solving large (53000 equations) systems of linear ODEs with Mma

*To*: mathgroup at yoda.physics.unc.edu*Subject*: solving large (53000 equations) systems of linear ODEs with Mma*From*: Roque Donizete de Oliveira <oliveria at engin.umich.edu>*Date*: Wed, 25 Nov 92 14:29:50 -0500

I have a systems of 53000 linear ODEs: P'(t) = A P(t) where P(t) is a vector of length 53000, and A is a very large matrix (constant) with only 5-diagonal of non-zero elements. I've solved this problem using standard Runge-Kutta (in C and Fortran) and using the Krylov approximation method for the matrix exponential operator. I'd like to do the same thing (i.e., solve the system above) using Mma, if possible. So I ask you fellow Mma users: Has anyone used Mma for this purpose ? Can Mma solve (numerically, of course) such a large number of linear ODEs ? Thanks for any information. Roque oliveria at engin.umich.edu Nuclear Eng Dept, The U. of Michigan