Application of high performance computing platforms to tomographic particle image velocimetry

Authors

  • V.A. Lozhkin
  • Yu.A. Lozhkin
  • M.P. Tokarev

Keywords:

tomography
particle image velocimetry
Tomo PIV
GPU
computing cluster
OpenCL
MPI
high performance computing

Abstract

Tomographic particle image velocimetry (Tomo PIV) is a new method to study gas and liquid flow. Tomo PIV allows measuring the three-dimensional flow characteristics in the measurement volume. One of the main problems of this measurement method is the high computing resource demands to its data processing algorithms. Several approaches to increase the computational performance of this method using high performance computing platforms, such as graphic processing units (GPU) and compute clusters, are discussed. Some problems connected porting the tomographic reconstruction and correlation algorithms onto many-core computing platforms and the proposed solutions are also discussed. The practical results of porting the tomographic reconstruction algorithm to GPU using OpenCL technology are described. The time estimates obtained experimentally for data processing are given for various computing platforms.

New computational technologies

Downloads

Published

2012-03-19

Issue

Vol. 13 (2012): Issue 1.

Section

Section 2. Programming

Author Biographies

V.A. Lozhkin

Kutateladze Institute of Thermophysics of SB RAS
• Assistant

Yu.A. Lozhkin

Kutateladze Institute of Thermophysics of SB RAS
• Engineer

M.P. Tokarev

Kutateladze Institute of Thermophysics of SB RAS
• Researcher

References

  1. Elsinga G.E., Scarano F., Wieneke B., van Oudheusden B.W. Tomographic particle image velocimetry // Experiments in Fluids. 2006. 41. 933-947.
  2. Stanislas M., Okamoto K., Kahler C.J., Westerweel J. Main results of the third international PIV challenge // Experiments in Fluids. 2008. 45. 27-71.
  3. Adrian R.J. Twenty years of particle image velocimetry // Experiments in Fluids. 2005. 39. 159-169.
  4. Бильский А.В., Ложкин В.А., Маркович Д.М., Токарев М.П., Шестаков М.В. Оптимизация и тестирование томографического метода измерения скорости в объеме потока // Теплофизика и аэромеханика. 2011. 18, № 4. 1-12.
  5. Atkinson C., Soria J. An efficient simultaneous reconstruction technique for tomographic particle image velocimetry // Experiments in Fluids. 2009. 47. 553-568.
  6. Бильский А.В., Маркович Д.М., Токарев М.П. Адаптивные алгоритмы обработки изображений частиц для расчета мгновенных полей скорости // Вычислительные технологии. 2007. 12, № 3. 109-131.
  7. Информационно-вычислительный центр НГУ. Описание комплекса (http://nusc.ru).
  8. Jang B., Kaeli D., Do S., Pien H. Multi GPU implementation of iterative tomographic reconstruction algorithms // Proc. IEEE Int. Symposium on Biomedical Imaging. Boston, 2009. 185-188.
  9. Xu F., Mueller K. Towards a unified framework for rapid 3D computed tomography on commodity GPUs // Nuclear Science Symposium Conference Record. 2003. 4. 2757-2759.
  10. Maar S., Batenburg K., Sijbers J. Experiences with Cell-BE and GPU for tomography // Proc. of the 9th Int. Workshop on Embedded Computer Systems: Architectures, Modeling, and Simulation. Berlin: Springer, 2009. 298-307.
  11. NVIDIA Corporation. CUDA C Programming Guide Version 4, 2011 (http://nvidia.com).
  12. Khronos OpenCL Working Group. The OpenCL Specification Version 1.1, 2011 (http://www.khronos.org/opencl/).