With the support of around 40 researchers, the University has allocated a UDF grant for establishing a new High Performance Computing Cluster for scientific computation. I am glad to report that the installation of the Cluster is well underway; and we plan to make the Cluster available for use of our research community in a few months’ time.
The Cluster consists of 104 general purpose nodes of 20 cores and 96GB RAM each; 3 nodes of 40 cores and 512GB RAM each; 4 nodes with NVidia GPUs and 2 nodes with Intel Xeon Phi CPU-accelerators; 150 Terabyte disk storage; and 40Gb/s InfiniBand internode connection. The Cluster has a theoretical performance of 104 Tera-FLOPS (FLoating-point Operation Per Second), which is roughly 8.5 times the power of the University’s biggest “Gridpoint” HPC cluster currently in production use. Moreover, the Cluster is equipped with many new features that our existing systems do not possess.
Interested researchers may contact the countersigned for details.
System Diagram of the New HPC Cluster
High Performance Computing Team
Tel: 3917 8631