Druckansicht der Internetadresse:

Faculty of Mathematics, Physics & Computer Science

Scientific Computing, Master of Science (M.Sc.)

Seite drucken

Supercomputers

Scientific Computing students may use all of the University of Bayreuth's computer clusters for their studies and research. Additionally they have exclusive access to the new ENB-Cluster.

A Quick Start Guide to supercomputing at the University of Bayreuth can be found here. A detailed introduction will be provided during "Practical Course on Parallel Numerical Methods" which is one of the key components of Scientific Compinting master's programme.

ENB-Cluster

The ENB-Cluster went fully operational in January 2021. It consists of 13 via InfiniBand network connected compute nodes each featuring two AMD Epyc Processors (2nd generation) with 24 cores each. One node also contains a NVIDIA Tesla V100 data center GPU. This GPU-node is equipped with 512 GB of RAM, the other 12 CPU-nodes with 256 GB. Due to these specifications the ENB-cluster is well suited for AI and other memory intensive HPC workloads like simulations and data analysis.

Login nodes

  • btrzx1-1.rz.uni-bayreuth.de
  • btrzx1-2.rz.uni-bayreuth.de

Compute nodes (state January 2021)                                                                                                   

  • 12 CPU- nodes:   
    - 2x AMD Epyc 7352 @ 2.30GHz, 24 cores each, in total 48 pyhsical cores 
    - 16x 16GB of RAM, in total 256 GB 
    - local disc space of 256 GB
       
  • 1 GPU-node:     
    - 2x AMD Epyc 7352 @ 2.30GHz, 24 cores each, in total 48 pyhsical cores
    - 16x 32GB of RAM, in total 512 GB
    - local disc space of 256 GB 

    - NVIDIA Tesla V100 16 GB data center GPU

Queues

WAP
Default job priority
Wall time limit: 24 hours
Restrictions: no

Network

  • InfiniBand (56 Gbit/s)

User file space (network and local)

  • Panasas file system (/scratch & /home)
  • Local disk (/tmp, 500 GByte)

Commissioning & Extension

  • January 2021

Resource Manager & Scheduler

Operating system


Verantwortlich für die Redaktion: Thomas Rau

Facebook Twitter Youtube-Kanal Instagram Blog UBT-A Contact