Full-Text Search:
Home|Journal Papers|About CNKI|User Service|FAQ|Contact Us|中文
《电子学报(英文)》 2018-05
Add to Favorite Get Latest Update

Memory Request Priority Based Warp Scheduling for GPUs

ZHANG Jun;HE Yanxiang;SHEN Fanfan;LI Qing'an;TAN Hai;School of Software,East China University of Technology;Computer School,Wuhan University;  
High performance of GPGPU comes from its super massive multithreading, which makes it more and more widely used especially in the field of throughputoriented. Data locality is one of the important factors affecting the performance of GPGPU. Although GPGPU can exploit intra/inter-warp locality by itself in part, there is still large improvement space for that. In our work, we analyze the characteristics of different applications and propose memory request based warp scheduling to better exploit inter-warp spatial locality. This method can make some warps with good inter-warp locality run faster, which is beneficial to improve the whole performance. Our experimental results show that our proposed method can achieve24.7% and 11.9% average performance improvement over LRR and MRPB respectively.
【Fund】: supported by the National Natural Science Foundation of China(No.61662002 No.61373039 No.61462004);; the Science and Technology Project of the Educational Deparment in Jiangxi Province China(No.GJJ150605);; the Natural Science Foundation of Jiangxi Provice(No.20151BAB207042 No.20161BAB212056);; the Key Research and Developement Plan of the Scientific Department in Jiangxi Province China(No.20161BBE50063)
【CateGory Index】: TP333
Download(CAJ format) Download(PDF format)
CAJViewer7.0 supports all the CNKI file formats; AdobeReader only supports the PDF format.
©2006 Tsinghua Tongfang Knowledge Network Technology Co., Ltd.(Beijing)(TTKN) All rights reserved