st: rewriting ML's d2 functions in Mata

["Zurab Sajaia" <zsajaia@hotmail.com>]

Hi everybody,


I have several estimation programs written that ar implemented as a d2 
functions for ML. Although having analytical second derivatives makes whole 
estimation process a lot faster, for more complicated models (but still 
useful) d2 gets really long and ugly, and as a result slow.

I was thinking if there were a chance to make programs more efficient by 
rewriting parts of it in Mata. So was wondering if anybody had experience of 
doing that. Since there are not that many matrix operations but instead lot 
of generate/replace. So will there be a enough gain to justify rewriting 
long programs?

Thanks in advance,
Zurab


*
*   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/