Inside the Message Passing Interface

Chapter 5 Optimization

In this chapter you will learn how to optimize MPI internally using advanced implementation techniques, available special hardware. We discussed optimization targets briefly in Chapter 1 (see Section 1.4.7) and occasionally practiced the art of optimization across the previous three chapters. Here we will drive everything together and review the entire field, noting those things that we have done before simply in passing, and focusing chiefly on performance as the optimization target.

Due to the sheer volume of this topic, we will start programming less and looking more into what other people have done, still providing an example or two of the technologies discussed in application to the MPI subset developed in Chapter ...