Week 8: Overview

This week we’ll be focusing on speeding up our code by taking advantage of parallelism, i.e. splitting up a problem and giving different pieces to different processors. In these notes we’ll introduce the following concepts:

• What is parallel programming?
  • Why do we need parallel programming?
  • What are shared- and distributed-memory parallelism?
  • What are weak and strong scaling?
• An introduction to OpenMP
  • Parallelising loops
  • Reductions
  • Data sharing
  • Thread control
  • Schedules
  • Parallelising nested loops
• OpenMP example: Fractals
• Cache Performance in Shared Memory
• Summary