STFC DiRAC Project Office 2014-2017
Lead Research Organisation:
University College London
Department Name: Physics and Astronomy
Abstract
It is now accepted that "computational science, the scientific investigation of physical processes through modelling and simulation on computers," is the third pillar of science, complementing and extending theory and experimentation" (International Review of Research using HPC in the UK, 2005).
The STFC DIRAC High Performance Computing Facility supports the simulation and modelling requirements of those communities that are represented by the Particle Physics Advisory Panel, Nuclear Physics Advisory Panel, Particle Astrophysics Advisory Panel, Astronomy Advisory Panel and the Solar System Science Advisory Panel.
Theoretical research in Particle Physics, Particle Astrophysics, Nuclear Physics, Astrophysics, Solar System and Planetary Science are world leading and DiRAC-enabled research has been published in high-impact refereed journals; 160 articles in 2010, 381 in 2011 and 2012 and over 250 papers produced in 2013 (see our publications list in our annual reports along with Science highlights at http://www.dirac.ac.uk/science.html ).
The STFC DIRAC High Performance Computing Facility supports the simulation and modelling requirements of those communities that are represented by the Particle Physics Advisory Panel, Nuclear Physics Advisory Panel, Particle Astrophysics Advisory Panel, Astronomy Advisory Panel and the Solar System Science Advisory Panel.
Theoretical research in Particle Physics, Particle Astrophysics, Nuclear Physics, Astrophysics, Solar System and Planetary Science are world leading and DiRAC-enabled research has been published in high-impact refereed journals; 160 articles in 2010, 381 in 2011 and 2012 and over 250 papers produced in 2013 (see our publications list in our annual reports along with Science highlights at http://www.dirac.ac.uk/science.html ).
Planned Impact
The high-performance computing applications supported by DiRAC typically involve new algorithms and implementations optimised for high energy efficiency which impose demands on computer architectures that the computing industry has found useful for hardware and system software design and testing.
DiRAC researchers have on-going collaborations with computing companies that maintain this strong connection between the scientific goals of the DiRAC Consortium and the development of new computing technologies that drive the commercial high-performance computing market, with economic benefits to the companies involved and more powerful computing capabilities available to other application areas including many that address socio-economic challenges.
The DiRAC Facility is designed to enable breakthrough science in the areas of lattice quantum chromodynamics, hadron physics, beyond the standard model physics, black hole Physics, whole system modelling of stars, solar systems, galaxies, the local universe and large scale structures; the early Universe and its evolution, cosmology and the nature of dark energy and dark matter, the formation and evolution of stars and planets and larger structures, and the chemistry of the Cosmos.
DiRAC researchers have on-going collaborations with computing companies that maintain this strong connection between the scientific goals of the DiRAC Consortium and the development of new computing technologies that drive the commercial high-performance computing market, with economic benefits to the companies involved and more powerful computing capabilities available to other application areas including many that address socio-economic challenges.
The DiRAC Facility is designed to enable breakthrough science in the areas of lattice quantum chromodynamics, hadron physics, beyond the standard model physics, black hole Physics, whole system modelling of stars, solar systems, galaxies, the local universe and large scale structures; the early Universe and its evolution, cosmology and the nature of dark energy and dark matter, the formation and evolution of stars and planets and larger structures, and the chemistry of the Cosmos.
Publications
Becker C
(2020)
Proca-stinated cosmology. Part I. A N -body code for the vector Galileon
in Journal of Cosmology and Astroparticle Physics
Beg R
(2022)
Evolution, Structure, and Topology of Self-generated Turbulent Reconnection Layers
in The Astrophysical Journal
Bena I
(2019)
Holographic dual of hot Polchinski-Strassler quark-gluon plasma
in Journal of High Energy Physics
Bending T
(2022)
Supernovae and photoionizing feedback in spiral arm molecular clouds
in Monthly Notices of the Royal Astronomical Society
Benitez-Llambay A
(2020)
The detailed structure and the onset of galaxy formation in low-mass gaseous dark matter haloes
in Monthly Notices of the Royal Astronomical Society
Bennett E
(2020)
S p ( 4 ) gauge theories on the lattice: Quenched fundamental and antisymmetric fermions
in Physical Review D
Bennett E
(2020)
Color dependence of tensor and scalar glueball masses in Yang-Mills theories
in Physical Review D
Bennett E
(2023)
Update on SU(2) with one adjoint Dirac flavor
Bennett E
(2019)
Sp (4) gauge theories on the lattice: Nf = 2 dynamical fundamental fermions
in Journal of High Energy Physics
| Description | We support all of PPAN science. The main highlights have been the modelling of the first Gravitaitonal Waves (the discovery of which won the 2017 Nobel Prize for Physics), the accurate properties of the B quark, and the most detailed models of Galaxy formation bythe VIRGO consortium |
| Exploitation Route | They have opened up new areasin Physics, particularly gravitational waves, calculating particle properties to be measured by the LHC and the use of baryonic physics in comsological models. Our advances in software and hardware design are applicable to all fields of study and commerce. a growing list of technology companies (ARM, Dell, HPE, IBM, Intel, Mellanox, Nvidia). We have appointed an Innovation Director (Dr. Jeremy Yates) and used the 2017/18 BEIS capital investment to kick-start new strategic industrial collaborations. DiRAC's innovation strategy is aligned with the UK Industrial Strategy white paper. In particular, our strategy aims to: 1. Enhance engagement with industry around the challenges of machine learning and data intensive science. 2. Enhance engagement with broader Industry, the Hartree Centre, and other sectors on the exploitation and use of new technologies and of DiRAC computational resources. 3. Expand innovation programmes with industry partner(s) at all sites, including both component-level and system-level co-design and optimisation. 4. Further expand our comprehensive, in-house HPC skills training portfolio by working with industry partners and the eight STFC Centres for Doctoral Training in Data Intensive Science , thereby enhancing HPC skills training across the UK and increasing the net out-flow of upskilled workers into the UK economy. 5. Expand engagement with industrial partners on software engineering work for particle physics, astronomy and general HPC codes. 6. Increase industrial income and explore potential options for innovative business models to support the recurrent costs of services. |
| Sectors | Digital/Communication/Information Technologies (including Software) |
| URL | http://www.dirac.ac.uk |
| Description | DiRAC has set up three Intel parallel computing centres which are being used to drive system design for heterogeneous architectures with SGI, writing maths libraries for intel and writing fine grained parallel task management libraries for intel. The work with SGI won an award at the SC15 for Best use of High Performance Data Analytics..• 4 EPSRC iCASE studentships co-funded by Intel/HP (Leicester), Intel/Cray (UCL), Intel/SGI (Cambridge), Intel/Lenovo (Durham); all in areas of advanced software design for new technologies. • DiRAC has secured 2 four-year PhD studentships with Mellanox/UCL and Lenovo/UCL in the areas of advanced cluster switch design and high volume and velocity machine learning, respectively. • DiRAC's Chief Technical Officer (CTO), Professor Peter Boyle, is the Co-Design leader for the Intel-Alan Turing Institute (ATI) Many Core Architecture Design team based in Edinburgh (one of only three teams worldwide). DiRAC's expertise provides the foundation upon which the ATI secured this unique international partnership with Intel's HPC architecture group, which has placed five hardware architects in Edinburgh, with Professor Boyle appointed both an Alan Turing Fellow and Co-Design Leader for the ATI. The two hardware architects hired by Intel to work with the ATI were two former DiRAC Research Software Engineers.• Support for STFC Centres for Doctoral Training (CDT) in Data Intensive Science - DiRAC is a partner in five of the eight of the newly established STFC CDTs, and is actively engaged with them in developing industrial partnerships. DiRAC is also offering placements to CDT students interested in Research Software Engineering roles. • Innovate UK - DiRAC was invited to join an InnovateUK bid entitled "High Performance Cloud for Artificial Intelligence (HiPerCloud-AI)" with StackHPC, Concertim Infrastructure Management, VScaler and Intelligent Voice. The bid, submitted in response to the "Emerging and Enabling" call for proposals in November 2017, focussed on work aimed at the elimination of the disablingly large latency caused by Cloud Operating Systems when parallel codes are run on cloud-operated clusters. If successful, it will give the UK practical experience in building, deploying and running weakly and strongly parallel applications on cloud platforms.The £9M DiRAC-2.5x intervention alone generated more than £1.5M of inward investment in addition to significant vendor discounts. These investments include: • Installation of an ARM-based cluster at the University of Leicester as a co-design project to increase the value of ARM technology in HPC (value: approx. £1M); • Support for co-design work with Intel at Edinburgh (value: approx. £0.25M per annum); • Co-funding for DiRAC Technical Manager position at Edinburgh (value: £45k); • Co-funding for a DiRAC Research Software Engineer at Leicester (value: £45k); • Support for co-design projects in flash technologies to accelerate storage access (Cambridge/Dell), flash technologies to allow simulation check-pointing (Durham/Dell), Authentication, Authorisation & Accounting Infrastructure (Edinburgh/HPE), Hierarchical Storage Management (Leicester/HPE); • Support for skills training workshops for DiRAC users and technical support staff. |
| First Year Of Impact | 2016 |
| Sector | Digital/Communication/Information Technologies (including Software) |
| Impact Types | Economic |
| Description | DiRAC 2.5x Project Office 2017-2020 |
| Amount | £300,000 (GBP) |
| Organisation | Science and Technologies Facilities Council (STFC) |
| Sector | Public |
| Country | United Kingdom |
| Start | 02/2018 |
| End | 03/2020 |
| Title | Citation analysys and Impact |
| Description | Use of IT to determineacademic impact of eInfrastructure |
| Type Of Material | Improvements to research infrastructure |
| Year Produced | 2017 |
| Provided To Others? | Yes |
| Impact | Understood emerging trends in DiRAC Science and helped decide the scale and type of IT investments and direct us to develop new technologies |
| URL | http://www.dirac.ac.uk |
| Description | AAAI for the UK NeI |
| Organisation | Jisc |
| Country | United Kingdom |
| Sector | Public |
| PI Contribution | PI of RCUK pilot project for AAAI |
| Collaborator Contribution | Software development and testing at 8 UK HEIs and ROs |
| Impact | SAFE+ASSET AAAI service |
| Start Year | 2016 |
| Description | DiRAC Intel Parallel Computing Centres |
| Organisation | Intel Corporation |
| Department | INTEL Research |
| Country | United States |
| Sector | Private |
| PI Contribution | Organised dicussions which lead to the setting up of 3 INTEL Parallel Computing Centres, at Durham, Edinburgh and Cambridge. Co-ordinated bids and held pre-application discussions with INTEL and the three HEIs to ensure INTEL buy-in to the proposed projects |
| Collaborator Contribution | Awards of £450k to fund 3 programmers for 2 years to i) build maths libraries for the KNL processor, (ii) build fine grain parallel job management libraries to mitigate the load balancing issue for weak balancing codes and (iii) develop heterogeneous architectures for data intensive problems |
| Impact | multi-disciplinary In progress |
| Start Year | 2014 |
| Description | STFC Centres for Doctoral Training in Data Intensive Science |
| Organisation | University of Leicester |
| Department | STFC DiRAC Complexity Cluster (HPC Facility Leicester) |
| Country | United Kingdom |
| Sector | Academic/University |
| PI Contribution | Support for STFC Centres for Doctoral Training (CDT) in Data Intensive Science - DiRAC is a partner in five of the eight of the newly established STFC CDTs, and is actively engaged with them in developing industrial partnerships. DiRAC is also offering placements to CDT students interested in Research Software Engineering roles. |
| Collaborator Contribution | Students to work on interesting technical problems for DiRAC |
| Impact | This is the first year |
| Start Year | 2017 |
| Title | Collaboration with Atempo |
| Description | Tape to Tape data transfter between DiRAC sites. |
| Type Of Technology | Software |
| Year Produced | 2019 |
| Open Source License? | Yes |
| Impact | Proof of COncept that data could be read from Tape stores remotely via a remote file system |
| Description | Member of UKRI E-Infrastructure Expert Panel 2017-2019 |
| Form Of Engagement Activity | A formal working group, expert panel or dialogue |
| Part Of Official Scheme? | No |
| Geographic Reach | National |
| Primary Audience | Policymakers/politicians |
| Results and Impact | Created 7 white papers for UKRI which detailed a Roadmap for future e-Infrastructure funding in the UK |
| Year(s) Of Engagement Activity | 2017,2018,2019 |
| Description | NeI Project Directors Group |
| Form Of Engagement Activity | A formal working group, expert panel or dialogue |
| Part Of Official Scheme? | No |
| Geographic Reach | National |
| Primary Audience | Policymakers/politicians |
| Results and Impact | Reports on AAAI, Data E-Infrastructure, Using Cloud for Research The National NeI Survey 2014, 2015, 2016 Report on Gender in HPC BEIS e-Infrastructure Business Case Integration activities of the NeI |
| Year(s) Of Engagement Activity | 2014,2015,2016,2017 |
| URL | https://neipdg.ac.uk/ |