Joint Schedule and Layout Autotuning for Sparse Matrices with Compound Entries on GPUs

Johannes Mueller-Roemer, André Stork, Dieter W. Fellner

Research output: Chapter in Book/Report/Conference proceedingConference paperpeer-review

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 5:5x. In comparison to cuSPARSE, we achieve speedups of up to 4:7x.
Original languageEnglish
Title of host publicationVision, Modeling, and Visualization
PublisherUniversity of Rostock
ISBN (Electronic)978-3-03868-098-7
Publication statusPublished - 2019

Publication series

NameVision, Modeling, and Visualization / von/by Schulz, Hans-J\örg [Ed.] [et al.]. - European Association for Computer Graphics (Eurographics): University of Rostock. - 978-3-03868-098-7 (ISBN). - (2019)
PublisherUniversity of Rostock

Keywords

  • Lead Topic: Digitized Work
  • Research Area: (Interactive) simulation (SIM)
  • General Purpose Computation on Graphics Processing Unit (GPGPU)
  • GPU computing
  • Linear systems
  • Code generation

Fields of Expertise

  • Information, Communication & Computing

Cite this