DDG nodes¶
We have 2 Lenovo ThinkSystem SD530 nodes with Cascade Lake CPUs, each with a single V100 GPU.
| DDG | |
|---|---|
| Count | 2 |
| Processor | 2 x 24 Core Intel Xeon Platinum 8268 (Cascade Lake) |
| Clock speed | Base 2.90GHz ; Turbo 3.90GHz |
| Cores/Node | 48 |
| RAM | 384GB |
| Accessible RAM | ~362GB |
| TMP Size | 745GB |
| Interconnect | 25GB Ethernet/ConnectX-6 |
| Purchase date | February 2021 |
| GPU | NVIDIA Tesla V100 |
| GPU architecture | Volta |
| CUDA cores | 5,120 per chip |
| GPU Memory | 16GiB per GPU |
| CUDA Compute | 7.0 (CUDA version 9 or greater required) |
Accessing the ddg nodes¶
Access to GPU nodes is permitted to QMUL researchers after demonstrating successful use of GPU nodes using the 1-hour queue. Note that access to GPU nodes is not permitted for Undergraduate/MSc students and external collaborators. Please see the using GPUs section for advice on submitting GPU jobs. If you need to limit your job to run on this older class of GPU (for example, if your code cannot run on newer GPUs), then you can specify `-l gpu_type=volta' in your job script.