TU Darmstadt / ULB / TUbiblio

Analysis of Schedule and Layout Tuning for Sparse Matrices With Compound Entries on GPUs

Mueller-Roemer, J. S. ; Stork, A. ; Fellner, D. (2020):
Analysis of Schedule and Layout Tuning for Sparse Matrices With Compound Entries on GPUs.
In: Computer Graphics Forum, ISSN 01677055,
DOI: 10.1111/cgf.13957,
[Article]

Abstract

Large sparse matrices with compound entries, i.e. complex and quaternionic matrices as well as matrices with dense blocks, are a core component of many algorithms in geometry processing, physically based animation and other areas of computer graphics. We generalize several matrix layouts and apply joint schedule and layout autotuning to improve the performance of the sparse matrix-vector product on massively parallel graphics processing units. Compared to schedule tuning without layout tuning, we achieve speedups of up to 5.5×. In comparison to cuSPARSE, we achieve speedups of up to 4.7×.

Item Type: Article
Erschienen: 2020
Creators: Mueller-Roemer, J. S. ; Stork, A. ; Fellner, D.
Title: Analysis of Schedule and Layout Tuning for Sparse Matrices With Compound Entries on GPUs
Language: English
Abstract:

Large sparse matrices with compound entries, i.e. complex and quaternionic matrices as well as matrices with dense blocks, are a core component of many algorithms in geometry processing, physically based animation and other areas of computer graphics. We generalize several matrix layouts and apply joint schedule and layout autotuning to improve the performance of the sparse matrix-vector product on massively parallel graphics processing units. Compared to schedule tuning without layout tuning, we achieve speedups of up to 5.5×. In comparison to cuSPARSE, we achieve speedups of up to 4.7×.

Journal or Publication Title: Computer Graphics Forum
Uncontrolled Keywords: General Purpose Computation on Graphics Processing Unit (GPGPU), Parallel computing, Matrix operations
Divisions: 20 Department of Computer Science
20 Department of Computer Science > Interactive Graphics Systems
Date Deposited: 07 May 2020 09:47
DOI: 10.1111/cgf.13957
Official URL: https://doi.org/10.1111/cgf.13957
Additional Information:

Early View-Online Version of Record before inclusion in an issue

Export:
Suche nach Titel in: TUfind oder in Google
Send an inquiry Send an inquiry

Options (only for editors)
Show editorial Details Show editorial Details