A study of implementation schemes for vectorized sparse EBE matrix-vector multiplication |
| |
| Affiliation: | 1. Faculty of Information Technology and Bionics, Pazmany Peter Catholic University, Budapest, Hungary;2. Department of Computer Science, University of Warwick, Coventry, UK;1. High Energy Physics Division, Argonne National Laboratory, Lemont, IL 60439, USA;2. Mathematics and Computer Science Division, Argonne National Laboratory, Lemont, IL 60439, USA;3. Kavli Institute for Cosmological Physics, The University of Chicago, 5640 S. Ellis Ave., Chicago, IL 60637, USA;4. Computation Institute, The University of Chicago, Chicago, IL 60637, USA;5. Argonne Leadership Computing Facility, Argonne National Laboratory, Lemont IL 60439, USA;6. Department of Physics, University of Chicago, Chicago, IL 60637, USA;7. Computer, Computational, and Statistical Sciences Division, Los Alamos National Laboratory, Los Alamos, NM 87545, USA;8. Kitware, 28 Corporate Drive, Clifton Park, NY 12065, USA;9. Computational Research Division, Lawrence Berkeley National Laboratory, Berkeley, CA 90095, USA;10. Department of Electrical Engineering and Computer Science, Northwestern University, 2145 Sheridan Road, Evanston, IL 60208, USA |
| |
| Abstract: | A study of implementation schemes for vectorized sparse element-by-element (EBE) matrix-vector multiplication used in the iterative equation solving of finite element problems is presented. The Hughes, Ferencz and Hallquist scheme and a variation are compared with Torigaki's approach. Computer timing and required memory for each scheme on the IBM 3090/VF vector computer regarding the analysis of large scale three-dimensional solid mechanics problems are presented and results compared. |
| |
| Keywords: | |
| 本文献已被 ScienceDirect 等数据库收录！ |
|
