Our group has equipped with very good computational and storage resources.
We have one CPU-GPU cluster with 27 nodes, among which 25 nodes are for computation and the other two nodes are for management and data storage purposes. Each computational node has 2 CPUs (AMD Opteron 6272，8 cores per CPU), one PCI-E X16 GPU card（ Nvidia Tesla™ C2050 3GB and Nvidia Tesla™C2075 6GB Fermi CUDA framework with 448 CUDA cores). This cluster is mainly used for MPI and/or GPU type of computation. Each node has 32 G memory. This CPU-GPU cluster is equiped with 24T (2x12 T) storage disks. This cluster is shared with a few other faculty members in the geophysics program. [USTC internal IP: 126.96.36.199]
We also have three CPU servers (great for OpenMP):
Server 1: 2 CPUs (AMD Opteron 6272，2.1 GHz, 32cores = 2*16 cores per CPU), 128 G memory, and 24 T (2x12 T) storage disks. [USTC internal IP: 188.8.131.52]
Server 2: 2 CPUs (AMD Opteron 6376，2.3 GHz, 32 cores = 2*16 cores per CPU), 128 G memory, and 8 T (2x4 T) storage disks. [USTC internal IP: 184.108.40.206]
Server 3: 4 CPUs (AMD Opteron 6376，2.3 GHz, 64 cores = 4*16 cores per CPU), 256 G memory, and 60 T storage disks. [USTC internal IP: 220.127.116.11]
For account application for the CPU-GPU cluster or the CPU server, please contact our server administrator Shican Li (63601095, email@example.com).
These cluster/servers can be only logged in through USTC IP. If you want to access these clusters from outside, please contact the administrator.
The detailed usage of these computational resources can be downloaded here: CPU-GPU cluster user manual and CPU server manual.