The Centre helps Physics researchers choose, develop and run the computing approach that best matches their scientific problem. You do not need to know in advance whether your project needs high-throughput computing, GPUs, cloud resources or specialist hardware.
Starting a new project?
Eligible projects may be able to use a time-limited seedcorn allocation for development and testing before applying for larger resources.
What we help with
Large-scale and high-throughput processing
Batch analysis, simulation and repeated processing across many independent workloads, on departmental systems and through the grid.
AI and GPU computing
Developing and running machine-learning and GPU-accelerated workloads on suitable departmental and College systems.
Storage and data movement
Planning for active datasets, bulk storage and transfers between local and external facilities.
Cloud and distributed computing
Virtualised and federated environments for projects that need flexible services or access across several sites.
Real-time and specialist systems
Connecting researchers with relevant departmental expertise in low-latency computing, instrumentation, FPGAs and other specialist architectures.
Software and workflows
Onboarding, containerisation, software distribution, performance testing and reproducibility, with routes to research software engineering support.
How access works
- Contact us to discuss the scientific objective, expected workload, data requirements and timescale.
- We help identify the most suitable systems, whether departmental, College or national.
- Eligible projects can start with a seedcorn allocation, a time-limited allocation for testing and development, subject to availability and the rules of the relevant service.
- We help with onboarding, software environments, optimisation and data movement.
- For larger-scale work after the seedcorn stage, we help researchers prepare applications to GridPP or IRIS, or identify the appropriate registration and allocation route through the Research Computing Service.
Get in touch by emailing data-intensive-physics@imperial.ac.uk.
Computing resources
Resources are organised in layers with different owners, purposes and access arrangements.
| Resource layer | Primary role | Access route |
|---|---|---|
| Departmental systems | Development, testing, teaching and specialist workloads within Physics. | Managed by the Centre. Contact us to discuss eligibility. |
| Research Computing Service | University-wide high-performance computing, GPU platforms, storage and research software engineering. | RCS registration and allocations. |
| GridPP | Distributed high-throughput computing and storage for data-intensive science. | GridPP eligibility, with seedcorn and formal allocation routes. |
| IRIS | Digital research infrastructure for STFC-supported science, developed by a cooperative of UK institutions. | IRIS eligibility, with seedcorn and formal allocation routes. |
Infrastructure we operate
The Grid Cluster
Imperial’s GridPP site contributes to LondonGrid, one of the UK’s four regional Tier-2 centres within GridPP and the Worldwide LHC Computing Grid. It supports workloads for international collaborations including the LHC experiments, neutrino experiments, dark matter searches and the Euclid and PLATO space missions. Access is through standard grid interfaces.
The GridPP DIRAC service
We host the DIRAC service that GridPP provides to help smaller research communities manage workloads across the UK grid.
The IRIS OpenStack Cloud
Our cloud service supports IRIS projects.
Departmental batch and storage
We operate local batch processing and storage for departmental projects.
All services follow the security procedures of the infrastructures they belong to. Access to the Research Computing Service is covered by the RCS access policy.