HPC-UGent Tier-2 Clusters#

The Stevin computing infrastructure consists of several Tier2 clusters which are hosted in the S10 datacenter of Ghent University. This infrastructure is co-financed by FWO and Department of Economy, Science and Innovation (EWI).

Compute clusters#

For most recent information about the available resources and cluster status, please consult https://www.ugent.be/hpc/en/infrastructure .

CPU clusters#

The HPC-UGent Tier-2 infrastructure currently included several standard CPU-only clusters, of different generations (listed from old to new).

For basic information on using these clusters, see our documentation.

cluster name	# nodes	Processor architecture	Usable memory/node	Local diskspace/node	Interconnect	Operating system
doduo (default)	128	2 x 48-core AMD EPYC 7552 (Rome @ 2.2 GHz)	250 GiB	180 GB SSD	HDR-100 InfiniBand	RHEL 9
gallade	16	2 x 64-core AMD EPYC 7773X (Milan-X @ 2.2 GHz)	940 GiB	1.5 TB NVME	HDR-100 InfiniBand	RHEL 9
shinx	48	2 x 96-core AMD EPYC 9654 (Genoa @ 2.4 GHz)	370 GiB	500GB NVME	NDR-200 InfiniBand	RHEL 9

Interactive debug cluster#

A special-purpose interactive debug cluster is available, where you should always be able to get a job running quickly, without waiting in the queue.

Intended usage is mainly for interactive work, either via an interactive job or using the HPC-UGent web portal.

This cluster is heavily over-provisioned, so jobs may run slower if the cluster is used more heavily.

Strict limits are in place per user:

max. 5 jobs in queue
max. 3 jobs running
max. of 8 cores and 27GB of memory in total for running jobs

For more information, see our documentation.

cluster name	# nodes	Processor architecture	Usable memory/node	Local diskspace/node	Interconnect	Operating system
donphan	16	2 x 18-core Intel Xeon Gold 6240 (Cascade Lake @ 2.6 GHz) + 1x shared NVIDIA Ampere A2 GPU (16GB GPU memory)	738 GiB	1.6 TB NVME	HDR-100 InfiniBand	RHEL 9

GPU clusters#

GPU clusters are available in the HPC-UGent Tier-2 infrastructure, with different generations of NVIDIA GPUs.

These are well suited for specific workloads, with software that can leverage the GPU resources (like TensorFlow, PyTorch, GROMACS, AlphaFold, etc.).

cluster name	# nodes	Processor architecture	Usable memory/node	Local diskspace/node	Interconnect	Operating system
joltik	10	2 x 16-core Intel Xeon Gold 6242 (Cascade Lake @ 2.8 GHz) + 4x NVIDIA Volta V100 GPUs (32GB GPU memory)	256 GiB	800 GB SSD	double EDR Infiniband	RHEL 9
accelgor	9	2 x 24-core AMD EPYC 7413 (Milan @ 2.2 GHz) + 4x NVIDIA Ampere A100 GPUs (80GB GPU memory)	500 GiB	180GB SSD	HDR-100 InfiniBand	RHEL 9
litleo	8	1x 48 core AMD EPYC 9454P (Genoa @ 2.75 GHz) + 2x NVIDIA H100 NVL (96GB GPU memory)	315 GiB	1.4TB SSD	NDR-200 Infiniband	RHEL 9

Shared storage#

Filesystem name	Intended usage	Total storage space	Personal storage space	VO storage space (+)
$VSC_HOME	Home directory, entry point to the system	90 TB	3GB (fixed)	(none)
$VSC_DATA	Long-term storage of large data files	1.9 PB	25GB (fixed)	250GB
$VSC_SCRATCH	Temporary fast storage of ‘live’ data for calculations	1.7 PB	25GB (fixed)	250GB
$VSC_SCRATCH_ARCANINE	Temporary very fast storage of ‘live’ data for calculations (recommended for very I/O-intensive jobs)	70 TB NVME	(none)	upon request

(+) Storage space for a group of users (Virtual Organisation or VO for short) can be increased significantly on request.

For more information, see our HPC-UGent tutorial: https://www.ugent.be/hpc/en/support/documentation.htm .

User documentation#

Please consult https://www.ugent.be/hpc/en/support/documentation.htm .

In case of questions or problems, don’t hesitate to contact the HPC-UGent support team via hpc@ugent.be, see also https://www.ugent.be/hpc/en/support .