Attention

The ShARC HPC cluster was decommissioned on the 30th of November 2023 at 17:00. It is no longer possible for users to access that cluster.

ShARC decommissioning

ShARC will be decommissioned on the 30th of November 2023, after which time users will no longer be able to access that cluster and any jobs running or queueing at that time will be cancelled.

Once ShARC has been decommissioned and all data securely deleted the system is to be donated to the Igor Sikorsky Kyiv Polytechnic Institute in Kyiv, Ukraine.

We recommend users move to the Stanage cluster which is a new, much larger system than previous TUOS clusters with drastically larger numbers of CPU cores, numbers of cores per node, memory per node (especially with Stanage’s 1TB and 2TB nodes) and a large number of GPUs.

Please get in touch with us if you have any concerns regarding this announcement via HPC drop-in sessions (every Friday) or research-it@sheffield.ac.uk.

Hint

We are running some virtual sessions to assist researchers in migrating workflows to and familiarising themselves with Stanage. If you would like to attend a Stanage HPC onboarding session, you can register for our next one via Google Forms.

Key things you need to know

Data must be migrated

Any user data stored in /home, /data or /fastdata directories on ShARC that you want to retain will need to be migrated elsewhere e.g.

to equivalent areas on Bessemer or Stanage or
to /shared/GROUPNAME areas, which can be made accessible from Bessemer or from Stanage login nodes if they aren’t already.

For guidance on moving your data please see our page on transferring files.

ShARC specific shared filestorage areas will no longer be accessible

Shared application installations within /usr/local on ShARC (including /usr/local/packages and /usr/local/community) will not be accessible after the decommissioning date.

Equivalent shared application installation areas will be made available on the Stanage cluster prior to the ShARC decommissioning.

Stanage is our recommended target for migration

Stanage is our newest HPC cluster and is the logical successor to ShARC; it is well suited to running larger parallel jobs spanning multiple nodes and benefits from a high-performance network interconnect between nodes.

As Bessemer only permits single node execution and is several years older than Stanage, we recommend that ShARC users preferentially migrate workloads to Stanage instead.

User installed software may require recompiling

If you have compiled and installed your own applications on ShARC you should recompile your applications on the newer clusters when migrating your workloads in order to ensure compatibility and potentially significant performance increases.

If you currently use IT Services-managed applications on ShARC which are not currently available on Stanage or Bessemer then you can submit a request to research-it@sheffield.ac.uk.