Re: st: mata error: unable to allocate --- memory issue?

[László Sándor <sandorl@gmail.com>]

Oh, sorry, also:

Could we get some rough guidance of what the trade-off is (what the
trade-offs are...) with -mata: mata set matafavor speed-? I am trying
to figure out whether I could keep more covariates if I wait longer
for results. But how much longer? Or maybe the memory benefit is
small, it won't save my 10K x 10K matrices?

I know I could always experiment, but maybe there are more general
lessons here where we could all benefit from.

Thanks!

On Fri, May 25, 2012 at 12:54 PM, László Sándor <sandorl@gmail.com> wrote:
>
> Hi,
>
> I am running my own Mata code under Stata 11 MP for Windows. An
> attempt of mine ran out of memory, but on second attempt I got an
> "unable to allocate..." error. Is this still a memory issue (basically
> the same thing), or something else?
>
> FYI/FWIW, some explanation: It is most likely an overkill what I did
> with allowing for factor variables where they had hundreds of values.
> This can and did result in trying to handle matrices of size tens of
> thousands by tens of thousands. So if it is a hopeless memory issue, I
> am happy to backtrack and give up these specific factor variables. But
> as my code passed the first hurdle, I was wondering if the memory
> issue is solved now and the "unable to allocate" is something else...
>
> Thanks!
>
> Laszlo
>
> *
> *   For searches and help try:
> *   http://www.stata.com/help.cgi?search
> *   http://www.stata.com/support/statalist/faq
> *   http://www.ats.ucla.edu/stat/stata/

*
*   For searches and help try:
*   http://www.stata.com/help.cgi?search
*   http://www.stata.com/support/statalist/faq
*   http://www.ats.ucla.edu/stat/stata/