Lecture 25

Learning objectives

After this class, you should be able to:

1. Write simple code that combines MPI and CUDA, and compile and run it on XSEDE machines.

Reading assignment

1. Lec25 example code.
2. UIUC Lecture 27.
3. Chapter 19 of text.

Exercises and review questions

• Exercises and review questions on current lecture's material
  1. Modify the vec-add program so that it uses zero-copy. Compare its performance against that of the code without zero-copy and report it on the discussion board. If you don't report the results of this task, then you will need to combine CUDA and MPI in your project.

• Preparation for the next lecture
  1. What is the OpenCL equivalent of CUDA's threadIdx.x?
  2. What is the OpenCL equivalent of CUDA's blockDim.x?

---

Last modified: 11 Apr 2013