ExCALIBUR-HEP
Lead Research Organisation:
Imperial College London
Department Name: Physics
Abstract
Abstracts are not currently available in GtR for all funded research. This is normally because the abstract was not required at the time of proposal submission, but may be because it included sensitive information such as personal details.
| Description | The resources provided were used to design and develop a working group that will help towards the software challenges expected in the next decade for particle physics experiments. Particle physics experiments have a large software and computing infrastructure that is needed to collect, store, curate and analyse the data collected and to simulate the experiment behaviour to compare it with theory. One area the wider proposal addressed was the challenge of writing software for FPGAs using high level languages and developing more complex algorithms for particle reconstruction. A production level workload was benchmarked and the computationally intensive areas identified. These were studied very carefully and a deep and quantitative understanding of the challenges was obtained. As part of the project we also engaged with the Intel FPGA group, including receiving a briefing on new technologies and discussions on joint areas of work for the future. |
| Exploitation Route | Existing techniques do not appear to offer a ready solution and avenues for further study, for example using Artificial Intelligence techniques (Graph Neural Networks), have been identified and will be pursued in the context of future applications. |
| Sectors | Digital/Communication/Information Technologies (including Software) |
| Description | SoftWare InFrastructure and Technology for High Energy Physics experiments (2020) at Imperial |
| Amount | £57,173 (GBP) |
| Funding ID | ST/V005936/1 |
| Organisation | Science and Technologies Facilities Council (STFC) |
| Sector | Public |
| Country | United Kingdom |
| Start | 10/2020 |
| End | 03/2022 |
| Description | Collaboration with Intel UK |
| Organisation | Intel Corporation |
| Department | Intel Corporation (UK) Ltd |
| Country | United Kingdom |
| Sector | Private |
| PI Contribution | We explained the challenges that experimental particle physics faces in coming decade with new and upgraded experiments generating significantly increased volumes of data and in particular the realtime processing challenge associated with this data. Giving Intel insight into an expanding market, in particular for their FPGA offerings which are not widely used in our field. |
| Collaborator Contribution | Intel UK briefed us on their newest FPGA technology and plans. They also made available their beta-cloud facility for prototyping and made engineers available to assist. |
| Impact | The main outcome is the establishing of a connection between the partners and briefing each other on challenges and potential solutions and joint participation in workshops. In the future we intend to collaborate more closely via hackathons to drive the collaboration forward to concrete results. |
| Start Year | 2021 |
| Description | Parallel computing training school |
| Form Of Engagement Activity | Participation in an activity, workshop or similar |
| Part Of Official Scheme? | No |
| Geographic Reach | National |
| Primary Audience | Postgraduate students |
| Results and Impact | We organised a training school on parallel programming and data parallelism, in collaboration with another working group in the ExCalibur project and The Hartree Centre. The school was primarily for post-graduate students and early career researchers, but open to all. It was advertised widely in the UK particle physics and lattice QCD communities. A bespoke set of courses and exercises was developed addressing techniques of particular relevance to the two fields of the working groups. Due to the Covid pandemic the school was held by video conference. Feedback was very good, with those who completed the course finding it very helpful. This is a key area looking forward to data processing with accelerators (GPUs, FPGAs etc.) especially and we aim to build on this for the future. Around 10-20 people attended. |
| Year(s) Of Engagement Activity | 2021 |