DocumentCode
2580580
Title
A New Software Architecture for Ultra-large-scale Rendering Cloud
Author
Weini, Zhou ; Yongquan, Lu ; Pengdong, Gao ; Chu, Qiu ; Quan, Qi
Author_Institution
High Performance Comput. Center, Commun., Univ. of China, Beijing, China
fYear
2012
fDate
19-22 Oct. 2012
Firstpage
196
Lastpage
199
Abstract
A kind of new five-level software architecture is presented in this paper, which is designed elaborately for ultra-large-scale rendering cloud. This software architecture consists of five levels, infrastructure layer, rendering service layer, rendering application layer, service management and access interface. Based on the proposed software architecture, we have developed a new type of rendering management software called Golden Farm Cluster Rendering Platform. And comprehensive tests have been carried on the super computer, Tian He 1A. Experimental results show that Golden Farm can effectively manage the rendering jobs on ultra-large-scale high performance cluster with up to 2000 computing nodes and also obtain good performance. In addition, the platform has created an important record, rendering a 3D model by using 24000 CPU cores in parallel.
Keywords
parallel machines; rendering (computer graphics); software architecture; solid modelling; 3D model; CPU cores; TianHe 1A; access interface; five-level software architecture; golden farm cluster rendering platform; infrastructure layer; rendering application layer; rendering service layer; service management; super computer; ultra-large-scale rendering cloud; Animation; Computational modeling; Computers; Rendering (computer graphics); Scheduling; Software; Software architecture; Golden Farm; animation rendering; cluster rendering management platform; five-level architecture; rendering cloud;
fLanguage
English
Publisher
ieee
Conference_Titel
Distributed Computing and Applications to Business, Engineering & Science (DCABES), 2012 11th International Symposium on
Conference_Location
Guilin
Print_ISBN
978-1-4673-2630-8
Type
conf
DOI
10.1109/DCABES.2012.10
Filename
6385270
Link To Document