Chi-Tsai Yeh, Ching-Yuan Lin, W. Chiou, Ing-Jer Huang
{"title":"Exploration of tile/triangle-based rendering approaches for 3D graphics SoC","authors":"Chi-Tsai Yeh, Ching-Yuan Lin, W. Chiou, Ing-Jer Huang","doi":"10.1109/ISNE.2010.5669157","DOIUrl":null,"url":null,"abstract":"Three dimension (3D) graphics applications have been widely used in handheld devices which are an inevitable tendency in the future. In this paper, we model a complex 3D graphics SoC hardware by using SystemC and run some testbenches on this platform. For speeding up the simulation time, we adopt Transaction Level Model (TLM) and implement two type of rendering approaches, tile-based and triangle-based, to find out the optimal architecture. According to experimental results, triangle-based rendering is 9–24% faster than tile-based rendering in smaller objects. Considering the energy consumption, the number of memory access of the former is 6–43% less than the latter.","PeriodicalId":412093,"journal":{"name":"2010 International Symposium on Next Generation Electronics","volume":"24 1","pages":"0"},"PeriodicalIF":0.0000,"publicationDate":"2010-12-17","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":"0","resultStr":null,"platform":"Semanticscholar","paperid":null,"PeriodicalName":"2010 International Symposium on Next Generation Electronics","FirstCategoryId":"1085","ListUrlMain":"https://doi.org/10.1109/ISNE.2010.5669157","RegionNum":0,"RegionCategory":null,"ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":null,"EPubDate":"","PubModel":"","JCR":"","JCRName":"","Score":null,"Total":0}
引用次数: 0
Abstract
Three dimension (3D) graphics applications have been widely used in handheld devices which are an inevitable tendency in the future. In this paper, we model a complex 3D graphics SoC hardware by using SystemC and run some testbenches on this platform. For speeding up the simulation time, we adopt Transaction Level Model (TLM) and implement two type of rendering approaches, tile-based and triangle-based, to find out the optimal architecture. According to experimental results, triangle-based rendering is 9–24% faster than tile-based rendering in smaller objects. Considering the energy consumption, the number of memory access of the former is 6–43% less than the latter.