Full Description
High Performance Computing (HPC) is the area of supercomputers, the fastest and best systems of their time. Most supercomputers are used to solve scientific problems, be it weather forecasting, simulation of car crashes or wind tunnels, modeling of proteins in bio-chemical research, or for simulation of atom-bomb tests. The book deals with systems that are in practical use for such application areas. In the first part, the history of supercomputers is described, and it is shown which properties and trends will remain in the future. The second part describes all aspects of High Performance Computing, from chip technology over computer and cluster architecture, up to software, algorithms, and applications. The third part explores the practice of HPC planning: system selection, benchmarks, acceptance tests, and implementation of HPC environments. Finally, the book closes with an outlook on current trends and visions like Grid Computing.
Contents
1. History.- History of Computing.- Development of Supercomputers and Clusters.- Seymor Cray, Amdahl and others.- 2. Architecture, Chips, Clusters and Nodes.- HPC Application Stack.- Chip Development Past and Future.- Cluster Systems.- Operating Systems.- 3. Software Aspects.- Introduction into High Performance Programming.-Optimization.- Parallelization Concepts.- Fortran and c as Programming Languages.-OpenMP and MPI for Parallelization.- Amdahl's Law.- 4. Practise (how to configure and purchase).- Benchmarks.- System Configuration.- Purchasing.
