On Tue, Dec 09, 2008 at 11:54:15PM +0100, Jakob Malm wrote:Hi Jakob
> I am doing the following with a large (256^2 x 300, sometimes sparse)
> matrix, X:
> [ U, S, P ] = svd ( X, 0 );
> I am only interested in S and P. Can I skip the U somehow, so that it
> does not get assigned, something like
> [ ..., S, P ] = svd ( X, 0 );
> The reason for wanting to do this is because U becomes the same size as
> X, and takes up memory.
This is not possible I think, the best solution for you would be to
reassign U to a smaller value immediately after calculating svd, that
should free up your memory.