Research Infrastructure

RI Update, Jan 2015

The Research Infrastructure Team

If you have any questions about the items below, please contact the RI team at

CSF Refresh

The Computational Shared Facility (CSF) is the University flagship computational resource. It started life four years ago and is now running 1000s of jobs for over 650 users from 35 research groups representing all faculties.

The CSF is getting a New Year refresh. This will secure the service for years to come. The refresh was part-funded by a successful bid to The University (ISSC) and therefore represents a further commitment to Research IT infrastructure at Manchester.

All fileservers, login nodes and some other pieces of infrastructure are being replaced:

  • scratch space will be bigger and faster;
  • login nodes will have much more available memory.
  • We will also have Intel Haswell-based compute nodes (around 500 cores) in production soon.

All the above will be connected via Infiniband (very fast network) — it is envisaged that a much higher proportion of the CSF will be Infiniband-connected in the future.

K20 and K40 Nvidia GPUs

Zrek (aka the zCSF) now contains 12 Nvidia K20 GPUs and also two K40 GPUs. (And some Xeon Phi cards and FPGAs as well.)

Priority use of these cards is for those who contributed (paid for!) them. However, all researchers at The University (and their collaborators) may use the cards for an evaluation or pump-priming period, and when contributors are not using them. If you would like to make use of these GPUs, please contact us at the above email address or visit

Dedicated Research IT VM Infrastructure

A recent successful bid was made to The University (ISSC) for virtual machine (VM) infrastructure dedicated to research use. These funds will be used to seed a contribution-based VM service running along the same lines as the CSF: any researcher contributing funds will receive VMs equivalent to the hardware so procured. These VMs can be used for group development areas, Web servers and databases and for moderate, standalone computational work.

We also plan:

  • to use this infrastructure to provide job submission gateways to the CSF and Hydra (e.g., using Galaxy);
  • and explore the possibility of offering a central visualization capability.

More details about this new infrastructure are contained in a dedicated article.

Big Data: Apache Spark

We are running an experimental Apache Spark service on Hydra

If you would like to try it out, send us an email!

Comments are closed.