{"title":"基于混合数字系统的低成本浮点可编程顶点处理器的设计","authors":"Shen-Fu Hsiao, Chan-Feng Chiu, Chia-Sheng Wen","doi":"10.1109/ICICDT.2011.5783231","DOIUrl":null,"url":null,"abstract":"Recent OpenGL ES 2.0 API Specification for embedded systems graphics operations requires programmable vertex shader and fragment shader to process vertex and pixel data. Calculation of dot-product for two vectors and transcendental functions for a scalar are two fundamental arithmetic operations in the vertex processing. Since some complicated arithmetic operations in binary number system (BNS) turn into simple operations of addition and/or multiplication in the logarithmic number system (LNS), we present a low-cost design of a floating-point programmable vertex processor based on hybrid BNS and LNS. The proposed design achieves at least the same (or even higher) precision with much lower cost compared with recent similar implementations.","PeriodicalId":402000,"journal":{"name":"2011 IEEE International Conference on IC Design & Technology","volume":"9 1","pages":"0"},"PeriodicalIF":0.0000,"publicationDate":"2011-05-02","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":"5","resultStr":"{\"title\":\"Design of a low-cost floating-point programmable vertex processor for mobile graphics applications based on hybrid number system\",\"authors\":\"Shen-Fu Hsiao, Chan-Feng Chiu, Chia-Sheng Wen\",\"doi\":\"10.1109/ICICDT.2011.5783231\",\"DOIUrl\":null,\"url\":null,\"abstract\":\"Recent OpenGL ES 2.0 API Specification for embedded systems graphics operations requires programmable vertex shader and fragment shader to process vertex and pixel data. Calculation of dot-product for two vectors and transcendental functions for a scalar are two fundamental arithmetic operations in the vertex processing. Since some complicated arithmetic operations in binary number system (BNS) turn into simple operations of addition and/or multiplication in the logarithmic number system (LNS), we present a low-cost design of a floating-point programmable vertex processor based on hybrid BNS and LNS. The proposed design achieves at least the same (or even higher) precision with much lower cost compared with recent similar implementations.\",\"PeriodicalId\":402000,\"journal\":{\"name\":\"2011 IEEE International Conference on IC Design & Technology\",\"volume\":\"9 1\",\"pages\":\"0\"},\"PeriodicalIF\":0.0000,\"publicationDate\":\"2011-05-02\",\"publicationTypes\":\"Journal Article\",\"fieldsOfStudy\":null,\"isOpenAccess\":false,\"openAccessPdf\":\"\",\"citationCount\":\"5\",\"resultStr\":null,\"platform\":\"Semanticscholar\",\"paperid\":null,\"PeriodicalName\":\"2011 IEEE International Conference on IC Design & Technology\",\"FirstCategoryId\":\"1085\",\"ListUrlMain\":\"https://doi.org/10.1109/ICICDT.2011.5783231\",\"RegionNum\":0,\"RegionCategory\":null,\"ArticlePicture\":[],\"TitleCN\":null,\"AbstractTextCN\":null,\"PMCID\":null,\"EPubDate\":\"\",\"PubModel\":\"\",\"JCR\":\"\",\"JCRName\":\"\",\"Score\":null,\"Total\":0}","platform":"Semanticscholar","paperid":null,"PeriodicalName":"2011 IEEE International Conference on IC Design & Technology","FirstCategoryId":"1085","ListUrlMain":"https://doi.org/10.1109/ICICDT.2011.5783231","RegionNum":0,"RegionCategory":null,"ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":null,"EPubDate":"","PubModel":"","JCR":"","JCRName":"","Score":null,"Total":0}
引用次数: 5
摘要
最近针对嵌入式系统图形操作的OpenGL ES 2.0 API规范要求可编程顶点着色器和片段着色器来处理顶点和像素数据。两个向量的点积计算和标量的超越函数计算是顶点处理中的两个基本算术运算。针对二进制数系统(BNS)中一些复杂的算术运算在对数数系统(LNS)中转化为简单的加法和乘法运算的问题,提出了一种基于BNS和LNS混合的浮点可编程顶点处理器的低成本设计。与最近的类似实现相比,所提出的设计以更低的成本实现了至少相同(甚至更高)的精度。
Design of a low-cost floating-point programmable vertex processor for mobile graphics applications based on hybrid number system
Recent OpenGL ES 2.0 API Specification for embedded systems graphics operations requires programmable vertex shader and fragment shader to process vertex and pixel data. Calculation of dot-product for two vectors and transcendental functions for a scalar are two fundamental arithmetic operations in the vertex processing. Since some complicated arithmetic operations in binary number system (BNS) turn into simple operations of addition and/or multiplication in the logarithmic number system (LNS), we present a low-cost design of a floating-point programmable vertex processor based on hybrid BNS and LNS. The proposed design achieves at least the same (or even higher) precision with much lower cost compared with recent similar implementations.