Paper Abstract and Keywords |
Presentation |
2017-07-26 11:45
Preliminary evaluation of large-scale computer-generated hologram computation time by using multi-GPU cluster Shinpei Watanabe (Utsunomiya Univ.), Boaz-Jessie Jackin (NICT), Takeshi Ohkawa, Kanemitsu Ootsu, Takashi Yokota, Yoshio Hayasaki, Toyohiko Yatagai, Takanobu Baba (Utsunomiya Univ.) CPSY2017-20 |
Abstract |
(in Japanese) |
(See Japanese page) |
(in English) |
Computer generated hologram (CGH) is a technology for realizing 3D displays. Large-scale CGH has
a merit that it resolves problems existing 3D displays have. However, generating CGH requires memory space
and calculation time, and these are in proportion to pixel number. Thus, large-scale CGH can not be generated
in real time and it is the reason why CGH is not suit to practical use yet. Computation of CGH is comprised of
data-independent operations and GPU has many processing cores. We had accelerated the CGH generation by using
single GPU and object decomposition method. However that still takes long time. Thus, we accelerate the CGH
generation by using multi-GPU cluster. The results show that we have achieved 134.7 times speed-up compared to
sequential program execution by CPU. |
Keyword |
(in Japanese) |
(See Japanese page) |
(in English) |
CGH / multi-GPU / cluster / Object decompositon method / / / / |
Reference Info. |
IEICE Tech. Rep., vol. 117, no. 153, CPSY2017-20, pp. 25-30, July 2017. |
Paper # |
CPSY2017-20 |
Date of Issue |
2017-07-19 (CPSY) |
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 |
CPSY2017-20 |
|