Data locality optimization for synthesis of efficient out-of-core algorithms
Document Type
Article
Publication Date
1-1-2003
Abstract
This paper describes an approach to synthesis of efficient out-of-core code for a class of imperfectly nested loops that represent tensor contraction computations. Tensor contraction expressions arise in many accurate computational models of electronic structure. The developed approach combines loop fusion with loop tiling and uses a performance-model driven approach to loop tiling for the generation of out-of-core code. Experimental measurements are provided that show a good match with model-based predictions and demonstrate the effectiveness of the proposed algorithm. © Springer-Verlag Berlin Heidelberg 2003.
Publication Source (Journal or Book title)
Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
First Page
406
Last Page
417
Recommended Citation
Krishnan, S., Krishnamoorthy, S., Baumgartner, G., Cociorva, D., Lam, C., Sadayappan, P., Ramanujam, J., Bernholdt, D., & Choppella, V. (2003). Data locality optimization for synthesis of efficient out-of-core algorithms. Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics), 2913, 406-417. https://doi.org/10.1007/978-3-540-24596-4_44
