Table of Contents

What Are My Compute Options?

Consider the three options below and the questions asked for each:

Desktop/Laptop

High Performance Computing Facility

Dedicated Compute

  • You only need to run a few tens of computations or they can only be sequential e.g. the next step needs the result of the previous step.
  • Your workload fits entirely within the hardware resources of your device
  • Your workflow is partially or fully interactive (e.g. you interact with an application/windows/buttons, or provide data entry at several points in the process)
  • Your application is only for Microsoft Windows or Apple Mac OS

Using Desktop/Laptop Equipment

  • Do you need more hardware resources (CPU cores, RAM) than available on your existing desktop/laptop?
  • Do you need to access specific resources that are not available on your existing desktop/laptop (e.g. high-end GPU)?
  • Do you need very high throughput disk/file storage during computation?
  • Does your computation need to run outside of office hours?
  • Does your computation run for many hours or days?
  • Do you need to run other software while your computation is running?
  • Do you have regular periods/processes at which your computation runs (e.g. end of month, over a weekend, etc)?
  • Your computation run under Linux and you do not have a Linux environment available on an existing desktop/laptop?
  • Do you need to run the same application/code many times over (possibly hundreds or even thousands of times) with different data?
  • Do you want to help contribute to shared resources used by your colleagues across the University?

Using HPC Facilities

  • Do you need more resources than available on an existing Desktop/Laptop, but which are not catered for by our HPC? (e.g. specific CPU architecture)
  • Do you have a workflow or process which is not compatible with a shared facility?
  • Are you tied to a proprietary application not available on Linux? (e.g. a commercial package on Microsoft Windows)
  • Do you need to run long-running application services accessed by client software? (e.g. databases, web services)
  • Do you need to accept, process or return data in real-time?
  • Do you need to directly interface with other, specialist equipment or instrumentation?

Using Dedicate Local Compute Equipment

Desktop/Laptop Equipment

In some cases it can be more effective to use your existing University provided desktop/laptop device, especially if the computation you need to undertake is infrequent or fits entirely within the resources of your existing device.

In these cases, whilst porting the workflow to a HPC environment may generate some performance improvements, the (human) resources needed to set up that HPC environment and training necessary may not be a sensible use of time.

In the case where your computation requirement is partly or wholly based upon Windows/Mac OS software, then it is likely impractical to port to an HPC environment unless the advantages of moving to that environment are substantial (for example if the initial work for a project was built on prototype software for Windows, but the project will run for several years and the redevelopment time needed to port to Linux would allow a substantial increase in quantity of results to be obtained over the lifetime of that project).

In this case we would recommend speaking to the RSE team first to establish the practicalities of this undertaking.

HPC Facilities

If you answered “Yes” to any of the High Performance Computing Facility questions above, then you should probably be using HPC to run your workloads:

There are a few examples where the use of a shared HPC facility may not be the most viable option. Check that the following do not apply to you:

Reconsider Using HPC If…

  • The information you are processing is sensitive (e.g. health data, commercially sensitive)
  • The information you are processing is restricted in some way (e.g. commercial restrictions)
  • The work being undertaken is for commercial gain or a commercial third-party (e.g. outside of regular teaching or funded research activity)

If you have decided that HPC is the best solution for you, then you have two possible routes to access HPC:

Newcastle University HPC facilities

The Newcastle University HPC facilities (as of 2025, this includes Rocket and the incoming Comet facility) are available for all members of staff and students to use.

Rocket

Rocket is a legacy system, due to be replaced in 2025, which is available free-of-charge to all members of staff and students. It is largely a CPU-based system, with approximately 5400 CPU cores, though has a small amount of older Nvidia GPU cards.

Rocket is a Linux based system, using the Slurm workload manager. There are very minimal graphical capabilities provided by X11 tunnelling over SSH, but these should not be relied upon to run your workloads.

You do not need to allocate any funds to use Rocket, and it is a first-come, first-served system, where all users have equal access to all resources.

Comet

Comet is our new HPC facility, introduced in 2025, and operates under a split access model. All staff and students may access the facility without charge, but may only access a certain percentage of the available resources; the remainder of the resources are reserved for projects which have contributed funding to the facility.

Comet is built from a number of different server compute 'node' types, including a total of approximately 15000 CPU cores, a number of nodes with very large RAM, as well as substantial number of very powerful GPU cards.

Like Rocket, Comet is a Linux based system using the Slurm workload manager. Unlike Rocket however, Comet has fully supported graphical environments available to run more interactive computation workloads (RStudio, Jupyter as well as full Linux X11 desktop sessions).

The cost of using the reserved resources can be estimated using our Cost Calculator & Resources tooling on this website.

National/Regional HPC facilities

Whilst we provide a range of HPC facilities within Newcastle University suitable for most purposes, there may be times where you have a specific requirement we are unable to satisfy, or need access to more resources than we are able to provide.

At this point you may want to consider the use of a regional Tier 2 HPC Facility, or even a national Tier 1 system. These are (usually) larger systems (though sometimes they are instead more specialised) that are made available to researchers and academics from across the UK. Depending on your type of work, funding and area of discipline, these may be viable options for you to consider.

Dedicated Local Compute

If you have a need for more advanced compute facilities than can be provided by existing, standard desktop or laptop equipment, and you believe that those requirements cannot be met by our HPC facilities, then there is a third option of requesting dedicated local compute.

Whilst the general principles that we work under are to make the most effective use of existing resources, and to provision new resources in such a way to make them available to the widest range of users, there are a small number of cases where this is not always feasible. In those cases, the University provides a means to capture those requirements and discuss them in an academic-lead forum.

If you have considered the other available options and feel that this is the most suitable way of delivering a solution for your requirements, you can follow the process for Advanced & High Performance equipment requests, below.

Note!

This route should not be used for the provision of dedicated, single-user workstations. Please follow the existing NUIT self-service system for the request of new personal IT devices.

Back to Getting Started