Achieving High Throughput Sequencing with Graphics Processing Units
Department of Computer Science, Arkansas State University, Jonesboro, AR 72467, USA
The 2011 International Conference on Parallel and Distributed Processing Techniques and Applications (PDPTA’11), 2011
@article{chen2011achieving,
title={Achieving High Throughput Sequencing with Graphics Processing Units},
author={Chen, S. and Zhang, C. and Shen, F. and Bai, L. and Jiang, H. and Herman, D.},
year={2011}
}
High throughput sequencing has become a powerful technique for genome analysis after this concept was raised in recent years. Currently, there is a huge demand from patients that have genetic diseases which cannot be satisfied due to the limitation of computation power. Though several softwares are developed using currently most efficient algorithm to deal with various types of sequencing problems, the CPU seems to be too expensive to process endless data economically because CPUs are not designed adaptive for data parallel problem. The latest Fermi architecture released by NVIDIA provides considerable number of streaming processors, bigger size of register file and 1 MB cache, which makes it very competitive for data parallel processing. This paper tries a simple sequence alignment method on GPU and compared the real world performance between CPU and GPU. Experiment shows that GPU may have a good potential with similar problems.
December 10, 2011 by hgpu
