图形处理器中纹理贴图算法的性能分析  被引量:2

Performance characterization analysis of texture mapping algorithms used in graphics processors

在线阅读下载全文

作  者:樊萌 蒋林 邓军勇 刘洋 FAN Meng;JIANG Lin;DENG Junyong;LIU Yang(School of Electronic Engineering,Xi'an University of Posts and Telecommunications,Xi'an 710121,China;Integrated Circuit Design Laboratory,Xi'an University of Science and Technology,Xi'an 710054,China)

机构地区:[1]西安邮电大学电子工程学院,陕西西安710121 [2]西安科技大学集成电路设计实验室,陕西西安710054

出  处:《西安邮电大学学报》2020年第2期74-79,共6页Journal of Xi’an University of Posts and Telecommunications

基  金:国家自然科学基金项目(61834005,61772417,61802304,61272120,61602377,61634004);陕西省科技统筹创新工程项目(2016KTZDGY02-04-02);陕西省国际科技合作计划项目(2018KW-006);陕西省重点研发计划项目(2017GY-060)。

摘  要:针对在可重构图形处理器中根据不同性能特征选择最优纹理贴图算法的问题,采用Coffee Lake架构处理器上的硬件性能计数器,分析了最近邻点采样、双线性滤波、Mipmap结合线性滤波等3种算法在立方体、球体、圆环和茶壶等4种渲染场景对象中的数据移动量、计算量、功耗、数据缓存以及各指标的相关性。仿真结果表明,为了提高图形处理器性能,点采样和Mipmap结合线性滤波算法重构可减少数据移动量,点采样和双线性滤波算法重构可减少计算量,采用点采样算法可降低功耗,使用双线性滤波算法可提高缓存命中率。How to choose the optimal texture mapping algorithm based on different performance characteristics in reconfigurable graphics processor is discussed.The hardware performance counters on the Coffee Lake architecture processor are used to analyze the data movement,calculation,power consumption,storage access,and correlation between performance/energy and performance indicators of the nearest neighbor sampling,bilinear filtering,Mipmap combined with linear filtering algorithms in multiple rendering objects.Experiment results show that in order to improve the performance of the graphics processor,reconstruction of point sampling and Mipmap combined with linear filtering algorithm can reduce the amount of data movement,reconstruction of point sampling and bilinear filtering algorithm can reduce the computation,point sampling algorithm can reduce power consumption,and bilinear filtering algorithm can improve the cache hit ratio.

关 键 词:可重构计算 图形处理器 纹理贴图算法 

分 类 号:TP302[自动化与计算机技术—计算机系统结构]

 

参考文献:

正在载入数据...

 

二级参考文献:

正在载入数据...

 

耦合文献:

正在载入数据...

 

引证文献:

正在载入数据...

 

二级引证文献:

正在载入数据...

 

同被引文献:

正在载入数据...

 

相关期刊文献:

正在载入数据...

相关的主题
相关的作者对象
相关的机构对象