About Isambard
The GW4 Isambard is a supercomputing collaboration between the GW4 universities (University of Bath, Bristol, Cardiff and Exeter) and the Met Office.
This Tier2 high performance computing service provides access to multiple advanced architectures in order to enable evaluation and comparison across a diverse range of hardware platforms.
Isambard is a Cray XC50 system which comprises 10,496 cores and is one of the world's first production Arm-based supercomputers. Each of the 164 compute nodes contains two 32-core Cavium ThunderX2 processors running at 2.1 GHz and has 256 GB of DDR4-2666 memory. The nodes are connected via Cray Aries interconnect with a Dragonfly topology. A Cray Sonexion 3000 storage cabinet provides 480 terabytes of high-performant Lustre storage.
The supercomputer is named after the renowned Victorian engineer Isambard Kingdom Brunel and is funded by EPSRC.
Applying for an allocation on Isambard
As part of our involvement in the project, Bath has 15% of the compute time on the machine.
Researchers can apply for an account to explore the performance of their codes and to make a technical case for their use of Isambard. In order to manage our allocation we distribute time on Isambard via a lightweight allocation panel which reviews the scientific and technical aspects of proposed projects.
If you are interested in running a project on Isambard, please get in touch with the Research Computing team.
Isambard 3 supercomputer announced for early 2024
GW4 has secured £10 million UKRI funding to create the next-generation NVIDIA Grace™ Isambard 3 supercomputer for AI and cutting-edge scientific innovations.
The new high performance computing (HPC) system utilising breakthrough technologies such as the new Arm®Neoverse™-based NVIDIA Grace CPU Superchip will have at least 55,000 cores to support large scale scientific experiments and will be ready for use in early 2024.
Further information
Further information about GW4 Isambard HPC can be found here: http://gw4.ac.uk/isambard/.
Isambard user documentation can be found here: https://gw4-isambard.github.io/docs/.