Standard GPU node as found in the Comet HPC facility. These nodes have the same general compute performance as non-GPU nodes, but are fitted with multiple GPU cards for accelerated compute.

These nodes are also connected by low-latency, high speed, 200gbit/sec Infiniband.

It should be noted that the L40S cards are optimised for FP32 and FP16 compute - if your code utilises 64bit floating point, then performance on these cards may be sub-optimal. Consider the use of the larger GPU-L nodes with H100 cards if 64bit performance of your GPU code is essential.

All registered users of the Comet facility may access any of the hours of compute available in the unfunded queues, as listed below. However, your jobs may need to wait longer to run due to the number of resources available. Consider building in costing for your research activity or project for dedicated HPC resources to enable access to the larger pool of resources dedicated for funded queues.

GPU Resources - Unfunded Projects

The maximum hours of GPU compute available across all users of unfunded queues is:
(4 Nodes) x (8 Cards) x (730 Hours/Month) x (25%) = 5840.0 GPU Compute Hours/Month

GPU Resources - Funded Projects

The maximum hours of GPU compute available across all users of funded queues is:
(4 Nodes) x (8 Cards) x (730 Hours/Month) x (75%) = 17520.0 GPU Compute Hours/Month

You can use the Advanced Calculator for the Comet facility to model the potential contributions towards accessing these resources.

Note: You must be logged in to access the costing tool.