Greetings all HPC Users, We have made some minor adjustments to the HPC default job settings. All jobs by default will now set –hint=nomultithread when submitted, including all sbatch, salloc and srun jobs. Since more than 95% of the jobs run faster without multithreading enabled, we think it is better to Read more…
Greetings all HPC Users, We will be removing direct access to the HPC Login Node soon alongside the implementation of VPN. Direct SSH into the HPC login node will no longer be possible once VPN is fully implemented and enforced. Here, we prepared some information that all HPC users would need Read more…
Greeting all HPC Users, In conjunction with the enforcement of DICC HPC Service Guideline, we will be removing Amber14 and Amber18 from the managed application list in the HPC cluster. These mentioned Amber versions will no longer be available for use starting 12am, 08 of June, 2022. Amber20 will remain unchanged Read more…
Greeting all HPC Users, We would like to announce that the first version of DICC HPC Service Guideline is now available and accessible from the DICC website, or through this link. The purpose of the HPC Service Guideline is to describe the scope of services provided by DICC, to the UM research communities. The Read more…
Ever since we set up the HPC cluster in 2015, our cluster has been using the Environment Modules system in managing all the applications and modules in the system. The platform was running fine since the first day of implementation with some occasional hiccups due to misconfiguration, but it falls Read more…
Dear all HPC Users, There will be scheduled maintenance on 11 May 2022, 7.00am to perform system update and critical fixes. Access to the HPC cluster will be temporarily unavailable during the maintenance, and the maintenance is expected to complete within an hour. All running jobs should not be affected and will Read more…
Dear all HPC Users, We are currently experiencing problems with home storage. The cluster has to be immediately brought down temporarily to perform necessary fixes. We will be updating on the progress once it is up and running again. We are sorry for the inconvenience caused. Update: The HPC cluster is now back Read more…
Dear HPC Users, We have recently redesigned some of the script modules located in the HPC login node (umhpc). These modules now have an color-coded interface that is much easier to read. The following are some screenshots of all the redesigned modules: Message of the Day (Login Banner) Lustre Cleanup Read more…
Dear all HPC Users, We have recently identified a bug in the OpenOnDemand HPC Dashboard module (XDMod) which prevents users with username shorter than 5 characters to be unable to access their job account details in XDMod. If you are one of those affected users previously, you should now be able Read more…
Dear all HPC Users, In order to allow users to better plan their job submissions in the HPC cluster, we have decided to allow users to view all the jobs in the queue. All users should now see jobs other than your own jobs when you input squeue command. If you have Read more…