Lecture 9

Learning objectives

After this class, you should be able to:

1. Given a simple problem, develop a parallel algorithm for it and implement parallel MPI code.
2. Explain how MPI processes typically perform I/O.

Reading assignment

1. Lecture 9 slides.
2. Lecture 9 example code.
3. Section 3.2-3.3.
4. Page 103.

Exercises and review questions

• Questions on current lecture's material
  1. (Post on the discussion board) Run the the lecture 9 parallel pi example. Give the number of intervals, the number of processes used, and the estimate of pi output.
  2. Write parallel MPI code for estimating pi using the trapezoidal method. The number of intervals should be a command line argument. Use a block decomposition, as used in the book.

• Questions on next lecture's material
  1. Which MPI call can we use to replace the sequential addition of the sums received from each process by process 0 in the MPI code for calculating pi?