Hardware¶
Our HPC platform is located at ITS MCR (P404). It includes the following components:
| Item | Qty | Purpose |
|---|---|---|
| NVIDIA DGX A800 | 3 | Compute nodes with total 24 A800 GPU |
| Dell PowerEdge R650xs Server | 2 | Login nodes for job submission |
| Dell PowerScale A300L NAS | 1 | Centalised Storage |
| Dell PowerSwitch S5232F-ON | 1 | Network interconnect between all devices |
| Mellanox MQM8700 Infiniband Switch | 1 | Network interconnect between all compute nodes |
Compute Node¶
| Item | Description |
|---|---|
| Name | NVIDIA DGX A800 |
| GPU | 8x NVIDIA A800 80GB SXM Tensor Core GPU |
| CPU | Dual AMD Rome 7742 |
| Memory | 2TB |
| Networking | 8x Sinlge-Port NVIDIA ConnectX-6VPI 200Gb/s InfiniBand 2x Dual-Port NVIDIA ConnectX-6 VPI 10/25/50/100/200 Gb/s Ethernet |
| Storage | OS: 2 x 1.92TB M.2 NVME drives Internal Storage: 8 x 3.84TB U.2 NVMe drives in RAID0 |
| OS | Ubuntu 22.04.4 LTS |
Login Node¶
| Item | Description |
|---|---|
| Name | Dell PowerEdge R650xs |
| CPU | Dual Intel Xeon Silver 4316 @2.2GHz |
| Memory | 128GB |
| Network | 2 x 100GbE QSFP28 |
Storage System¶
| Item | Description |
|---|---|
| Name | Dell PowerScale A300L NAS |
| Disk | 8 x 800GB Cache SSD 60 x 12TB HardDrive |
| Usable Space | ~600TB |