System hardware

The Apollo HPC cluster is the University's latest central HPC system, purchased in 2025.  

Apollo offers:

  • 81 general-purpose compute nodes each with 64 cores (AMD EPYC 9354, 3.25 GHz) and 384 GB of memory.
  • 3 high-memory/visualisation nodes each with 64 cores (AMD EPYC 9354, 3.25 GHz), 2304 GB memory and a NVIDIA L4 GPU.
  • 11 GPGPU computation nodes, each with 64 cores (AMD EPYC 9354, 3.25 GHz), 1536 GB memory and 4 x NVIDIA L40S GPU.
  • 2.5 PB of storage delivered via the Lustre high-performance filesystem.
  • All connected by a 200 Gb/s Infiniband (high-speed, low-latency) network.

As a description of Apollo for grant applications we suggest: "The Apollo HPC system has 81 compute nodes each with 64 cores and 384 GB of memory, 3 high-memory nodes each with 64 cores, 2304 GB of memory and an L4 GPU, and 11 GPU nodes each with 64 cores, 1536 GB of memory and 4 L40S GPUs, all connected by Infiniband networking and backed by 2.5 PB of storage".