MCSR currently has three supercomputers: Sequoia, Catalpa, and Maple.
Sequoia is a cluster and MCSR’s oldest supercomputer. It consists of 124 nodes with a total of 1304 CPU cores interconnected with a mixture of DDR and QDR Infiniband. Nodes have up to 35GB of memory and a varying number of CPU cores. It has the most software installed, and is a good system to start with if you don’t need a lot of memory.
Catalpa is MCSR’s only single-image, shared memory system, and should only be used for large memory jobs that are unable to be run across multiple nodes on one of the clusters. It is an SGI UV 2000 with 320 processor cores and 2.5 terabytes of shared memory.
Maple is a Cray cluster with a theoretical peak performance of 116 TFLOPS, 73 compute nodes, a total of 2,164 CPU cores, 29 NVIDIA Kepler K20 GPUs, 4 NVIDIA Tesla P100 GPUs, and 6.6TB of memory. It is the future workhorse cluster of MCSR, but does not yet have the installed software base of Sequoia and Catalpa. This cluster is grant funded in part by the National Science Foundation (CHE-1338056).
All systems run SUSE Linux Enterprise Server 12 and are connected to MCSR’s storage system.