Optimizing HPC Applications with Intel Cluster Tools

Optimize Parallel Programs w/Intel Cluster Studio XE takes the reader on a tour of the quickly growing area of High Performance Computing: optimization of hybrid programs. These programs combine distributed memory and shared memory programming models and use the Message Passing Interface (MPI), OpenMP, Intel(R) Thread Building Blocks (TBB), Intel(R) Cilk Plus, and other ways to achieve ultimate performance and power consumption goals on enterprise-class workstations and computational clusters.§The book combines an easygoing and informal style with a systematic methodology of parallel program optimization proven by years of experience, in the context of Intel's platforms and software-development tool environments.§Real-life application examples convey in a practical, hands-on manner the best known optimization methods applicable in both Linux and Microsoft(R) Windows environments.§Although primary attention is focused on hybrid applications and heterogeneous installations that mix Intel(R) Xeon and Intel(R) Xeon Phi(TM) processors, the optimization of purely distributed memory, shared memory, and even sequential applications on homogeneous clusters and workstations built on Intel(R) Xeon(R) is also covered. The reader can easily focus on the applicable sections clearly marked by respective '147;trails.'148; For example, there is a trail for MPI programmers, a trail for OpenMP programmers, and so on. They can be combined at will to suit the reader's needs.§On top of the tutorial and reference content, the authors address, unveil, and refute many myths and misconceptions surrounding the topic. The text is augmented and enriched by descriptions of real-life situations, sometimes funny and always instructive in their bearing.§