Increasingly competitive market conditions in the technology industry mean that bringing products to market faster and with more complex systems is a key factor for success. Creating optimal software plays a key role in this success; for instance, reducing the software size by a mere five percent can be enough to allow another feature to be added to the product, therefore making it significantly more competitive. To determine where in the software to begin optimizations, a developer requires feedback on code performance on the target.

This paper explains how ARM’s RealView Profiler graphically highlights the bottlenecks in software based on real use cases reflective of how an end-user would run the application. The Profiler also shows how these bottlenecks relate back to source code so that appropriate changes can be made to either compiler optimizations or to the source code.