 |
[Date Prev][Date Next][Thread Prev][Thread Next][Date index][Thread index]
st: Upgrading from v9 to v10mp
Dear Statalist,
There's several reasons why I'd like to upgrade from Stata 9 to
Stata 10. But I have a question about MP and whether this is
something I should be looking at (as I have dual core CPU).
From the info I've found (http://www.stata.com/statamp/) MP is
suitable and *should* be faster for most analyses. The same info
also states that not all processes will run in parallel, however.
It would seem silly for me to pay extra only to find that the most
computationally intensive work I do (which isn't a great deal)
cannot make use of this extra processing power. So I was wondering
whether anyone on Statalist might be able to offer some thoughts?
Probably the most intensive work involves such things as xtlogit
and I've also looked at reoprob - with a large survey and
bootstrapping would MP work through user written commands (i.e.
reoprob) any faster? Other things that can take time include
cycling through large -foreach num- lists and using postfile
(neither reoprob nor postfile are covered in report.pdf at http://
www.stata.com/statamp/) to generate a new dataset. Would these be
noticeably quicker under MP?
What's my MP criteria? I suppose shaving 1 or 2 seconds of
something that takes 1 or 2 seconds wouldn't be that useful, I just
don't get another sip of coffee. Similarly, shaving a few hours off
something that takes more 24 hours to run probably wouldn't mean
that much as I'd be doing something else anyway. It's the 'not
quite long enough to go do something useful' tasks where I'd find
the extra speed most useful... are the time savings linear over
problem size?
Any thoughts greatly appreciated
Simon
Since desktop computers will eventually all have multiple cores,
eventually all serious statistical environments will too. Why would
you want to only take advantage on one CPU? It's only a question now
of cost. In the transition, companies will charge a premium to take
advantage of multiple CPUs, because they can.
I think all survey commands are parallel. This would be a good reason
to upgrade if you work with large surveys.
In my circumstance I need to perform genome scans with categorical
dependent variables and nested data structure, so I have decided to
buy MP; $900 EDU License to obtain the MP upgrade and all
documentation seems an excellent deal, compared to the $900 grant
dollars I spend on other equipment and services.
--
David C. Airey, Ph.D.
Research Assistant Professor
*
* 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/