ABSTRACT
In this chapter, we look at the SPEComp® benchmark suite [14]. ese tests give us a wide variety of examples, some run very well and others do not run so well. In each of the cases, we will examine the code and determine the reasons for the obtained performance. e hope is that these codes would give the reader an idea of the issues involved when using OpenMPTM on the node. e eective use of OpenMP on the node is a precursor to obtaining good Hybrid code, which was discussed in Chapter 7.
