Abstract
Matrix eigenvalue theory has become an important analysis tool in scientific computing. Sometimes, people do not need to find all eigenvalues but only the maximum eigenvalue. Existing algorithms of finding the maximum eigenvalue of matrices are implemented sequentially. With the increasing of the orders of matrices, the workload of calculation is getting heavier. Therefore, traditional sequential methods are unable to meet the need of fast calculation for large matrices. This paper proposes a parallel algorithm named PA-ST to find the maximum eigenvalue of positive matrices by using similarity transformation which is implemented by CUDA (Computer Unified Device Architecture) on GPU (Graphic Process Unit). To the best of our knowledge, this is the first CUDA based parallel algorithm of calculating maximum eigenvalue of matrices. In order to improve the performance, optimization techniques are applied in this paper such as using the shared memory rather than the global memory to improve the speed of computation, avoiding bank conflicts by setting the span index, satisfying the principle of coalesced memory access, and by using single-precision floating-point arithmetic and the pinned memory to reduce the copy operation and obtain higher data transfer bandwidth between the host and the GPU device. The experimental results show that the similarity transformation technique can significantly shorten the running time compared to the sequential algorithm and the speedup ratio is nearly stable when the number of iterations increases. As the matrix order increases, the running time of the sequential algorithm and PA-ST increases correspondingly. Experiments also show that the speedup ratio of the PA-ST is between 2.85 and 35.028.
This is a preview of subscription content, log in via an institution.
Buying options
Tax calculation will be finalised at checkout
Purchases are for personal use only
Learn about institutional subscriptionsPreview
Unable to display preview. Download preview PDF.
References
Wang, T., Guo, L., Li, G., Li, J., Wang, R., Ren, M., He, J.S.: Implementing the jacobi algorithm for solving eigenvalues of symmetric matrices with cuda. In: NAS, pp. 69–78 (2012)
Tsourakakis, C.E.: Fast counting of triangles in large real networks without counting: Algorithms and laws. In: ICDM, pp. 608–617 (2008)
Gaidhane, V.H., Hote, Y.V., Singh, V.: Article: A new approach for estimation of eigenvalues of images. International Journal of Computer Applications 26(9), 1–6 (2011)
Griffiths, J.W.R.: Adaptive array processing: A tutorial. IEE Proceedings F Communications, Radar and Signal Processing 130(1), 3–10 (1983)
Luo, X., Lin, J., Wu, W.: A prediction-correction dynamic method for large-scale generalized eigenvalue problems. Abstract and Applied Analysis 2013(SI), 1–8 (2013)
Hall, C., Porsching, T.: Computing the maximal eigenvalue and eigenvector of a positive matrix. SIAM J. Numer. Anal. 5(2), 269–274 (1968)
Oepomo, T.: Survey of power, qr, and oepomos iterative methods for solution of largest eigenvalue of essentially positive matrices. International Journal of Management Science and Engineering Management 4(1), 3–19 (2009)
Faddeev, D., Faddeeva, V.: Computational Methods of Linear Algebra. W. H. Freeman and Company (1973)
Wilkinson, J.: Convergence of lr, qr and related algorithms. Comp. Jour. 8(1), 77–84 (1966)
Oepomo, T.: A contribution to collatzs eigenvalue inclusion theorem for nonnegative irreducible matrices. ELA 10(1), 31–45 (2003)
NVIDIA Corporation: NVIDIA CUDA Programming Guide version 4.2 (2012), http://developer.download.nvidia.com/compute/DevZone/docs/
Garland, M., Grand, S.L., Nickolls, J., Anderson, J., Hardwick, J., Morton, S., Phillips, E., Zhang, Y., Volkov, V.: Parallel computing experiences with cuda. IEEE Micro 28(4), 13–27 (2008)
Spampinato, D.G., Elster, A.C.: Linear optimization on modern gpus. In: IPDPS, pp. 1–8 (2009)
Yeh, L.: Inequalities for the maximal eigenvalue of a nonnegative matrix. Glasgow Mathematical Journal 39(3), 276–284 (1997)
Harris, M.: Optimizing parallel reduction in cuda. Technical report, NVIDIA Developer Technology Website/projects/reduction/doc/reduction.pdf (2007), http://developer.download.nvidia.com/compute/cuda/1_1/
Author information
Authors and Affiliations
Editor information
Editors and Affiliations
Rights and permissions
Copyright information
© 2014 Springer International Publishing Switzerland
About this paper
Cite this paper
Tian, N., Guo, L., Ai, C., Ren, M., Li, J. (2014). GPU Acceleration of Finding Maximum Eigenvalue of Positive Matrices. In: Sun, Xh., et al. Algorithms and Architectures for Parallel Processing. ICA3PP 2014. Lecture Notes in Computer Science, vol 8631. Springer, Cham. https://doi.org/10.1007/978-3-319-11194-0_18
Download citation
DOI: https://doi.org/10.1007/978-3-319-11194-0_18
Publisher Name: Springer, Cham
Print ISBN: 978-3-319-11193-3
Online ISBN: 978-3-319-11194-0
eBook Packages: Computer ScienceComputer Science (R0)