Re: Complexity of 'eigs'

[c.]

On 13 Aug 2012, at 12:01, Søren Hauberg wrote:

> Thanks, that helped quite a bit, actually. I am fine with having the 
> complexity of each iteration, which I understood to be O(D^3) if the matrix 
> is dense DxD.

If your matrix is dense I am not sure using "eigs" rather than "eig" is 
convenient, the deirect algorithm in "eig" should be O(D^2):
http://web.eecs.utk.edu/~dongarra/etemplates/node93.html#sec:dense

> Thanks
> Søren
c.