Skip to ContentSkip to Navigation
Society/business Center for Information Technology Research and Innovation Support Services High Performance Computing Facilities

Peregrine HPC cluster

Peregrine falcon (photo: Keven Law, LA, USA, CC BY-SA 2.0 license)
Peregrine falcon (photo: Keven Law, LA, USA, CC BY-SA 2.0 license)

Peregrine is the university compute cluster. The cluster has been delivered by Dell.

The Peregrine falcon is the fastest animal on the planet. It can reach a speed of more than 320 km/h when stooping down on its prey.

The cluster can be used by scientists from the university. It is useful for solving computational problems for which a single PC is not powerful enough. For these problems the 5740 cpu-cores and 220000 CUDA-cores Peregrine cluster can be used.

Computers of Peregrine during the installation
Computers of Peregrine during the installation


Compute nodes

The cluster nodes come in three variants. There are 210 'standard' nodes with

  • 24 or 28 Intel Xeon 2.5 GHz cores or 64 AMD EPYC 7601 cores
  • 128 GB internal memory or 512 GB internal memory

Furthermore 42 nodes with special accelerator cards are available.

  • 6 with 2 Nvidia K40 GPU cards
  • 36 virtual nodes with one Nvidia V100 GPU card

For jobs that need extra memory 7 nodes are available with

  • 48 Intel Xeon 2.6 GHz cores
  • 1024 or 2048 GB internal memory

Attached to the cluster is 463TB of disk space. Each node also has internal disk space of about 1 TB. Furthermore the Datahandling facilities, with 3PB of disk space, are also available to Peregrine.


The nodes are coupled using an 56 Gb/s Infiniband network. This network has both a high bandwidth (large amounts of data per second) and a low latency (minimal amount of time needed for sending small messages). This makes this network especially suited for tasks using more than one machine.

The system is also equipped with a 10 Gb/s Ethernet network, which is useful when accessing external data.

Parallel computing

The machines can be used to submit a large number of jobs. It is also possible to make use of the fast network between the nodes to submit a large parallel job. These jobs can run much faster than when running on a single core. In order to make use of this option programs often need to be adapted, for example by adding adding MPI calls. The CIT can assist you in making your program parallel.


The CIT offers a beginners course for using the cluster. An advanced course is offered twice a year as well. More information can be found on the course page.


More information can be found on the pages with  documentation

Request account

To request an account for the HPC cluster, please fill in the request form. Please enter all required information so that the CIT Servicedesk can respond as soon as possible.

Please note!: As employee of the university, you can apply for an account for one of your students. If you do so, please fill in the student's number, name, email etc., otherwise fill in your own info.

Maintenance and Problems

For a list of planned maintenance and current problems have a look at the status page

Last modified:17 March 2020 4.56 p.m.
printView this page in: Nederlands