Welcome to the High-Performance Computing (HPC) wiki

The High-performance Computing (HPC) cluster is part of workpackage 5 (infrastructure) of the Research ICT program of the UMC Utrecht and aims to provide high-performance computing power to researchers within the UMC Utrecht.

The first part of the wiki introduces the HPC cluster to those of you who may be interested to join. Take your time and have a look around.

The second part of the wiki contains useful practical information for users of the HPC cluster.

(for a high-resolution version of the HPC flyer, click on the thumbnail on the left)

News

Thursday, 17th Oct 2013

Yesterday evening, we performed some maintenance.

Among others, the network connection of the first submit host (hpcs01.op.umcutrecht.nl) was upgraded. It is now the same as its sister (hpcs02): two gigabits/s, both to the storage, and to the rest of the network. The machine has a new IP address as well: 143.121.195.5; your ssh-client may notice this change.

A somewhat older change: on both submit hosts, the memory limits for interactive work have been relaxed: you can now use 10GB ram, plus 2GB swapspace.

Thursday, 30th May 2013

To further facilitate basic interactive usage of the HPC cluster, we installed a second login/submission server (see here).

General information

The HPC cluster currently consists of 46 compute nodes (552 cores, 5.75TB working memory) and 160TB of HPC storage, and will grow further in the near future. A dedicated Linux administrator and part-time bioinformatician are funded by the research ICT program. In addition, the program subsidizes the HPC storage capacity.

Participating groups. Currently, thirteen research groups are actively using the HPC cluster.

HPC user council. To steer future directions of the HPC infrastructure together with the participating research groups, a HPC user council has been setup. We aim to meet every two months to discuss the usage of the HPC cluster as well as new developments.

How to get involved. Research groups can participate by funding the hardware required for their own computational needs. HPC storage capacity can be rented on a per Terabyte, per year basis. For testing purposes and/or trying out the HPC infrastructure, free trial accounts can also be arranged that have (limited) access to the HPC resources, contact us if you are interested.

Contact details

The working team HPC is responsible for setting up and maintaining the HPC infrastructure, as well as for helping out with HPC related user questions. For details and contact information, go here.