2023-12-20 — Introductory courses Spring 2024

At the beginning of the Spring semester, we offer our series of introductory courses to supercomputing and the use of the VSC clusters. More information on our training events webpage.

  • Linux Introduction: 12, 13 February 2024, 9.30-13.30 h
  • Supercomputers for starters: 20, 22 February 2024, 9.30-13.00 h
  • HPC@UAntwerp Introduction: 26, 27 February 2024, 13.00-17.30 h

Do not forget to take a look at VSC’s training page as well: you will find an overview of many other interesting trainings organized by VSC, LUMI and EuroCC.

2023-12-20 — Cut-offs for project proposals for Tier-1, LUMI and EuroHPC

Web pages, forms and regulations will be updated in the coming weeks, but this way you are already informed about it.

2023-11-10 — Downtime from Thursday November 16 to Monday November 20

There will be a downtime for all CalcUA clusters between Thursday November 16 (8 am) and Monday November 20 (6 pm). This is needed for planned work on the high-voltage network on CGB on Saturday. Therefore, we need to power down all our equipment, clusters as well as storage.

During the downtime, access to other VSC Tier-2 sites will not be possible. Access to VSC Tier-1 is possible but the share folders $VSC_HOME and $VSC_DATA will not be available.

We have decided to combine this downtime with other maintenance tasks (that would otherwise require yet another downtime). Afterwards, we will run some tests before we release the clusters again.

Our apologies for the inconvenience.

Update Monday 20 September

Due to an issue with the upgrade of our scratch filesystem, we unfortunately need to extend our downtime. We will keep you informed about the availability.

Update Monday 27 September

We were (finally) able to fix our scratch filesystem without data loss. The CalcUA clusters are now back online and the shared filesystems are available again on all VSC clusters.

Thank you all for your patience during the extended downtime.

2023-08-23 — Hopper nodes replaced by Breniac nodes

When the Hopper cluster was moved out in the summer of 2020, we kept 24 compute nodes having two 10-core Intel E5-2680v2 Ivy Bridge generation CPUs with 256 GB RAM to form a separate ivybridge partition. This partition was well suited to run jobs that lack enough parallelism to fully exploit the Leibniz or Vaughan compute nodes, including batches of single core jobs, as well as for jobs that do not fit in a maximum wall time of 3 days and cannot be restarted cheaply. The single user policy was not enabled on this partition, thus offering shared (simultaneous multi-user) access to the compute nodes. The Hopper nodes had CentOS 7.9 installed.

In December 2022, the (now previous) Tier-1 machine BrENIAC located at the University of Leuven that was in operation since 2016, was finally decommissioned. We had the opportunity to obtain a bunch of its compute nodes with dual 14-core Xeon Gold 6132 Skylake generation CPUs and 192 GB RAM. These Breniac nodes are now available in the new skylake partition, with shared access and a maximum wall time of 7 days. There is also a dedicated login node, accessible via login-breniac.hpc.uantwerpen.be.

All CalcUA cluster nodes are now running the same version of Rocky Linux 8.

The Hopper nodes and partition are no longer available.

2023-06-05 — New AMD Zen3 nodes added to Vaughan

In June 2023, the Vaughan cluster was extended with an extra rack, consisting of 40 new compute nodes with two 32-core AMD Epyc 7543 Zen3 generation CPUs. These new compute nodes are now available in the following partitions:

  • zen3 partition: 24 AMD Zen3 compute nodes with 256 GB RAM
  • zen3_512 partition: 16 AMD Zen3 compute nodes with 512 GB RAM