A Synchronization Mechanism between CUDA Blocks for GPU

CORC > 沈阳自动化研究所 > 中国科学院沈阳自动化研究所 > 数字工厂研究室

	A Synchronization Mechanism between CUDA Blocks for GPU
	Wang, Bingru ; Feng, Jun ; Wang F(王锋); Zhang, Changyou
	2017
会议名称	2nd International Conference on Control, Automation and Artificial Intelligence (CAAI)
会议日期	June 25-26, 2017
会议地点	Sanya, CHINA
关键词	GPU synchronization mechanism SSSP parallel computing delta-stepping CUDA
页码	251-254
通讯作者	Feng, Jun
中文摘要	GPU(Graphic Processing Unit) provides a promising solution with massive threads and its advantage is high performance computing. The emergence of CUDA(Compute Unified Device Architecture) opens the door of using GPU's powerful computing power. However, because of the limitation of CUDA itself, direct communication is not supported between SMs(streaming multiprocessors) on GPU. It is time-consuming by atomic operation or barrier synchronization. A synchronization mechanism has been proposed in this paper, that is, on the premise of result available, the times of kernel launched should be reduced. Each kernel launched, it should be computed enough on GPU, the results back to the CPU. Based on SSSP, the validity of this method is illustrated by delta-stepping. For facebook dataset, compared with atomic operation, the speedup ratio is about 1.8. For New York map dataset, compared with atomic operation and barrier synchronization, the speedup ratio is about 9.3 and 1.7 separately.
收录类别	CPCI（ISTP）
产权排序	3
会议主办者	Sci & Engn Res Ctr
会议录	PROCEEDINGS OF THE 2017 2ND INTERNATIONAL CONFERENCE ON CONTROL, AUTOMATION AND ARTIFICIAL INTELLIGENCE (CAAI 2017)
会议录出版者	ATLANTIS PRESS
会议录出版地	PARIS
语种	英语
ISSN号	1951-6851
WOS记录号	WOS:000426676200056
内容类型	会议论文
源URL	[http://ir.sia.cn/handle/173321/21555]
专题	沈阳自动化研究所_数字工厂研究室
作者单位	1.Institute of Software, Chinese Academy of Science, Beijing, China 2.Shenyang Institute of Automation, Chinese Academy of Science, Shenyang, China 3.Shijiazhuang Tiedao University, Shijiazhuang, China
推荐引用方式 GB/T 7714	Wang, Bingru,Feng, Jun,Wang F,et al. A Synchronization Mechanism between CUDA Blocks for GPU[C]. 见:2nd International Conference on Control, Automation and Artificial Intelligence (CAAI). Sanya, CHINA. June 25-26, 2017.