Strategies for Choosing Analytics and Visualization Software and Hardware at NERSC

Data analysis (or analytics) and visualization are two steps in data understanding, often interleaved and symbiotic, so that many of the available tools characterized as one category, end-up having some functionalities of the other. Bellow find links to software tools grouped under Analytics or Visualization, but have in mind that their functions may be interchangeable.

Choosing the Right System

• Small datasets:
  
  • interactive: Euclid
  • batch: Carver's 1 TB memory node
• Large datasets:
  
  • for parallel visualization: use Franklin or Hopper
  • for GPU: use Dirac

Choosing the Right Software Tools

• Visualization of large datasets: VisIT
• Data Analysis using statistical computing and graphics: R