You have 4 Gb of RAM, but it's typical that Windows only provides
about half to Stata. You need more RAM, i.e. a hardware change, to run
precisely this model. Otherwise you try using fewer groups or shutting
other applications (which probably won't make much difference).
Nick
On Sun, Dec 9, 2012 at 10:28 AM, Murod Aliyev <ma.statalist@yahoo.com> wrote:
> I have read a lot of threads on this issue, but still can not resolve my problem:
>
> I am trying to run FGLS (xtgls) with 49,000 observations, 8 years unbalanced panel with around 7,500 groups, 26 varibles in the model.
> After first trying xtgls I get a message saying "matsize too small - should be at least 7369". So, I set matsize to 8000, then I get a message r(909) "op. sys. refuses to provide memory". Could you please help me understand what is causing the problem here, and how to reslve it? I presume it is not related to the number of observations or variables, but rather caused by the way FGLS works.
>
> Resources in use:
>
> - 32 bit Windows 7
> - RAM 4g
> - Stata/SE 12.1
> - Stata "About" shows Available Physical Memory 1,945,452 KB
> - max_memory set to .
>
> Memory report is as follows:
>
> . memory
> Memory usage
> used allocated
> ---------------------------------------------------------------------
> data (incl. buffers) 35,809,764 67,108,864
> var. names, %fmts, ... 32,668 52,140
> overhead 1,064,972 1,065,172
> Stata matrices 0 0
> ado-files 114,691 114,691
> saved results 65,115 65,115
> Mata matrices 34,884 34,884
> Mata functions 14,040 14,040
> set maxvar usage 1,391,728 1,391,728
> other 2,687 2,687
> ---------------------------------------------------------------------
> total 38,449,909 69,849,321
>
*
* For searches and help try:
* http://www.stata.com/help.cgi?search
* http://www.stata.com/support/faqs/resources/statalist-faq/
* http://www.ats.ucla.edu/stat/stata/
