Warning

Sheffield’s access to Bede ended on 31 July 2026. The service continues; see the operator’s website for current access information.

Bede (Tier 2 GPU cluster)

Bede is a EPSRC-funded ‘Tier 2’ (regional) GPU-equipped HPC cluster. Sheffield researchers previously had access through the N8 Research Partnership. This page retains information about that service; current eligibility and specifications are maintained on the N8 CIR Bede site.

NB the system was previously known as NICE-19.

Suitable workflows

This system is particularly well suited to supporting:

  • Jobs that benefit from distributing work between multiple GPUs and possibly multiple nodes.

  • Jobs that require much movement of data between CPU and GPU memory.

  • In particular deep learning and machine learning workflows that meet either of the above criteria.

Noteworthy features of the system

The following specifications are retained for reference. Consult the operator’s documentation for current hardware availability.

  • 32x GPU nodes (IBM AC922 nodes) each with

    • 2x IBM POWER9 CPUs

    • 2x NVIDIA V100 GPUs per CPU

    • Each CPU is connected to its two GPUs via high-bandwidth (150GB/s), low-latency interconnects (NVLink), which helps if you need to move lots of data to/from GPU memory

    • 512 GB RAM

  • 4x ‘inference’ nodes (IBM IC922 nodes) each with

  • 1x ‘ghlogin’ node with

  • 7x NVIDIA Grace Hopper Superchip (GH200 480GB) nodes each with

    • 1x NVIDIA Grace CPU (72 Arm Neoverse V2 cores)

    • 1x NVIDIA H100 96GB GPU

    • The CPU and GPU are connected via 900 GB/s NVLink-C2C interconnect, which helps if you need to move lots of data to/from GPU memory or over-subscribe the GPU

    • 480 GB LPDDR5X RAM

  • 1x dual Grace Hopper superchip node with

  • High-bandwidth, low-latency networking between nodes (100 Gb/s EDR Infiniband)

  • High-performance parallel file system (Lustre)

  • Slurm job scheduler

Further information

See the N8 CIR’s Bede site for:

  • Documentation on how to use the system

  • Information on support for participating institutions

  • How to register a project

  • Hardware specifications

  • Available software

  • How to acknowledge Bede and the N8 CIR in publications

  • Bede and N8 CIR logos

For current Sheffield HPC access options, see external HPC facilities.