Paper Abstract and Keywords |
Presentation |
2012-09-27 13:25
A Study of Acceleration of Conjugate Gradient Method Using GPU Hajime Katsuda, Keisuke Konno, Qiang Chen, Kunio Sawaya (Tohoku Univ.), Kei Yokogawa, Qiaowei Yuan (Sendai National College of Technology), Toshihiro Sezai (JAXA) AP2012-86 |
Abstract |
(in Japanese) |
(See Japanese page) |
(in English) |
In method of moments (MoM), large computing time consumption is a serious problem for solving a large-scale matrix equation. To overcome this problem, Graphics Processing Units (GPU) have been used for acceleration of matrix equation solver in recent years. In GPU computing, it is assumed that assignment of numerical operation to each processor, memory access scheme by each processor and assignment of numerical operation between CPU and GPU should be optimized for reduction of computing time. In this report, Conjugate Gradient (CG) method is accelerated by optimized GPU and its computing time is quantitatively evaluated. |
Keyword |
(in Japanese) |
(See Japanese page) |
(in English) |
MoM / GPU / Conjugate Gradient method / / / / / |
Reference Info. |
IEICE Tech. Rep., vol. 112, no. 216, AP2012-86, pp. 25-29, Sept. 2012. |
Paper # |
AP2012-86 |
Date of Issue |
2012-09-20 (AP) |
ISSN |
Print edition: ISSN 0913-5685 Online edition: ISSN 2432-6380 |
Copyright and reproduction |
All rights are reserved and no part of this publication may be reproduced or transmitted in any form or by any means, electronic or mechanical, including photocopy, recording, or any information storage and retrieval system, without permission in writing from the publisher. Notwithstanding, instructors are permitted to photocopy isolated articles for noncommercial classroom use without fee. (License No.: 10GA0019/12GB0052/13GB0056/17GB0034/18GB0034) |
Download PDF |
AP2012-86 |
|