Parallel Computing: Lecture 13

Learning objectives

After this class, you should be able to:

1. Given speedup data for a code, compute the Karp-Flatt metric and determine the cause of the bottleneck to parallelization.
2. Given a parallel algorithm, determine its iso-efficiency function and its scalability function.

Reading assignment

1. Chapter 7, sections 7.5 - 7.7.
2. Chapter 8, page 179.

Exercises and review questions

• Questions on current lecture's material
  1. Compute the Karp-Flatt metric for columns 1 and 2 of exercise 7.10.
  2. Exercise 7.13, a, b, c.

• Questions on next lecture's material
  1. (Post your solution on the discussion board) Give an example of a 3x3 matrix and a vector, and show the result of the matrix-vector product, using the algorithm in fig 8.1.