RE: st: increasing time burden during resampling

["Feiveson, Alan H. (JSC-SK311)" <alan.h.feiveson@nasa.gov>]

Vince - 

Thanks for looking at this problem. Actually, there are some matrices that I form before the iterations that I would like to preserve. Instead of "matrix drop _all", is there some easily referenced group of matrices (not my own) that should be dropped? I would rather avoid saving those original matrices in a data set if possible.

Al Feiveosn

-----Original Message-----
From: owner-statalist@hsphsun2.harvard.edu [mailto:owner-statalist@hsphsun2.harvard.edu] On Behalf Of Vince Wiggins, StataCorp
Sent: Wednesday, December 30, 2009 2:46 PM
To: statalist@hsphsun2.harvard.edu
Subject: Re: st: increasing time burden during resampling

In a continuing thread, Al Feiveson <alan.h.feiveson@nasa.gov> has written
that repeated invocations of -test- or -lincom- with factor variables causes
Stata to slow down on commands that subsequently use factor variables.

We asked for Al's data yesterday and have been able to reproduce the problem.

We will fix this soon (within the next few weeks) in an update.

The speed decrease is negligible unless you issue many hundreds or thousands
of -test- or -lincom- statements.  This means that only those doing
simulations or other heavily iterated processes need to be concerned.

As Al noted, the speed decrease grows approximately linearly and without
bound.  You can reclaim the original speed in any Stata session by typing
-matrix drop _all-.  Unless Al is collecting results in a matrix, this is a
safe short-term workaround for the problem.


-- Vince 
   vwiggins@stata.com

 
*
*   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/