CUDA加速的地图代数并行算法
|
摘要
针对传统地图代数实现方法应用于海量栅格数据计算时效率低下的问题,在一种全新的GPU并行编程模型CUDA上,利用地图代数算子体现出来的基于栅格点集、处理流程相对固定、数据处理具有内在的并行性等特点,将传统的串行算法映射到GPU并行处理架构上,旨在从串行算法的并行化映射、计算机图形处理器资源的自适应参数调整等多角度来研究地图代数空间并行算法的实现机制,为空间分析算法的优化研究提供一种新的解决思路。
To improve the efficiency in traditional method of map algebra calculation for gigantic raster data,arithmetic operators characteristics are applied,including relatively fixed process flow and inherently parallel specialty,and a new GPU parallel programming model named CUDA is selectsed as technique supports.The realization mechanism surrounding parallel mapping of serial algorithms is discussed in adaptive parameter adjustments on computer graphic processor resources,thus providing a new solution for optimized research of spatial analytic algorithms.
To improve the efficiency in traditional method of map algebra calculation for gigantic raster data,arithmetic operators characteristics are applied,including relatively fixed process flow and inherently parallel specialty,and a new GPU parallel programming model named CUDA is selectsed as technique supports.The realization mechanism surrounding parallel mapping of serial algorithms is discussed in adaptive parameter adjustments on computer graphic processor resources,thus providing a new solution for optimized research of spatial analytic algorithms.
引文
[1]胡鹏,游涟,胡海.地图代数概论[M].北京:测绘出版社,2008.
[2]吴连贵,易瑜,李肯立.基于CUDA的地震数据相干体并行算法[J].计算机应用,2009,29(3):912-914.
[3]Maquire D J,Goodchild M F,Rhind D W.Geographic In-formation Systems:Principles and Applications[M].Lon-don:Longman,1991:361-374.
[4]Pullar D.MapScript:a map algebra programming languageincorporating neighborhood analysis[J].GeoInformatica,2001,5(2):145-163.
[5]Ostlander N.Interoperable services for web-based spatial de-cision support[C]//7th AGILE Conference on GeographicInformation Science.Heraklion:AGILE,2004.
[6]Mennis J,Viger R,Tomlin C D.Cubic map algebra func-tions for spatio-temporal analysis[J].Cartography and GIS,2005,32(1):17-32.
[7]Frank A.Map algebra extended with functors fortemporal data[C]//Perspectives in Conceptual Modeling:ER2005 Work-shops,Klagenfurt:Springer,2005.
[8]Gilberto C,Danilo P,Ricardo C.Towards a generalized mapalgebra:principles and data types[C]//Proceedings ofGeoinfo 2005,2005:66-81.
[9]Amara Y,Marsault X.A GPU tile-load-map architecture forterrain rendering:theory and applications[J].Vis Comput,2009,25:805-824.
[10]Ruiz A,Ujaldon M,Cooper L,etal.Non-rigid registrationfor large sets of microscopic images on graphics processors[J].Sign Process Syst,2009,55:229-250.
[11]Nistor A,Chin WN,Tan TS,et al.Optimizing the paral-lel computation of linear recurrences using compact matrixrepresentations[J].Paralle Distrib Comput,2009,69:373-381.
[12]Liu W G,Schmidt B,Voss G,et al.Accelerating molecu-lar dynamicssimulations using graphics processing units withCUDA[J].Computer Physics Communications,2008,179:634-641.
[2]吴连贵,易瑜,李肯立.基于CUDA的地震数据相干体并行算法[J].计算机应用,2009,29(3):912-914.
[3]Maquire D J,Goodchild M F,Rhind D W.Geographic In-formation Systems:Principles and Applications[M].Lon-don:Longman,1991:361-374.
[4]Pullar D.MapScript:a map algebra programming languageincorporating neighborhood analysis[J].GeoInformatica,2001,5(2):145-163.
[5]Ostlander N.Interoperable services for web-based spatial de-cision support[C]//7th AGILE Conference on GeographicInformation Science.Heraklion:AGILE,2004.
[6]Mennis J,Viger R,Tomlin C D.Cubic map algebra func-tions for spatio-temporal analysis[J].Cartography and GIS,2005,32(1):17-32.
[7]Frank A.Map algebra extended with functors fortemporal data[C]//Perspectives in Conceptual Modeling:ER2005 Work-shops,Klagenfurt:Springer,2005.
[8]Gilberto C,Danilo P,Ricardo C.Towards a generalized mapalgebra:principles and data types[C]//Proceedings ofGeoinfo 2005,2005:66-81.
[9]Amara Y,Marsault X.A GPU tile-load-map architecture forterrain rendering:theory and applications[J].Vis Comput,2009,25:805-824.
[10]Ruiz A,Ujaldon M,Cooper L,etal.Non-rigid registrationfor large sets of microscopic images on graphics processors[J].Sign Process Syst,2009,55:229-250.
[11]Nistor A,Chin WN,Tan TS,et al.Optimizing the paral-lel computation of linear recurrences using compact matrixrepresentations[J].Paralle Distrib Comput,2009,69:373-381.
[12]Liu W G,Schmidt B,Voss G,et al.Accelerating molecu-lar dynamicssimulations using graphics processing units withCUDA[J].Computer Physics Communications,2008,179:634-641.
版权所有：© 2023 中国地质图书馆 中国地质调查局地学文献中心