Scientific computing with multicore and accelerators [electronic resource] / edited by Jakub Kurzak, David A. Bader, Jack Dongarra.
Contributor(s): Kurzak, Jakub | Bader, David A | Dongarra, J. J.Material type: BookSeries: Chapman & Hall/CRC computational science series: 10.Publisher: Boca Raton : CRC Press, 2011Description: xxxiii, 480 p. : ill.ISBN: 9781439825372 (ebook : PDF).Subject(s): Science -- Data processing | Engineering -- Data processing | High performance computing | MultiprocessorsGenre/Form: Electronic books.Additional physical formats: No titleOnline resources: Distributed by publisher. Purchase or institutional license may be required for access. Also available in print edition.
Includes bibliographical references and index.
1. Dense linear algebra -- 2. Sparse linear algebra -- 3. Multigrid methods -- 4. Fast Fourier transforms -- 5. Combinatorial algorithms -- 6. Stencil algorithms -- 7. Bioinformatics -- 8. Molecular modeling -- 9. Complementary topics.
"The current trend in microprocessor architecture is toward powerful multicore designs in which a node contains several full-featured processing cores, private and shared caches, and memory. The IBM Cell Broadband Engine (B.E.) and Graphics Processing Units (GPUs) are two accelerators that are used for a variety of computations, including signal processing and quantum chemistry. This is the first reference on the use of Cell B.E. and GPUs as accelerators for numerical kernels, algorithms, and computational science and engineering applications. With contributions from leading experts, the book covers a broad range of topics on the increased role of these accelerators in scientific computing"-- Provided by publisher.
Also available in print edition.
Mode of access: World Wide Web.