SC EX supercomputers and for InfiniBand NDR or 400 Gbps Ethernet for use with HPC/AI compute clusters.
Double the read throughput: The new storage controllers enable up to 190 GB/s read throughput per 2U which translates to 2.6 TB/sec in the base rack and 3.0 TB/s for each expansion rack.