Sequence Data Indexing Method Exploiting the Parallel Processing Resources of GPGPU
Dept. of Computer Engineering, Korea National University of Transportation, Chungbuk, Korea
Advances in Information Technology and Computer Science (ASTL), Volume 25, 2013
@article{kim2013sequence,
title={Sequence Data Indexing Method Exploiting the Parallel Processing Resources of GPGPU},
author={Kim, Taehwan and Song, Seokil and Choi, Dojin and Kwak, Yunsik and Goo, Bongen},
year={2013}
}
In this paper, we propose a new sequence retrieval method that utilizes the parallel processing power of General-Purpose Computation on Graphics Processing Unit (GPGPU). GPGPU has high parallel computing power, due to its characteristic (SIMD) we need to carefully design a method to attain the parallel computing power. Existing CPU based sequence retrieval method does not consider the feature of GPGPU, so even though it is running based on GPGPU it is not able to fully use the parallelism of GPGPU. In this paper, we propose a sequence retrieval method with considering the features of GPGPU, and implement our proposed method, existing method and some variants of the proposed method. Finally, we compare them through experiments with real data.
August 10, 2013 by hgpu
