Although a lot has been written about the Intel® Xeon Phi™ coprocessor’s performance comparison with GPGPUs, relatively little has been written on the comparison of available programming models and methodologies. However, with the release of Intel® Parallel Composer XE 2015 and HP Apollo 6000 ProLiant XL250a server, it’s now time for developers to see how following a step-by-step methodology in a structured framework can help them develop highly parallel applications that can run effectively in an integrated heterogeneous environment.
A new white paper from HP and Intel explains the five-step application model that is essential to extracting the maximum application performance from multilevel parallelism. The optimization framework turns a sequential legacy application into a highly parallel application to make use of the hardware resources both on the host CPU and on the accelerator devices to enable simultaneous heterogeneous computing. The benefit of such a framework is that it allows the developer to realize and quantify substantial performance improvements while concentrating on their application by implementing a straight-forward optimization method.
The paper also includes a European option pricing application case study that looks at how to apply this framework and adopt a structured methodology to take advantage of a heterogeneous computing environment.