How to Write Fast Numerical Code: A Small Introduction Export

@incollection{citeulike:5969490, abstract = {The complexity of modern computing platforms has made it extremely difficult to write numerical code that achieves the best possible performance. Straightforward implementations based on algorithms that minimize the operations count often fall short in performance by at least one order of magnitude. This tutorial introduces the reader to a set of general techniques to improve the performance of numerical code, focusing on optimizations for the computer's memory hierarchy. Further, program generators are discussed as a way to reduce the implementation and optimization effort. Two running examples are used to demonstrate these techniques: matrix-matrix multiplication and the discrete Fourier transform.}, author = {Chellappa, Srinivas and Franchetti, Franz and P\"{u}schel, Markus}, citeulike-article-id = {5969490}, citeulike-linkout-0 = {http://dx.doi.org/10.1007/978-3-540-88643-3_5}, citeulike-linkout-1 = {http://www.springerlink.com/content/lnr2462533470323}, doi = {10.1007/978-3-540-88643-3_5}, journal = {Generative and Transformational Techniques in Software Engineering II}, keywords = {floating\_point, high-performance\_computing, linear\_algebra, matrix, performance\_analysis, teaching}, pages = {196--259}, posted-at = {2009-10-19 08:31:37}, priority = {2}, title = {How to Write Fast Numerical Code: A Small Introduction}, url = {http://dx.doi.org/10.1007/978-3-540-88643-3_5}, year = {2008} }

TY - CHAP ID - citeulike:5969490 L3 - citeulike-article-id:5969490 N2 - The complexity of modern computing platforms has made it extremely difficult to write numerical code that achieves the best possible performance. Straightforward implementations based on algorithms that minimize the operations count often fall short in performance by at least one order of magnitude. This tutorial introduces the reader to a set of general techniques to improve the performance of numerical code, focusing on optimizations for the computer’s memory hierarchy. Further, program generators are discussed as a way to reduce the implementation and optimization effort. Two running examples are used to demonstrate these techniques: matrix-matrix multiplication and the discrete Fourier transform. JF - Generative and Transformational Techniques in Software Engineering II EP - 259 TI - How to Write Fast Numerical Code: A Small Introduction SP - 196 KW - floating_point KW - high-performance_computing KW - linear_algebra KW - matrix KW - performance_analysis KW - teaching AU - Chellappa, Srinivas AU - Franchetti, Franz AU - Püschel, Markus PY - 2008/// UR - http://dx.doi.org/10.1007/978-3-540-88643-3_5 ER -