User Tools

Site Tools


scbc9

Future of Scientific Computing

What will the future hold? Generally in plasma physics the driving force for larger and larger computations is the need to perform multiphysics simulations, coupling between many different models from the atomic scale through to full system descriptions. The methodology is evolving too - moving towards more of a validation activity than algorithm and physics description phase: the future entails more validation analysis than previously. Finally, the technology is evolving perhaps more rapidly than we can all keep up with: standard architectures are giving way to massively parallelized systems like GPUs. This session with entail a summary of some of the more forward-looking aspects of scientific computing.

Invited Speaker:

Dr. David Keyes Algorithmic Adaptations to Extreme Scale

Bulk synchronous parallelisms

Floating-point Operations Per Second (FLOPS)

Online resources

Argonne Training Program on Extreme-Scale Computing

Exascale Computing

America’s Next Generation Supercomputer: The Exascale Challenge

  • pdf page 7 - Background
  • pdf page 8 - Advanced Scientific Computing Research Program
  • pdf page 10 - Exascale Challenges

The Opportunities and Challenges of Exascale Computing

  • pdf page 3 - Finding: Making the transition to exascale poses numerous unavoidable scientific and technological challenges
  • pdf page 9 - 2.2.7 Fusion Energy
  • pdf page 21 - 3.5 Fusion Energy
  • pdf page 42 - 4.7 Fusion Energy
  • pdf page 49 - 5.2 The Hardware Challenges
  • pdf page 53 - 5.3 - 5.5, maybe 5.6 - Challenges → none of this is particularly helpful if we don't have more-clever algorithms etc. to take advantage of exascale computing

Additional Interest

scbc9.txt · Last modified: 2022/07/21 06:59 by 127.0.0.1