Remote Visualization (visgpu)

The remote visualization systems provide access to visualization software, such as uv-cdat, without requiring running X-Windows through ssh. The systems have powerful graphics cards and allow the user to render the image close to the data and send only the rendered image back to the user's desktop, eliminating the need to download large data files and reducing the time to display the visualizations.

Remote Visualization System Details

System Architecture

The remote visualization hardware consists of four systems each containing:

  • HP DL380 G8
  • 20 E5-2670 v2/2.50GHz cores
  • 128 GB of RAM
  • NVIDIA K5000 GPU card

In addition to the above, the visgpu systems have access to shared storage from the ADAPT and discover (read-only) filesystems.

Operating Environment

  • Operating System: CentOS
  • Job Scheduler: None
  • Compilers: gcc
  • Analysis applications: UV-CDAT, Octave, python, Panoply, GrADS, cdo, nco

Nodes available to users:

  • Visgpu02
  • Visgpu03

Users are assigned to one of the nodes to distribute load and resources.