GPU Based Spot Noise Parallel Algorithm for 2D Vector Field Visualization
Dept. of Comput. Sci., Ocean Univ. of China, Qingdao, China
International Conference on Optoelectronics and Image Processing (ICOIP), 2010
@conference{qin2010gpu,
title={GPU Based Spot Noise Parallel Algorithm for 2D Vector Field Visualization},
author={Qin, B. and Su, F. and Wu, Z. and Wang, J.},
booktitle={2010 International Conference on Optoelectronics and Image Processing},
pages={580–583},
year={2010},
organization={IEEE}
}
Graphic Processing Unit (GPU) has involved into a parallel computation for it’s massively multi threaded architecture. Due to its high computational power, GPU has been used to deal with many problems that can be easily parallelized. This paper will present a GPU based spot noise parallel algorithm for 2D vector field visualization. It uses spot noise method with GPU resources and compute unified device architecture (CUDA) to visualize 2D vector field. Vector field are partitioned to multiple thread so that a large number of data are processed simultaneously. Fast on-chip shared memory is used on GPU to optimize the performance and a data transformation mechanism between host and device is presented. The parallel algorithm applies these strategies to a 2D velocity field and obtains up to 16X speedup compared with conventional sequential computation. It is suitable for interactive applications and in-time remote visualization of vector fields.
March 28, 2011 by hgpu
