The University has a wide range of HPC and computational resources available to the research community. Detailed information about each of the resources listed below can be found on the Research IT website along with details of how to access the service, where to obtain further information and if the resource is free to use or if there is a cost involved.
The Computational Shared Facility (CSF)
The CSF comprises CSF3 and CSF4. They provide a batch-based and limited interactive on-campus platform with a large range of software for all research domains to enable computationally-intensive work.
The Interactive Computation Shared Facility (iCSF)
The iCSF is a service designed specifically for interactive computationally-intensive work. For example using applications such as Matlab, R, and Stata through their graphical user interface (GUI) to process / analyse large datasets. It has a large range of software for all research domains to enable computationally-intensive work.
HPC Pool
The HPC Pool provides a batch-based on-campus platform with a large range of software for all research domains to enable computationally-intensive work. It is a resource for true High Performance Computing jobs, i.e. those that scale and can use several compute nodes, typically via MPI.
HTCondor Pool
The HTCondor resource is designed for high throughput computing (HTC) – where you run the same application to process tens, hundreds or even thousands of different data files, or you run a simulation tens, hundreds of thousands of times with different input parameters, perhaps to find the “best” set of parameters. HTCondor can perform many of these runs at the same, giving you your results much sooner than if you had performed each run one after another on your PC/laptop.
N8 CIR (Bede)
N8 CIR provide a regional GPU machine (Bede) based at Durham University which is shared between the N8 universities including the University of Manchester. It comprises of 32 IBM Power 9 dual-CPU nodes (each with 4 NVIDIA V100 GPUs), and 3 NVIDIA Grace Hopper Superchip nodes (each with an NVIDIA Grace CPU and H100 GPU), and high performance interconnect. It is free to use for UoM researchers and dedicated support is available within Research IT to get you started.