矿山三维空间数据距离直方图算法优化及加速

Optimization and acceleration of distance histogram algorithm of three-dimensionspace data of coal mine

  • 摘要: 分析了三维空间数据距离直方图算法的性质及数据结构,提出了基于图形处理器的通用计算方法和基于FPGA的高性能计算方法,基于图形处理器的计算方法可用于实现三维空间数据距离直方图算法的单指令多数据并行优化;基于FPGA的计算方法可实现算法的硬件分块优化,使算法的硬件结构达到最优匹配。实验结果表明,利用基于图形处理器的计算方法可使算法达到平均18倍的性能加速,基于FPGA的计算方法可使算法达到平均30倍的性能加速,大大提升了算法的数据处理能力。

     

    Abstract: The properties and data structure of distance histogram algorithm of three-dimensional space data were analyzed, and a general computing method based on graphics processor and a high performance computing method based on FPGA were proposed. The calculation method based on graphics processor is used to implement single instruction multiple data parallel optimization of distance histogram algorithm of three-dimensional space data; the calculation method based on FPGA can realize hardware block optimization, and achieve the optimal matching of the hardware structure of the algorithm. The experimental results show that using the calculation method based on graphics processor can make the algorithm reach performance acceleration with an average of 18 times, and the computing method based on FPGA can make the algorithm reach performance acceleration with an average of 30 times, which greatly improve data processing ability of the algorithm.

     

/

返回文章
返回