Výsledky vyhledávání - "Communication-avoiding algorithms"

I/O-Optimal Algorithms for Symmetric Linear Algebra Kernels

Autor: Olivier Beaumont, Lionel Eyraud-Dubois, Julien Langou, Mathieu Vérité

Publikováno v: ACM Symposium on Parallelism in Algorithms and Architectures
ACM Symposium on Parallelism in Algorithms and Architectures, Association for Computing Machinery : SIGACT, SIGARCH, Jul 2022, Philadelphie, United States

International audience; In this paper, we consider two fundamental symmetric kernels in linear algebra: the Cholesky factorization and the symmetric rank-$k$ update (SYRK), with the classical three nested loops algorithms for these kernels. In additi

Externí odkaz: https://explore.openaire.eu/search/publication?articleId=doi_dedup___::912c292e1318ae5d58f91ecae139a6c7
https://hal.inria.fr/hal-03580531

Zobrazit plný text záznamu

Shifted CholeskyQR for computing the QR factorization of ill-conditioned matrices

Autor: Takeshi Fukaya, Yuji Nakatsukasa, Yusaku Yamamoto, Ramaseshan Kannan, Yuka Yanagisawa

The Cholesky QR algorithm is an efficient communication-minimizing algorithm for computing the QR factorization of a tall-skinny matrix X epsilon R-mxn, where m >> n. Unfortunately it is inherently unstable and often breaks down when the matrix is il

Externí odkaz: https://explore.openaire.eu/search/publication?articleId=doi_dedup___::21c257739756a62fcc1bf6e77b8bd6c8

Zobrazit plný text záznamu

A 3D Parallel Algorithm for QR Decomposition

Autor: Mathias Jacquelin, Nicholas Knight, James Demmel, Grey Ballard, Laura Grigori

Publikováno v: Ballard, Grey; Demmel, James; Grigori, Laura; Jacquelin, Mathias; & Knight, Nicholas. (2018). A 3D Parallel Algorithm for QR Decomposition. Proceedings of the 30th on Symposium on Parallelism in Algorithms and Architectures-SPAA '18, abs/1805.05278, 55-65. doi: 10.1145/3210377.3210415. Lawrence Berkeley National Laboratory: Retrieved from: http://www.escholarship.org/uc/item/6qz3r1gf
SPAA '18-30th ACM Symposium on Parallelism in Algorithms and Architectures
SPAA '18-30th ACM Symposium on Parallelism in Algorithms and Architectures, Jul 2018, Vienna, Austria
Proceedings of the 30th on Symposium on Parallelism in Algorithms and Architectures
SPAA
Proceedings of the 30th on Symposium on Parallelism in Algorithms and Architectures-SPAA 18
Proceedings of the 30th on Symposium on Parallelism in Algorithms and Architectures -SPAA '18

© 2018 Association for Computing Machinery. Interprocessor communication often dominates the runtime of large matrix computations. We present a parallel algorithm for computing QR decompositions whose bandwidth cost (communication volume) can be dec

Externí odkaz: https://explore.openaire.eu/search/publication?articleId=doi_dedup___::48a0ac093abc580499a0196112df0e03
http://www.escholarship.org/uc/item/6qz3r1gf

Zobrazit plný text záznamu

Recent Developments in Iterative Methods for Reducing Synchronization

Autor: Frédéric Magoulès, Qinmeng Zou

Publikováno v: 2019 18th International Symposium on Distributed Computing and Applications for Business Engineering and Science (DCABES)
2019 18th International Symposium on Distributed Computing and Applications for Business Engineering and Science (DCABES), Nov 2019, Wuhan, China. pp.163-166, ⟨10.1109/DCABES48411.2019.00048⟩

International audience; On modern parallel architectures, the cost of synchronization among processors can often dominate the cost of floating-point computation. Several modifications of the existing methods have been proposed in order to keep the co

Externí odkaz: https://explore.openaire.eu/search/publication?articleId=doi_dedup___::5af63843ea678c47ed80f0b10fa9946c
https://hal.archives-ouvertes.fr/hal-02399831v2/file/CA.pdf

Zobrazit plný text záznamu

Vyhledávací nástroje:

Upřesnit hledání