Loughborough University
Leicestershire, UK
LE11 3TU
+44 (0)1509 222222
Loughborough University

IT Services : High Performance Computing

Use Cases


Introduction

There are a number of use cases for the use of Hydra, some of which are indicated below, and use cases may overlap.

If you have a research computing requirement which seems to fall outside these use cases please contact the team (See Contacting Support ) to discuss your requirements.

Funding Use Cases and Allocation Policy

The below summarises the allocation policy. An updated policy will be available soon .

Funded Research

Research council or charity funded research that uses the capabilities of the system as suggested below.

PhD Research

PhD research that uses the capabilities of the system for PhD students

Seed Time

Up to 20,000 core hours can be awarded for 'Seed' time for unfunded exploration of the system to determine if the system provides benefits prior to a funded usage as above.

Technical Use Cases

See Hydra HPC System Capabilities for system technical information.

Parallel, single node usage

Multicore parallel simulations that run on a single processing node. I.e. 12 to 20 cores, using MPI , OpenMP , or other forms of parallelism, or using a package (See Installed Software that uses such parallelism).

Parallel, multinode usage

Parallel simulations that need many nodes, i.e. more than 20 cores, using MPI , OpenMP , or other forms of parallelism, or using a package (See Installed Software that uses such parallelism).

High throughput computing

Serial (single processor core) programs, or parallelism of fewer than 12 cores, but where many runs are needed, e.g. loosely coupled systems of jobs, ensemble models, large volumes of processing. See Highthroughput and trivially parallel for more information.

Visualisation

Visualisation of data or results using one of a number of fast GPU-accelerated visualisation node, e.g. using VTK or paraview

Post processing

The use of high memory nodes for processing that joins data together that may need high memory requirements for sifting and joining. See Hydra HPC System Capabilities for information on the specification of the nodes for ths purpose.