Logo: Leibniz Universität Hannover

Tesla

Tesla is a computer containing 4 Nvidia Tesla cards for GPGPU computing.  The Tesla-cards enable acceleration of a program by offloading compute-intensive parts of a program onto the GPU.  Specially written C and Fortran programs which have been compiled with either the Nvidia-C-Compiler or the PGI Fortran Compiler respectively can make use of the very large amount of computing power available on the GPU.

Tesla runs under control of the Torque/Maui batch system.  Jobs for Tesla can be sent from Orac, or Avon the login computers for the RRZN computing service.

Hardware and operating system

Number of nodes: 1
Processors per node: 8
Processor: Intel Xeon CPU E5540, 2.53GHz
Main memory per node: 24 GB
Local disk space: 388 GB
Number of Nvidia-Tesla cards: 4
Main memory per card: 4 GB
Number of multiprocessors per card: 30
Number of GPU-Cores per card: 240
File systems: Lustre; see also general information about file systemsFile systems
Operating system: Scientific Linux

Available software

Application software is available via the Module system.

Regionales Rechenzentrum für Niedersachsen - URL: www.rrzn.uni-hannover.de/tesla.html?&L=1
 
Dr. Gerd Brand, Last Change: 26.11.2010
Copyright Gottfried Wilhelm Leibniz Universität Hannover