Abstract
The authors present a novel modification of the HYB format — known from the CUSP library. The new format is suitable for sparse Markovian transition rate matrices and enables processing two times bigger matrices on single GPU, also improving computation performance at the same time. Particularly, the SpMV operation — that is the multiplication of a sparse matrix by a vector — is analyzed for this format on one GPU and two GPUs. Numerical experiments for transition rate matrices of Markov chains from [18] show that the proposed format allows to process matrices of sizes about \(3.6 \times 10^7\) rows with the use of single GPU (3 GB RAM). When the plain HYB format is used the matrices of these sizes do not fit in one GPUs memory. Moreover, the use of the modified HYB format can give the speedup even up to 13 times in comparison to multi-threaded CPU (12 cores).
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 subscriptionsReferences
Bell, N., Garland, M.: Efficient Sparse Matrix-Vector Multiplication on CUDA. NVIDIA Technical Report No. NVR-2008-004 (2008)
Bianchi, G.: Performance analysis of the IEEE 802.11 distributed coordination function. IEEE J. Sel. Areas Commun. 18, 535–547 (2000)
Bustamam, A., Burrage, K., Hamilton, N.A.: Fast parallel Markov clustering in bioinformatics using massively parallel computing on GPU with CUDA and ELLPACK-R sparse format. IEEE/ACM Trans. Comput. Biol. Bioinform. 9, 679–692 (2012)
Bylina, B., Bylina, J., Karwacki, M.: Computational aspects of GPU-accelerated sparse matrix-vector multiplication for solving Markov models. Theor. Appl. Inform. 23, 127–145 (2011)
Bylina, B., Karwacki, M., Bylina, J.: A CPU-GPU hybrid approach to the uniformization method for solving Markovian models – a case study of a wireless network. In: Kwiecień, A., Gaj, P., Stera, P. (eds.) CN 2012. CCIS, vol. 291, pp. 401–410. Springer, Heidelberg (2012)
Bylina, B., Karwacki, M., Bylina, J.: Multi-GPU implementation of the uniformization method for solving Markov models. In: Proceedings of Federated Conference on Computer Science and Information Systems (FedCSIS), pp. 533–537 (2012)
Bylina, J., Bylina, B.: A Markovian queuing model of a WLAN node. In: Kwiecień, A., Gaj, P., Stera, P. (eds.) CN 2011. CCIS, vol. 160, pp. 80–86. Springer, Heidelberg (2011)
Bylina, J., Bylina, B., Karwacki, M.: A Markovian model of a network of two wireless devices. In: Kwiecień, A., Gaj, P., Stera, P. (eds.) CN 2012. CCIS, vol. 291, pp. 411–420. Springer, Heidelberg (2012)
Fernandes, P., Plateau, B., Stewart, W.J.: Efficient descriptor-vector multiplication in stochastic automata networks. J. ACM 45, 381–414 (1998)
Philippe, B., Saad, Y., Stewart, W.J.: Numerical methods in Markov chain modelling. Oper. Res. 40, 1156–1179 (1992)
Stewart, W.J.: Introduction to the Numerical Solution of Markov Chains. Princeton University Press, Princeton (1994)
Wijs, A.J., Bošnački, D.: Improving GPU sparse matrix-vector multiplication for probabilistic model checking. In: Donaldson, A., Parker, D. (eds.) SPIN 2012. LNCS, vol. 7385, pp. 98–116. Springer, Heidelberg (2012)
ELLPACK. http://www.cs.purdue.edu/ellpack/
Intel Math Kernel Library. http://software.intel.com/en-us/articles/intel-mkl/
Stewart, W.J.: MARCA\_Models: a collection of Markov chain models. http://www4.ncsu.edu/~billy/MARCA_Models/MARCA_Models.html
Thrust. http://code.google.com/p/thrust/
Acknowledgments
This work was partially supported within the project N N516 479640 of the Ministry of Science and Higher Education (MNiSW) of the Polish Republic “Modele dynamiki transmisji, sterowania zatłoczeniem i jakością usług w Internecie”.
Author information
Authors and Affiliations
Corresponding author
Editor information
Editors and Affiliations
Rights and permissions
Copyright information
© 2014 Springer-Verlag Berlin Heidelberg
About this paper
Cite this paper
Bylina, J., Bylina, B., Karwacki, M. (2014). An Efficient Representation on GPU for Transition Rate Matrices for Markov Chains. In: Wyrzykowski, R., Dongarra, J., Karczewski, K., Waśniewski, J. (eds) Parallel Processing and Applied Mathematics. PPAM 2013. Lecture Notes in Computer Science(), vol 8384. Springer, Berlin, Heidelberg. https://doi.org/10.1007/978-3-642-55224-3_62
Download citation
DOI: https://doi.org/10.1007/978-3-642-55224-3_62
Published:
Publisher Name: Springer, Berlin, Heidelberg
Print ISBN: 978-3-642-55223-6
Online ISBN: 978-3-642-55224-3
eBook Packages: Computer ScienceComputer Science (R0)