Our new 8x A100 GPU system has arrived (Sept 2022)!
- 2x AMD EPYC Milan 7763
- 2TB ECC Ram
- 21 TB Scratch Space
- 8x A100 Nvidia GPUs
We also have a dedicated cluster for our lab, built by us from commodity parts bought individually. It is fast…
- Total of 544 cores/1088 threads
- Login node
- 48-thread 2-socket Xeon E5-2697v2 login node
- Cuda-based GPU capability NVIDIA Quadro K5000
- 128 GB ECC Ram
- 2TB SSD NFS-mounted
- 12TB HDD NFS-mounted
- Compute nodes
- 24-socket 960-thread Xeon E5-2698v4
- 512 Gb per node ECC Ram DDR4-2400
- 512GB SSD local scratch on each node
- Storage
- 80TB HDD Lustre filesystem with ZFS backend
- Storage server 2xE5-2630v4 512GB ECC DDR4-2400
- Metadata server 1xE5-2698v4 256GB ECC DDR4-2400
- Infiniband (Mellanox) FDR 56 Gbps Switch w/ QSFP Interconnects
- Software
- Centos Linux
- SLURM job management (with no time limits on runs!)
- Available Packages: Ganglia, Nagios, modules, Matlab, Mathematica, LAMMPS, GULP, Quantum Espresso, ANSYS, Parallel Studio, code development tools & libs, Jupyter, Atom, Spyder …
Contact Prof Chung for more information.
We also work with University and State computing resources.