Electromagnetic computation by GPGPU based CIP method : a study on thread block size
-
- HIRAUME Tomoaki
- Graduate School of Engineering, Kyushu Sangyo University
-
- MATSUOKA Tsuyoshi
- Faculty of Engineering, Kyushu Sangyo University
-
- OGATA Masato
- Faculty of Engineering, Kyushu Sangyo University
Bibliographic Information
- Other Title
-
- GPUを用いたCIP法による電磁界解析の高速化 : ブロック数とスレッド数の検討(学生若手発表,学生若手発表および一般)
- GPUを用いたCIP法による電磁界解析の高速化 : ブロック数とスレッド数の検討
- GPU オ モチイタ CIPホウ ニ ヨル デンジカイ カイセキ ノ コウソクカ : ブロックスウ ト スレッドスウ ノ ケントウ
Search this article
Abstract
GPU (graphics processing unit) has a lot of simple numerical computation cores with parallel operation, and the GPGPU (general purpose computing on GPU) is one of the candidates to improve computation efficiency. In this paper, we numericaly analyze the 1D electromagnetic wave propagation based on the CIP method by using the GPU. The computation times are evaluated by changing the thread size per one block and the results shows that the thread size is chosen to be larger than 64 for efficient computation.
Journal
-
- ITE Technical Report
-
ITE Technical Report 37.4 (0), 103-106, 2013
The Institute of Image Information and Television Engineers
- Tweet
Keywords
Details 詳細情報について
-
- CRID
- 1390282679506777728
-
- NII Article ID
- 110009594604
-
- NII Book ID
- AN1059086X
-
- ISSN
- 24241970
- 13426893
-
- NDL BIB ID
- 024272904
-
- Text Lang
- ja
-
- Data Source
-
- JaLC
- NDL
- CiNii Articles
-
- Abstract License Flag
- Disallowed