Policies

Who can use Prospero?

Prospero is available for use by all academic staff and postgraduate research students at LJMU. Collaborating researchers at external institutions are also eligible if they have a genuine need to use the facility, with cases reviewed on their individual merits. LJMU staff wishing for collaborators to have access should raise an enquiry with a Helpdesk ticket.

Cost of use

Prospero is free at the point of use to researchers. Users are currently grouped into three accounting groups to reflect the current usage of the facility, these are ARI (members of the Astrophysics Research Institute), FET (for all other members of the Faculty of Engineering and Technology) and LJMU (for members of other faculties). The batch queue scheduler system on Prospero allocates jobs with a priority that is balanced against the prior usage of these groups.

Conditions of use

Prospero is subject the university’s standard IT policy. There are a number of additional conditions that apply specifically to Prospero:

  • Users must not execute jobs directly other than for very brief testing. All production calculations must be submitted via the SLURM batch queue scheduler.

  • Users may log into compute nodes, but only for the purposes of examining the local use of resources (e.g. via top). As above, jobs may not be executed directly on compute nodes.

  • Note that the system is actively monitored and directly-executed jobs will be terminated without warning.

  • Personal data should not be stored on the system.

  • Anonymised survey data should not be stored on the system.

  • Prospero’s security features should not be attempted to be bypassed.

  • Users violating these rules will initially be warned by ITS. Repeat offenders will be banned from the system and their line manager and head of school informed.

Acknowledging Prospero

Please include the standard wording below, as appropriate, in the acknowledgements of all papers that make use of Prospero. This is critical as it enables research outputs enabled by the facility to be tracked, and by extension its productivity to be demonstrated to funders.

This study made use of Prospero high-performance computing facility at Liverpool John Moores University.