Improving the Performance of Tensor Matrix Vector Multiplication in Cumulative Reaction Probability Based Quantum Chemistry Codes
Cumulative reaction probability (CRP) calculations provide a viable computational approach to estimate reaction rate coefficients. However, in order to give meaningful results these calculations should be done in many dimensions (ten to fifteen). This makes CRP codes memory intensive. For this reason, these codes use iterative methods to solve the linear systems, where a good fraction of the execution time is spent on matrix-vector multiplication. In this paper, we discuss the tensor product form of applying the system operator on a vector. This approach shows much better performance and provides huge savings in memory as compared to the explicit sparse representation of the system matrix.
Unable to display preview. Download preview PDF.
- 1.Wyatt, R.E., Zhang, J.Z.H.: Dynamics of molecules and chemical reactions. CRC Press, Boca Raton (1996)Google Scholar
- 2.Manthe, U., Miller, W.H.: The cumulative reactions probability as eigenvalue problem. J. Chem. Phys., pp. 3411–3419 (1999)Google Scholar
- 4.Gropp, W.D., Kaushik, D.K., Keyes, D.E., Smith, B.F.: Toward realistic performance bounds for implicit CFD codes. In: Keyes, D., Ecer, A., Periaux, J., Satofuka, N., Fox, P. (eds.) Proceedings of Parallel CFD 1999, pp. 233–240. Elsevier, Amsterdam (1999)Google Scholar
- 7.Temam, O., Jalby, W.: Characterizing the behavior of sparse algorithms on caches. In: Proceedings of Supercomputing 1992, pp. 578–587. IEEE Computer Society, Los Alamitos (1992)Google Scholar
- 8.White, J., Sadayappan, P.: On improving the performance of sparse matrix-vector multiplication. In: Proceedings of the 4th International Conference on High Performance Computing (HiPC 1997), pp. 578–587. IEEE Computer Society, Los Alamitos (1997)Google Scholar
- 9.Balay, S., Buschelman, K.R., Gropp, W.D., Kaushik, D.K., Knepley, M.G., McInnes, L.C., Smith, B.F.: PETSc home page (2002), http://www.mcs.anl.gov/petsc
- 10.McCalpin, J.D.: STREAM: Sustainable memory bandwidth in high performance computers. Technical report, University of Virginia (1995), http://www.cs.virginia.edu/stream