How to get the eigenvalues ​​of a huge matrix (size: 2x10 ^ 5)

Question

I have a 200,000 x 200,000 matrix. I need to find my own values ​​for this. I have used matlab so far, but since the size of the matrix does not lend itself to MATLAB, I have moved to perl and now even perl is unable to handle this huge matrix, it says from memory. I would like to know if I can find out the eigenvalues ​​of this matrix using some other programming language that can handle such huge data. Elements are not zeros in the main, so there is no way to go to a sparse matrix. Please help me in resolving this issue.

Answer 1

I think you're lucky anyway with MATLAB. Take a look at their distributed computing tools . You will need some kind of parallel environment, a computing cluster.

If you don't have a compute cluster, you can look at the distributed methods for calculating eigenvalues ​​/ vectors that could be used on Amazon EC2 or similar .

It also discusses parallel eigenvalue methods here , which may direct you to better libraries and programming approaches than Perl.