This article was accepted into the corpus but its outbound wikilinks were never NER-processed — typical at the deepest BFS hop or when the run's entity cap was reached. No expansion funnel to show.
| DiRAC HPC Facility | |
|---|---|
| Name | DiRAC HPC Facility |
| Established | 2009 |
| Location | United Kingdom |
| Type | High-performance computing facility |
| Affiliation | Science and Technology Facilities Council |
DiRAC HPC Facility
The DiRAC HPC Facility is a distributed high-performance computing consortium serving computational science in the United Kingdom, providing specialist supercomputing resources for research in astrophysics, cosmology, particle physics, nuclear physics, and related fields. It supports large-scale simulations and data analysis for projects associated with major institutions and collaborations across the UK and internationally, enabling work that connects to observatories, laboratories, and missions.
DiRAC links national research communities including Science and Technology Facilities Council, STFC Rutherford Appleton Laboratory, University of Cambridge, University of Edinburgh, University of Durham, University of Oxford, and University of Leicester with specialist hardware for computational campaigns. The facility underpins projects aligned with European Space Agency, National Aeronautics and Space Administration, Large Hadron Collider, Square Kilometre Array, and Laser Interferometer Gravitational-Wave Observatory science, supporting workflows that involve institutes such as Institute of Physics, Royal Astronomical Society, Max Planck Society, Princeton University, and Harvard University. Governance and collaborative operation involve consortia including Engineering and Physical Sciences Research Council partners and national research infrastructures like Jisc.
DiRAC’s architecture encompasses heterogeneous clusters optimized for memory-intensive, throughput, and GPU-accelerated workloads, integrating hardware from vendors such as Cray Inc., Dell Technologies, Hewlett Packard Enterprise, NVIDIA, and Intel Corporation. The facility deploys interconnects and storage solutions from Mellanox Technologies and IBM subsystems, with parallel filesystems like Lustre and management tools influenced by OpenStack, SLURM Workload Manager, and Kubernetes paradigms. Nodes are provisioned to serve codebases including GADGET (code), ENZO (software), FLASH (code), CASTEP, VASP, and numerical libraries like ScaLAPACK, FFTW, PETSc, and MPI. Cooling and power strategies draw on practices from European Centre for Medium-Range Weather Forecasts data centres and standards promoted by International Organization for Standardization.
Researchers run simulations and analyses tied to observational programs such as Planck (spacecraft), Gaia (spacecraft), James Webb Space Telescope, Herschel Space Observatory, and experiments including ATLAS experiment, CMS experiment, ALICE (experiment), and LHCb experiment. Numerical relativity studies connect to Einstein Telescope science, while nuclear astrophysics modelling aligns with work at CERN, Oak Ridge National Laboratory, and Lawrence Berkeley National Laboratory. Cosmological structure formation, galaxy evolution, and dark matter simulations integrate with models used by teams at Kavli Institute for Cosmology, Cambridge, Max Planck Institute for Astrophysics, Institute of Astronomy, Cambridge, and Royal Observatory, Edinburgh. Climate and planetary modelling collaborations engage groups from Met Office, NASA Goddard Space Flight Center, and European Southern Observatory.
DiRAC is funded through grants and capital investments from bodies such as UK Research and Innovation, Engineering and Physical Sciences Research Council, and institutional contributions from universities like University of Manchester, University of Sussex, and Cardiff University. Operational oversight involves partnerships with Science and Technology Facilities Council and advisory input from panels including representatives from Royal Society, UK Space Agency, and international stakeholders like European Research Council. Procurement and lifecycle management follow frameworks used by CERN and European Grid Infrastructure, while training and outreach coordinate with Royal Astronomical Society and education initiatives at Institute of Physics.
Performance characterization uses community codes and benchmarks such as HPL (High-Performance Linpack), STREAM (benchmark), and suite tests familiar to groups at Top500 and Green500 listings. Results are compared against installations at Hartree Centre, ARCHER (supercomputer), Joliot-Curie (supercomputer), and international facilities including Oak Ridge Leadership Computing Facility, Argonne National Laboratory, and Lawrence Livermore National Laboratory. Optimization involves compilers and toolchains from GNU Project, Intel Parallel Studio, and accelerators programmed via CUDA and OpenCL, with profiling using TAU (software), Vampir (software), and Score-P.
The initiative began from community planning exercises involving stakeholders such as Royal Astronomical Society, Institute of Physics, Particle Physics and Astronomy Research Council, and universities including University of Cambridge and University of Edinburgh. Early procurement cycles engaged vendors like Dell Technologies and Cray Inc. and aligned with national compute roadmaps influenced by European Strategy Forum on Research Infrastructures. Subsequent upgrades paralleled developments at PRACE, HECToR, and national computing transitions witnessed by XSEDE and NVIDIA GPU generations, while scientific milestones intersected with discoveries reported by groups at Max Planck Society and collaborations such as LIGO Scientific Collaboration.
Access is granted through peer-reviewed allocation processes coordinated with consortia and panels drawing on expertise from UK Research and Innovation, STFC, and university computing directors. Users receive support via documentation, training workshops run with partners like Software Carpentry, Hartree Centre, and EPCC, and helpdesk operations modeled after services at PRACE and National Energy Research Scientific Computing Center. Software provisioning leverages modules and container ecosystems like Singularity and best practices from Open Science Grid, while collaboration tools interface with platforms used by GitHub, GitLab, and Jupyter (software).
Category:Supercomputer sites in the United Kingdom