Introduction to Parallel Scientific Computing

Parallel Scientific Computing


Two pillars of science

Theory and Experiments

This is not enough as, some experiments are1:

  • economically infeasible
  • ethically inappropriate
  • ecologically dubious
  • by construction impossible
  • so complex that we can’t solve them analytically - fluid dynamics

For these reasons simulation is proposed at the third pillar of science. Some disagree with this statement, but regardless, computer simulations are very important in modern science, be they as a pillar of their own or as an essential part of theory and experiment.

Computational Science and Engineering

More generally called Computational X to work for all disciplines

Computational X covers the broad range of challenges

  1. Model something with mathematical equations
  2. Translate the equations into something you can solve on the computer
  3. Design the algorithms to solve the equations
  4. Code the algorithms
  5. Make the algorithm fast
  6. Input and output data management
  7. Postprocessing
Scientific Computing

Developing the tools to allow others to make progress


  1. David E. Keyes:Computational Science. In N.J. Higham: The Princeton Companion to Applied Mathematics. pp. 335–350 (2015)