Research‎ > ‎

AClab GPU resources

Below is a list of ``small'' GPU servers for research and testing purposes of the laboratory, with details included:

Server0:
  • Codename ``Codebreaker'' and acquired in March 2012, in time for the celebrations of the centenary of Alan Turing.
  • CPU: Intel i7-2600 CPU @ 3.40GHz, 2 * 4 cores (8 virtual cores due to Hyperthreading), 8MBytes cache
  • Main memory: 8GBytes
  • Programming: gcc 4.8.4, g++ 4.8.4, python 2.7.6, R 3.0.2
  • GPU:
    • 2 * GeForce GTX 58016 Multiprocessors, 32 CUDA Cores/MP (512 CUDA Cores), 1.57 GHz (GPU max clock rate), 2010 Mhz (memory clock rate), 1536 MBytes memory, CUDA version 7.0
  • Operating system: Ubuntu 14.04 LTS (64 bits)
  • Provided by the equipment grant of the MSc in computer science grant of F. Cabarle from the DOST-ERDT project.
Server1:
  • Codename ``Enigma'' and acquired in January 2016.
  • CPU: Intel Core i7-4790CPU @ 3.60GHz (max: 4GHz), 2 * 4 cores (8 virtual cores due to Hyperthreading), 8MBytes cache
  • Main memory: 16GBytes
  • Programming: gcc 4.9.2, g++ 4.9.2, python 2.7.9, python 3.4
  • GPU:
    • 2 * GeForce GTX 750 @ 16 Multiprocessors, 512 CUDA Cores, 1084 MHz (GPU max clock rate), 2MByte cache, 2505 MHz (max memory clock rate), 2047 MBytes memory, CUDA version 7.5
  • Operating system: Ubuntu 15.04 64bits
  • Provided by the equipment grant of the PhD in computer science grant of J. Clemente from the DOST-ERDT project.

Note that at the moment, server0 is more suitable for ``production'' i.e. stable GPU computing, while server1 is more suitable for ``development'' or testing. This page will be regularly updated. The servers are utilized mostly by members of the AClab, and in conjunction with HPC clusters of our collaborators.
Comments