CHAPTER 3: Common Algorithmic Patterns

Several of the sample problems we have discussed this far have been embarassingly parallel.

Computations are largely independent of each other.

Introduce the notion of a dependency, and how dependenecies can limit the amount of parallelism inherent in a program.

In this chapter, we cover common algorithmic patterns and their parallel implementations to show readers what strategies to use when parallelizing more complex tasks.

• 3.1 Parallel Selection
  • 3.1.1 The Selection Task
  • 3.1.2 A Basic (Sequential) Solution
  • 3.1.3: A Parallel Selection Algorithm
  • 3.1.4 A Parallel Solution using OpenMP
  • 3.1.5 A Parallel Solution using Message Passing
  • 3.1.6 Performance Consideratons
  • 3.1.7 Summary
• 3.2 Shared Queue
  • Introduction
  • The Python multiprocessing module
  • Creating a shared queue
  • The Greatest Common Divisor (GCD)
  • Computing the GCD with a Shared Queue
  • Translating it to code: Serial Version
  • Translating it to code: Parallel Version

You have attempted of activities on this page