Batched one-sided factorizations of tiny matrices using GPUs: Challenges and countermeasures

Abdelfattah, Ahmad; Haidar, Azzam; Tomov, Stanimire; Dongarra, Jack

doi:10.1016/j.jocs.2018.01.005

Title: Batched one-sided factorizations of tiny matrices using GPUs: Challenges and countermeasures

Full Record
Other Related Research

Authors:: Abdelfattah, Ahmad; Haidar, Azzam; Tomov, Stanimire; Dongarra, Jack

Publication Date:: Tue May 01 00:00:00 EDT 2018

Sponsoring Org.:: USDOE National Nuclear Security Administration (NNSA)

OSTI Identifier:: 1582813

Resource Type:: Publisher's Accepted Manuscript

Journal Name:: Journal of Computational Science

Additional Journal Information:: Journal Name: Journal of Computational Science Journal Volume: 26 Journal Issue: C; Journal ID: ISSN 1877-7503

Publisher:: Elsevier

Country of Publication:: Netherlands

Language:: English

Citation Formats


                    Abdelfattah, Ahmad, Haidar, Azzam, Tomov, Stanimire, and Dongarra, Jack. Batched one-sided factorizations of tiny matrices using GPUs: Challenges and countermeasures.  Netherlands: N. p., 2018. 
Web.  doi:10.1016/j.jocs.2018.01.005.

Copy to clipboard


                    Abdelfattah, Ahmad, Haidar, Azzam, Tomov, Stanimire, & Dongarra, Jack. Batched one-sided factorizations of tiny matrices using GPUs: Challenges and countermeasures.  Netherlands.  https://doi.org/10.1016/j.jocs.2018.01.005

Copy to clipboard


                    Abdelfattah, Ahmad, Haidar, Azzam, Tomov, Stanimire, and Dongarra, Jack. Tue .  
"Batched one-sided factorizations of tiny matrices using GPUs: Challenges and countermeasures".  Netherlands.  https://doi.org/10.1016/j.jocs.2018.01.005.

Copy to clipboard


                    
@article{osti_1582813,

  title        = {Batched one-sided factorizations of tiny matrices using GPUs: Challenges and countermeasures},

  author       = {Abdelfattah, Ahmad and Haidar, Azzam and Tomov, Stanimire and Dongarra, Jack},

  abstractNote = {},

  doi          = {10.1016/j.jocs.2018.01.005},

  journal      = {Journal of Computational Science},

  number       = C,

  volume       = 26,

  place        = {Netherlands},

  year         = {Tue May 01 00:00:00 EDT 2018},

  month        = {Tue May 01 00:00:00 EDT 2018}

}

Copy to clipboard

Journal Article:

Free Publicly Available Full Text

Accepted Manuscript (Publisher)

Publisher's Version of Record
https://doi.org/10.1016/j.jocs.2018.01.005

Other availability

Search WorldCat to find libraries that may hold this journal

Citation Metrics:

Cited by: 6 works

Citation information provided by
Web of Science

Save / Share:

Export Metadata

Save to My Library

Similar Records in DOE PAGES and OSTI.GOV collections:

Accelerating the SVD bi-diagonalization of a batch of small matrices using GPUs

Journal Article Dong, Tingxing ; Haidar, Azzam ; Tomov, Stanimire ; ... - Journal of Computational Science
Cited by 7
https://doi.org/10.1016/j.jocs.2018.01.007
Implementation and Tuning of Batched Cholesky Factorization and Solve for NVIDIA GPUs

Journal Article Kurzak, Jakub ; Anzt, Hartwig ; Gates, Mark ; ... - IEEE Transactions on Parallel and Distributed Systems

Many problems in engineering and scientific computing require the solution of a large number of small systems of linear equations. Due to their high processing power, Graphics Processing Units became an attractive target for this class of problems, and routines based on the LU and the QR factorization have been provided by NVIDIA in the cuBLAS library. This work addresses the situation where the systems of equations are symmetric positive definite. The paper describes the implementation and tuning of the kernels for the Cholesky factorization and the forward and backward substitution. Targeted workloads involve the solution of thousands of linearmore »« less
https://doi.org/10.1109/tpds.2015.2481890
A Framework for Batched and GPU-Resident Factorization Algorithms Applied to Block Householder Transformations

Book Dong, Tingzing Tim ; Tomov, Stanimire Z ; Luszczek, Piotr R ; ...

As modern hardware keeps evolving, an increasingly effective approach to developing energy efficient and high-performance solvers is to design them to work on many small size and independent problems. Many applications already need this functionality, especially for GPUs, which are currently known to be about four to five times more energy efficient than multicore CPUs. We describe the development of one-sided factorizations that work for a set of small dense matrices in parallel, and we illustrate our techniques on the QR factorization based on Householder transformations. We refer to this mode of operation as a batched factorization. Our approach ismore »« less
ONE-SIDED NORMAL TOLERANCE FACTORS--NEW TABLES AND EXTENDED USE OF TABLES

Technical Report Burrows, G L
ONE-SIDED NORMAL TOLERANCE FACTORS. New Tables and Extended Use of Tables

Technical Report Burrows, G L

Similar Records