基于向量化的扰动引力矢量快速并行算法

doi:10.13278/j.cnki.jjuese.20200066

吉林大学学报(地球科学版) ›› 2021, Vol. 51 ›› Issue (6): 1863-1871.doi: 10.13278/j.cnki.jjuese.20200066

基于向量化的扰动引力矢量快速并行算法

黄炎¹, 王庆宾¹, 李国强², 冯进凯¹, 谭勖立¹

1. 信息工程大学地理空间信息学院, 郑州 450001;
2. 信息工程大学指挥系, 郑州 450001

收稿日期:2020-03-19 出版日期:2021-11-26 发布日期:2021-11-24
作者简介:黄炎(1994-),男,博士研究生,主要从事物理大地测量方面的研究,E-mail:781367531@qq.com
基金资助:
国家自然科学基金项目（41574020）；信息工程大学校自立课题（2105070232）

Fast Parallel Algorithm of Disturbing Gravity Vector Based on Vectorization

Huang Yan¹, Wang Qingbin¹, Li Guoqiang², Feng Jinkai¹, Tan Xuli¹

1. Institute of Geospatial Information, Information Engineering University, Zhengzhou 450001, China;
2. The Command Department, Information Engineering University, Zhengzhou 450001, China

Received:2020-03-19 Online:2021-11-26 Published:2021-11-24
Supported by:
Supported by the National Natural Science Foundation of China (41574020) and the Independent Project of Information Engineering University (2105070232)

摘要/Abstract

摘要： 无人飞行器在飞行过程中受地球外部扰动引力影响，对其精确控制需要计算飞行轨迹点扰动引力。为有效恢复地球外部引力场，常采用地球重力场位系数模型进行计算，但其计算耗时随模型阶数的升高而呈指数增加。本文提出一种扰动引力矢量计算的向量化方法，推导出利用地球重力场模型计算扰动引力三分量的向量化表达公式，并利用CUDA（compute unified device architecture）异构并行算法进行并行化以达到快速计算单点扰动引力的目的。通过仿真实验可知，本文所提方法可有效减少单点扰动引力计算耗时，利用跨阶次递推法进行单点计算加速比可达8以上，最高可达13.20；利用Belikov递推法进行单点计算加速比可达6以上，最高可达8.99。

关键词: 向量化, 并行计算, 扰动引力, 地球重力场位系数模型, 加速比

Abstract: Unmanned aerial vehicle (UAV) is affected by the disturbing gravity of the Earth during its flight. To control the UAV accurately, it is necessary to calculate the disturbing gravity of the flight trajectory points. In order to effectively restore the external gravitational field of the Earth, the Earth gravitational model is often used, but its calculation time increases exponentially with the increase of the order of the model. In this paper, an improved vectorization method for the calculation of perturbation gravity is proposed, and a CUDA(compute unified device architecture)heterogeneous parallel algorithm is used for parallelization to achieve the purpose of rapid calculation of single point perturbation gravity. The simulation results show that the proposed method can effectively reduce the time-consuming of single-point calculation of perturbation gravity:The acceleration ratio of single-point calculation can be more than eight times and the highest can be up to 13.20 times by using the step-by-step recursion method; By using Belikov recursion method, the acceleration ratio of single point calculation can be more than six times and the maximum can be 8.99 times.

Key words: vectorization, parallel computing, disturbing gravity, Earth gravitational model, acceleration ratio

中图分类号:

P221

黄炎, 王庆宾, 李国强, 冯进凯, 谭勖立. 基于向量化的扰动引力矢量快速并行算法[J]. 吉林大学学报(地球科学版), 2021, 51(6): 1863-1871.

Huang Yan, Wang Qingbin, Li Guoqiang, Feng Jinkai, Tan Xuli. Fast Parallel Algorithm of Disturbing Gravity Vector Based on Vectorization[J]. Journal of Jilin University(Earth Science Edition), 2021, 51(6): 1863-1871.

参考文献

[1] 范昊鹏,李姗姗.局部区域模型重力异常快速算法研究[J].大地测量与地球动力学,2013,33(6):28-30. Fan Haopeng, Li Shanshan. Study on a Fast Algorithm for Model Gravity Anomalies in Local Areas[J]. Journal of Geodesy and Geodynamics, 2013,33(6):28-30.
[2] Sandwell D T, Müller R D, Smith W H F, et al. New Global Marine Gravity Model from CryoSat-2 and Jason-1 Reveals Buried Tectonic Structure[J]. Science, 2014, 346:65-67.
[3] 冯进凯,王庆宾,黄炎,等. 基于局部重力场建模的Tikhonov正则化点质量核径向基函数方法[J].吉林大学学报(地球科学版),2019, 49(2):569-577. Feng Jinkai, Wang Qingbin, Huang Yan, et al. Point-Mass Kernel RBF Model Based on Tikhonov Regularization[J]. Journal of Jilin University (Earth Science Edition), 2019, 49(2):569-577.
[4] Khosro Ghobadi-Far, Mohammad Ali Sharifi, Nico Sneeuw. 2D Fourier Series Representation of Gravitational Functionals in Spherical Coordinates[J]. Journal of Geodesy, 2016, 90(9):871-881.
[5] 王庆宾,马国元,王永收,等.全球空间扰动引力快速FFT计算效能分析[C]//国家安全地球物理.西安:西安地图出版社,2015:55-59. Wang Qingbin, Ma Guoyuan, Wang Yongshou, et al. Analysis of Fast Computational Efficiency of Global Disturbing Gravity[C]//National Security Geophysics. Xi'an:Xi'an Map Press, 2015:55-59.
[6] 王昱. 扰动引力的快速计算及其对落点偏差的影响[D].长沙:国防科学技术大学,2002. Wang Yu. Fast Calculation of Perturbation Gravity and Its Effect on Landing Point Deviation[D].Changsha:National University of Defense Science and Technology, 2002
[7] 李红伟.地球外部空间扰动引力并行计算[J].指挥控制与仿真,2013,35(3):100-103. Li Hongwei. Parallel Method for the Disturbing Gravity of the Outside Space[J]. Command Control & Simulation, 2013, 35(3):100-103.
[8] 黄佳喜.扰动重力场快速(并行)计算方法研究[D].郑州:信息工程大学,2017. Huang Jiaxi. Research on Fast (Parallel) Computing of Disturbing Gravity[D]. Zhengzhou:Information Engineering University, 2017.
[9] René Forsberg. The Use of Spectral Techniques in Gravity Field Modelling:Trends and Perspectives[J]. Physics and Chemistry of the Earth,1998, 23(1):31-39.
[10] 周蓓,黄永忠,许瑾晨,等. 向量数学库的向量化方法研究[J].计算机科学,2019,46(1):320-324. Zhou Bei, Huang Yongzhong, Xu Jinchen, et al. Study on SIMD Method of Vector Math Library[J]. Cumputer Science,2019, 46(1):320-324.
[11] Gene D R, Amdahl M. Validity of the Single Processor Approach to Achieving Large Scale Computing Capabilities[Z]. 1967.
[12] 刘文志.并行算法设计与性能优化[M].北京:机械工业出版社,2015. Liu Wenzhi. Parallel Computing and Performance Optimization[M]. Beijing:Machinery Industry Press, 2015.
[13] Yu Zhihui. High Performance Parallel Programming with MPI[M]. Beijing:Tsinghua University Press, 2001.
[14] Nvidia. Nvidia Tesla P100-the Most Advanced Datacenter Accelerator Ever Built Featuring Pascal GP100, the World's Fastest GPU[Z]. Nvidia Whitepaper,2016, v01.1(Nvidia WP-08019-001).
[15] Girish Sharma, Abhishek Agarwala, Baidurya Bhattacharya. A Fast Parallel Gauss Jordan Algorithm for Matrix Inversion Using CUDA[J]. Computers & Structures,2013, 128:31-37.
[16] Christian Hirt, Sten Claessens, Thomas Fecher, et al. New Ultrahigh-Resolution Picture of Earth's Gravity Field[J]. Geophysical Research Letters,2013, 40(16):4279-4283.
[17] 欧阳明达,张敏利,于亮.采用Belikov列推和跨阶次递推方法计算超高阶缔合勒让德函数[J].测绘工程,2017,26(7):12-15. Ouyang Mingda, Zhang Minli, Yu Liang. Calculating the Ultra-High-Qrder Associated Legendre Functions by Belikov Column Method and Recursive Method Between Every Other Order and Degree[J]. Engineering of Surveying and Mapping, 2017, 26(7):12-15.
[18] Christopher J, Jong K L, Jay H K. On the Computation and Approximation of Ultra-High-Degree Spherical Harmonic Series[J]. Journal of Geodesy,2007, 81(9):603-615
[19] 吴星,刘雁雨. 多种超高阶次缔合勒让德函数计算方法的比较[J]. 测绘科学技术学报,2006(3):188-191. Wu Xing, Liu Yanyu. Comparison of Computing Methods of the Ultra-High Degree and Order[J]. Journal of Geomatics Science and Technology,2006(3):188-191.
[20] 范宏涛,郭春喜,王小华,等.超高阶重力场模型EIGEN-6C2适应性分析[J].测绘科学,2015,40(9):18-22. Fan Hongtao,Guo Chunxi,Wang Xiaohua,et al. Adaptability Analysis on Ultra-High-Gravity Model EIGEN-6C2 in Chinese Coastal Zone[J]. Science of Surveying and Mapping, 2015, 40(9):18-22.

Metrics

Viewed

Full text

Abstract

Cited

Shared

Discussed

基于向量化的扰动引力矢量快速并行算法

Fast Parallel Algorithm of Disturbing Gravity Vector Based on Vectorization

PDF (PC)

摘要/Abstract

引用本文

使用本文

参考文献

相关文章 3

Metrics

本文评价

推荐阅读 10

[1]	韩利, 韩立国, 李翔, 王德利, 崔杰. 二阶声波方程频域PML边界条件及频域变网格步长并行计算[J]. J4, 2011, 41(4): 1226-1232.
[2]	潘宏勋，方伍宝. 基于PC机群波动方程叠前深度偏移的并行计算策略[J]. J4, 2008, 38(4): 708-0712.
[3]	殷文,印兴耀, 张繁昌. 基于并行遗传算法的地震属性优化研究[J]. J4, 2005, 35(05): 672-676.