[R] SVD for a 500, 000* 500, 000 matrix (singular value decomposition)

[Uwe Ligges]

kayj wrote:
> Hi,
> 
> 
> I tried to run SVD on a  500,000* 500,000 matrix and i get a message that it
> can  not allocate a vector of length 270 mb 


Well, you will obviously need >> 1Tera(!)bytes of RAM just in order to 
store the matrix (or is it some sparse one?). I wonder how you managed 
that issue.

Uwe Ligges


> doe snayone know how to solve this problem? any ideas on other softwares
> where I can do this? 
> I appreciate your help
> thanks