For the purpose of common mid⁃point bin attribute analysis from high density exploration and three⁃dimensional seismic ge⁃ometry,this paper analyzes the calculation characteristics of bin analysis,and proposes an efficient parallel algorithm for bin attribute a⁃nalysis based on Open Computing Language ( OpenCL) heterogeneous programming model. The algorithm divides bin grid into several sub⁃regions under hardware memory constraint.Tasks of sub⁃regions are executed with coarse⁃grained parallelization by assigning task to different devices,whereas pairs of source⁃receiver in sub⁃region are executed with fine⁃grained parallelization. With HP's Z820 desktop workstation,experiment was carried out with high density three⁃dimensional geometry from seismic exploration.The experimental results show that the parallel algorithm takes full advantage of multiprocessor's parallel processing capability,and the processing speed is about 30 times faster than that on CPU⁃based implementation ( single core) under the condition that the precision is not changed.%针对高密度地震勘探下的三维观测系统海量面元属性计算需求,基于OpenCL并行编程模型,研究并实现了适用于异构环境下面元属性的并行计算方法。根据面元分析算法特点将整个网格分成多个子区域,子区域可由不同设备以粗粒度方式并行,而子区域内部以炮检对方式细粒度并行。采用HP Z820工作站,在CPU+GPU异构混合平台下的试算结果表明,异构多核处理速度是CPU(单核)的30倍以上,数据生成速率高于300 MB/s。
展开▼