Results 11 -
13 of
13
Vice Provost for Academic and International Programs ACKNOWLEDGMENTS
"... would like to thank Dr. Guang R. Gao for his advisement and support during the last two years. His insight and methodology on the research problems always teach me a lot. I really appreciate his comments and valuable guidance provided in the development of my thesis. I would also like to thank Dr. K ..."
Abstract
- Add to MetaCart
(Show Context)
would like to thank Dr. Guang R. Gao for his advisement and support during the last two years. His insight and methodology on the research problems always teach me a lot. I really appreciate his comments and valuable guidance provided in the development of my thesis. I would also like to thank Dr. Kevin Theobold, who guided me to the area of parallel system and computer architecture and taught me lots of computing skills during past three years. Also, I like to thank Yanwei Niu, Dr. Jizhu Lu, Chuan Shen, who worked with me for developing parallel system and parallel applications. One of important outcomes of our work { parallel HMMPFAM will be shown in this thesis. In addition, I would like to acknowledge Dr. Ziang Hu, Dr. Clement Leung, Yuan Zhang for helping to proof read and comment on my thesis at various stages of development. This research is funded in part by NSF, under the NGS grant 0103723, DOE, grant number DE-FC02-01ER25503. I also thank other US Federal agencies and
Recommended Citation
, 2006
"... Hybrid programming in high performance scientific computing ..."
(Show Context)
GPU Computing for Parallel Local Search Metaheuristic Algorithms
- IEEE TRANSACTIONS ON COMPUTERS
, 2012
"... Local search metaheuristics (LSMs) are efficient methods for solving complex problems in science and industry. They allow significantly to reduce the size of the search space to be explored and the search time. Nevertheless, the resolution time remains prohibitive when dealing with large problem ins ..."
Abstract
- Add to MetaCart
(Show Context)
Local search metaheuristics (LSMs) are efficient methods for solving complex problems in science and industry. They allow significantly to reduce the size of the search space to be explored and the search time. Nevertheless, the resolution time remains prohibitive when dealing with large problem instances. Therefore, the use of GPU-based massively parallel computing is a major complementary way to speed up the search. However, GPU computing for LSMs is rarely investigated in the literature. In this paper, we introduce a new guideline for the design and implementation of effective LSMs on GPU. Very efficient approaches are proposed for CPU-GPU data transfer optimization, thread control, mapping of neighboring solutions to GPU threads and memory management. These approaches have been experimented using four well-known combinatorial and continuous optimization problems and four GPU configurations. Compared to a CPU-based execution, accelerations up to ×80 are reported for the large combinatorial problems and up to ×240 for a continuous problem. Finally, extensive experiments demonstrate the strong potential of GPU-based LSMs compared to cluster or grid-based parallel architectures.
