Cluster Downtime
The HPC cluster will be down over the weekend of 22-23 March for scheduled network maintenance. We will be placing the cluster in draining mode on Friday 21st at 17:00 at which time no further jobs may be submitted. The cluster will be shut down on Saturday 22nd at 17:00 at which point all running jobs will be cancelled.
We hope to have the cluster back up by 09:00 Monday 24th.
Acknowledgements
If you have acknowledged use of the cluster in any publications please let us know.
Accounts
If you no longer require your UCTHPC account please let us know. Accounts that have never been logged into will be de-provisioned after a year.
Jobs
A reminder, please do not run high load processes on the head node. Software compiles, zipping or unzipping large quantities of data etc should be run on a worker node. This can be done via an interactive job. Even short tests of software should not be done on the head node.
Core reservations
Please when submitting jobs only reserve an appropriate number of cores. Short, interactive jobs seldom require 40 cores.