Parallel Scientific Computing in C++ and MPI: A Seamless...

Features

• Cover Type: Paperback with 628 pages
• Published by: Cambridge University Press; Pap/Cdr edition June 16, 2003
• Written in: English
• ISBN 10 Number: 0521520800
• ISBN 13 Number: 978-0521520805
• Book Dimensions: 8.9 x 7.6 x 1.2 inches
• Weighs: 2.5 pounds

Product Review
'This book is a valuable addition to the literature on numerical algorithms.' Numerical Algorithms

'There is plenty of material for a two-semester sequence, or selected chapters could be used for a one-semester course on numerical linear algebra. The presentation is clear, practical, and lively this text would be a very useful reference for statistics students and professionals) seeking to take their statistical computing skills into the parallel realm.' Journal of the American Statistical Association

Book Description
This book provides a seamless approach to numerical algorithms, modern programming techniques and parallel computing. These concepts and tools are usually taught serially across different courses and different textbooks, thus observing the connection between them. The necessity of integrating these subjects usually comes after such courses are concluded (e.g., during a first job or a thesis project), thus forcing the student to synthesize what is perceived to be three independent subfields into one in order to produce a solution. The book includes both basic and advanced topics and places equal emphasis on the discretization of partial differential equations and on solvers. Advanced topics include wavelets, high-order methods, non-symmetric systems and parallelization of sparse systems. A CD-ROM accompanies the text.

Reader Reviews
Scientific parallel computing is what this book is all about, and it does a very good job kneading MPI into the mathematical dough. The book assumes knowledge of mathematics (through Calculus, in some sections, and quite a bit of linear algebra) and little programming experience. This is not a book on C++ programming (not even close, and it does not pretend to be), nor is it a book on MPI programming, parallel computer design, or even the setting up of a suitable software development environment. In fact, it assumes all of the above, which leaves the way uncluttered for the exploration of the application of parallel algorithms upon familiar mathematical concepts. An introductory C++ section is provided to get things rolling, though the level of C++ in the book amounts to simple classes and cin / cout in lieu of C's `printf' nonsense. This is good news, because it maintains the book's focus on the parallelization of mathematical procedures rather than on the irrelevant details of how cute it would all look wrapped up in a needlessly complex object hierarchy. Basic applied C++ and MPI, as and when needed to get the job done. I personally use this text at home on my home-grown parallel Linux computer to investigate the partitioning of algorithms, such as going from a complex function to a Taylor series that may then be distributed to compute nodes. Further MPI, C/C++, and extremely high-level parallel concepts are introduced through the book in a natural progression, as the problems at hand require their introduction. This keeps the book from being bogged down and off-topic. Kudos to the authors for remaining on course through a sea of such tempting distractions. The book is the `how'. BYOW:) ps: my CD was damaged so couldn't evaluate it. 4-stars Comment | | (Report this)