A Structured Approach Using BSP Rob H. Bisseling deals with parallel scientific computation and provides a structured method for designing and implementing parallel algorithms for modern parallel computers. The textbook focuses on scientific and big data applications using the Bulk Synchronous Parallel (BSP) model.
The book discusses the use of BSP to develop parallel solutions, from problem description to actual programs in BSPlib. Each chapter combines theory with practical experiments and exercises, supported by extensive bibliographic references.
Features
This edition includes up-to-date algorithms for parallel scientific computation, including sorting, graphs, LU decomposition, and improved data partitioning for sparse matrix computations. Thanks to the hybrid BSP approach, both shared and distributed memory are used efficiently. The BSPedupack package, available on the author’s website, provides source code and test programs for modern BSPlib implementations such as MulticoreBSP and BSPonMPI.
Product specifications
- Author: Rob H. Bisseling (Professor of Scientific Computing, Utrecht University)
- Publisher: Oxford University Press
- Release date: 2020-09-30
- Number of pages: 410
- ISBN: 9780198788355
- Subject: Computer science
- BISAC: COMPUTERS / General
About the author
Rob Bisseling is a professor of Scientific Computing at Utrecht University and has extensive experience with applications of parallel computer technology in science. Previously, he worked as a researcher at Shell Amsterdam on parallel computations.