High Performance Computing -

. .

Résumé :
Newer computer architectures rely on multi-core, multi-chip designs to achieve the highest performance; programmers, therefore, need to utilize multi-threading and parallel programming techniques in their applications to achieve high performance on these new designs. This book provides application developers with a detailed understanding of how to effectively program for these new high performance architectures. The authors give a broad overview of the current state of hardware and software advances to support high performance applications. They focus on the more common and successful strategies for multi-threading and parallel programming using examples from actual codes.

Sommaire:
Introduction. Multi-Core Architectures for the Applications Programmer. Compiling for Multi-Core Architectures. Programming for Cache-Based Architectures. Programming for DDE Instructions. Programming for Distributed Memory Clusters. Programming for Multi-Core Distributed Memory Clusters. Using OpenMP and Pthreads across the Cores within the Node. What the Programmer Needs to Do? Message Passing Issues. Performance Analysis. Application Analysis.