Dear all
I am stymied by the following conversation with Stata:
====================
. xtgls depvar $indepvar, p(h)
matsize too small - should be at least 7864
r(908);
. set matsize 8000
Current memory allocation
current memory usage
settable value description (1M = 1024k)
--------------------------------------------------------------------
set maxvar 5000 max. variables allowed 1.733M
set memory 256M max. data space 256.000M
set matsize 8000 max. RHS vars in models 488.953M
-----------
746.685M
. xtgls depvar $indepvar, p(h)
op. sys. refuses to provide memory
r(909);
======================
Herewith some more details:
0. I have Stata 8.2 running on a Windows XP box with 1024M of ram.
1. The data set is a mere 135 MB which seems "small" relative to the >1GB limits discussed in the FAQ at http://www.stata.com/support/faqs/win/winmemory.html <http://www.stata.com/support/faqs/win/winmemory.html> .
2. I have roughly 8000 (unbalanced) panels with gaps, max. panel length<=30.
3. I read the bit in the manual entry on xtgls ([XT]p.95) which points to xtreg or xtgee when panels are many. I have run xtreg, pa c(uns) r, but in my application, the errors could well be correlated across panels.
4. I have tried out the bit of advice about typing "discard" and then trying to re-run the xtgls command again, but to no avail.
The FAQ seems to say that the lack of contiguous memory blocks business is not likely to be an issue with Stata 8, so I'm curious why I get the r(909) msg. Is it that
a. I am making a programming mistake?
b. the continguous memory space problem affects Stata 8 as well?
c. I'm making a logic error in that xtgls, p(h) is infeasible when id >> t?
If the answer is (a) or (b), I hope some kind soul can help me.
I would also like to know if the hotfix posted on the Stata website for Stata 9 will work for Stata 8 also? Or must I upgrade to Stata 9 to make this work? [I have 9 sitting in my office but would much prefer not to have to take the time to install it right now.]
If the answer is (b), that's not a Stata question, but can someone point me to where I might be able to read up on what to do next in such a setting (i.e. in cases where id much outnumbers t)?
As always, all pointers will be much appreciated.
Thanks!
Rajib Doogar,
Department of Accountancy,
The University of Illinois at Urbana-Champaign
1206 S. Sixth Street, Champaign, IL 61820
Ph: 217.244.8083, Fax: 217.244.0902
http://www.cba.uiuc.edu/doogar
*
* For searches and help try:
* http://www.stata.com/support/faqs/res/findit.html
* http://www.stata.com/support/statalist/faq
* http://www.ats.ucla.edu/stat/stata/
