Lecture 14

Learning objectives

After this class, you should be able to:

1. Time code using the rdtsc counter.
2. Use performance tools, such as gprof, Tau, and Vtune to profile code and identify its performance bottlenecks.
3. Dissasemble object code and read the assembly code to identify causes of performance problems.

Reading assignment

1. Lecture 14 slides.
2. Lecture 14 example codes.

Exercises and review questions

• Questions on current lecture's material
  1. Run the lecture 14 example codes on gpu and read the codes.
  2. Check if the assembly timer in the lecture 14 example gives correct results. If not, fix the error.
  3. Profile the ex1p code using gprof.
  4. (Post on the discussion board) Identify performance bottlenecks to any large code that you have written in the past, using any profiling tool. Summerize your conclusions on the discussion forum.

• Questions on next lecture's material
  1. None.