- Shopping Bag ( 0 items )
Posted August 1, 2013
Are you a programmer? If you are, then this book is for you! Aut
Are you a programmer? If you are, then this book is for you! Authors James Jeffers and James Reinders, have done an outstanding job of writing a book that gives you the keys to general parallel programming.Was this review helpful? Yes NoThank you for your feedback. Report this reviewThank you, this review has been flagged.
Authors Jeffers and Reinders, begin by Bringing together the essentials to high performance programming for an Intel Xeon Phi coprocessor. Next, they focus on accessing two of the three performance features of the many processing cores with floating point vector processing units and high-speed memory. Then, the authors examine the highly parallel programming with Intel Xeon Phi coprocessors, by using a more substantial yet still relatively simple code example that includes both computation and more complex memory access. They continue by giving an analogy to a real scientific code example that comes from the Riken Advanced Institute for computational Science in Japan. Next, the authors examine the rich set of options available for processors and Intel Xeon Phi coprocessors, with little specific Intel Xeon Phi knowledge required. Then, they show you why you need lots of task level parallelism; and, why you should first consider using OpenMP, Fortran 2008 DO CONCURRENT, Intel Threading Building Blocks, and Intel Cilk Plus. The authors continue by discussing offload models, which have a number of options that blur the distinction a little, specifically _Cilk_Shared; but, still manage distinct task/thread pools on each coprocessor and the host. Next, they examine key elements behind the design of the new Intel computing product line. Then, the authors look at software architecture and components that enable the coprocessor to operate seamlessly in a standard environment. They continue to focus more on knowledge useful to the administration of a system that contains Intel Xeon Phi coprocessors. Next, the authors show you how to use Intel MKL for the Intel Xeon Phi coprocessor. Then, they show you how the MPI programming model is a natural fit for clusters containing Intel Xeon Phi coprocessors. Finally, the authors include information on accurate timing, as it is related to the general topic of gaining insight into the behavior of applications.
In this most excellent book, the authors worked to demystify the Intel Xeon Phi coprocessor and provide the foundational information that would fit in one book. The authors also worked to explain common algorithms or patterns; and, how to implement them to utilize parallel programming.