Euler is a heterogeneous compute cluster that has been operating since January 2010. Its hardware makeup includes a variety of Intel x86 and IBM POWER processors, NVIDIA GPUs, and additional computational hardware. Euler currently hosts 932 CPU cores, 8.9 TB RAM, 76 NVIDIA GPUs, and 398TB of RAID 6+0 backed distributed storage. The estimated peak FLOP rate of this collective hardware is around 759 single precision and 127 double precision TFLOP/s.
Euler is curated by Dan Negrut’s Simulation Based Engineering Laboratory (SBEL) along with hardware contributions from other campus groups. It is currently the flagship supercomputer of the Euler Co-Op.
View the current status of the Euler Cluster.
Last updated: Mar 2018
MSCData is a relatively small hardware stack providing three homogeneous compute nodes, and a large, 144TB SAS-connected backend for the storage of Scanning Transmission Electron Microscopy (STEM) imaging data. Each compute node is a Dell PowerEdge R730 comprised of two Intel Xeon E5-2670v3 processors, 192GB RAM, and a NVIDIA Tesla M10 GPU (provides 4x logical GPUs to software).
MSCData is curated by SBEL on behalf of the Voyles Group. It was assembled in January 2018.
View the current status of MSCData.