As clock speeds for CPU’s have not been increasing as compared to a decade ago, chip designers have been enhancing the performance of both CPUs, such as the Intel Xeon and the Intel Xeon Phi coprocessor by adding more cores. New designs allow for applications to perform more work in parallel, reducing the overall time to perform a simulation, for example. However, to get this increase in performance, applications must be designed or re-worked to take advantage of these new designs which can include hundreds to thousands of cores in a single computer system.
Applications that have demonstrated the ability to scale on the main CPUs, such as the Intel Xeon, may scale well beyond what a CPU can offer, and require the high core count of a coprocessor. It is important to give developers the familiar development tools and languages to scale an application beyond the main CPU. If a developer has to learn a new programming environment in order to gain the performance of the coprocessor, then overall productivity can take a hit.
Since most applications have not been designed to run in a highly parallel mode, understanding how to take advantage of this increase in parallelism is as important as how to restructure a piece of code, or the entire application. Understanding the architecture of the system is critical to understanding how to gain maximum performance. While an Intel Xeon core will outperform a single Intel Xeon Phi coprocessor core, the overall performance of the Intel Xeon Phi coprocessor will be higher, due to having many more cores available for computation.
A platform that is ideal for modern HPC applications will consist of two main Intel Xeon CPUs and either one or two Intel Xeon Phi coprocessors. Familiarity is very important for developers who require the performance of the Intel Xeon Phi coprocessor. In fact, once on a system, a programmer can ssh into the coprocessor and run all the normal Linux commands on it. This is the first step in developing new and innovative applications that take advantage of the unique characteristics of the Intel Xeon Phi coprocessor.
Source: Intel, USA
Transform data into opportunity. Speed data analysis in your applications.