.. _UGentT2 hardware:

#########################
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).

Login nodes
===========

Log in to the HPC-UGent Tier-2 infrastructure via https://login.hpc.ugent.be or using SSH via ``login.hpc.ugent.be``.

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 <https://docs.hpc.ugent.be/Linux/running_batch_jobs/>`__.

=============== ========== =========================================================== ===================== =========================== ======================= ===================
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 <https://docs.hpc.ugent.be/Linux/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 <https://docs.hpc.ugent.be/Linux/interactive_debug/>`__.

=============== ========== =========================================================== ===================== =========================== ======================= ===================
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)   738 GiB               1.6 TB NVME                 HDR-100 InfiniBand      RHEL 9
                           + 1x shared NVIDIA Ampere A2 GPU (16GB GPU memory)
=============== ========== =========================================================== ===================== =========================== ======================= ===================

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)   256 GiB               800 GB SSD                  double EDR Infiniband   RHEL 9
                           + 4x NVIDIA Volta V100 GPUs (32GB GPU memory)
accelgor        9          2 x 24-core AMD EPYC 7413 (Milan @ 2.2 GHz)                 500 GiB               180GB SSD                   HDR-100 InfiniBand      RHEL 9
                           + 4x NVIDIA Ampere A100 GPUs (80GB GPU memory)
litleo          8          1x 48 core AMD EPYC 9454P (Genoa @ 2.75 GHz)                315 GiB 	             1.4TB SSD 	                 NDR-200 Infiniband 	 RHEL 9
                           + 2x NVIDIA H100 NVL (96GB GPU memory)
=============== ========== =========================================================== ===================== =========================== ======================= ===================

.. _UGent storage:

Shared storage
==============

.. include:: storage_quota_table.rst

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 .