Parallel Programming with Intel® Parallel Studio XE

Chapter 4

Producing Optimized Code

What's in This Chapter?

A seven-step optimization process

Using different compiler options to optimize your code

Using auto-vectorization to tune your application to different CPUs

This chapter discusses how to use the Intel C/C++ compiler to produce optimized code. You start by building an application using the /O2 compiler option (optimized for speed) and then add additional compiler flags, resulting in a speedup of more than 300 percent.

The different compiler options you use are the course-grained general options, followed by auto-vectorization, interprocedural optimization (IPO), and profile-guided optimization (PGO). The chapter concludes with a brief look at how you can use the guided auto-parallelization (GAP) feature to get additional advice on tuning auto-vectorization.

The steps in this chapter will help you to maximize the performance you get from the Intel compiler.

Most of the text of this chapter uses the Windows version of the compiler options. You can use the option-mapping tool to find the equivalent Linux option. The following example is used to find the Linux equivalent of /Oy-:

map_opts -tl -lc -opts /Oy- Intel(R) Compiler option mapping tool mapping Windows options to Linux for C++ ‘-Oy-’ Windows option maps to --> ‘-fomit-frame-pointer-’ option on Linux --> ‘-fno-omit-frame-pointer’ option on Linux --> ‘-fp’ option on Linux ...