Stata The Stata listserver
[Date Prev][Date Next][Thread Prev][Thread Next][Date index][Thread index]

Re: st: multithreading in Stata?


From   rossini@blindglobe.net (A.J. Rossini)
To   statalist@hsphsun2.harvard.edu
Subject   Re: st: multithreading in Stata?
Date   21 Oct 2002 20:25:23 -0700

>>>>> "stas" == Stas Kolenikov <skolenik@email.unc.edu> writes:

    stas> (This is probably a more technical question to support@, but it might be
    stas> of general interest, too.) Can Stata be tuned for multiprocessor /
    stas> multithread environment? My interest is inspired by an internal discussion
    stas> inside my department about the use of Beowulf cluster that the department
    stas> has access to. As long as Stata is an interpretative environment, I
    stas> suppose there is rather little one could do to multithread it. Am I
    stas> bluntly mistaken?

Yes.

Python, for example, is multithreaded (in a weird sense; so,
technically, it might not be what you mean, but it sort of is).  Perl
might be; it's been a while since I've had to touch it.  Some Lisp
interpreters are.

Another example would be the work of myself, Michael Li, and Luke
Tierney on simple approaches for parallel statistical computations,
using R.  It uses sockets, PVM, or MPI as backends. 

Work on threading R has been in progress for eons.  It might be, one
of these days.  

Again, this latter work should be replicable in theory by stata corp,
and wouldn't really require threading (thread-safe code isn't simple,
though perhaps, given the exceptional nature of Stata, it already
is). 

    stas> A clear example of where it would be beneficial is the Monte Carlo /
    stas> bootstrap situations. As long as both are implemented as ado files rather
    stas> than internal commands, there would be no benefits from trying to run it
    stas> in a multiprocessor environment, I am afraid :(.

Not really -- it is a matter of whether you can spawn processes or
threads.  If you can spawn one or the other, you are set.

(they aren't the same, but if there is enough computation to be done,
for obviously parallelizable statistical computation, they are pretty
much the same).

best,
-tony

-- 
A.J. Rossini				Rsrch. Asst. Prof. of Biostatistics
U. of Washington Biostatistics		rossini@u.washington.edu	
FHCRC/SCHARP/HIV Vaccine Trials Net	rossini@scharp.org
-------------- http://software.biostat.washington.edu/ ----------------
FHCRC: M: 206-667-7025 (fax=4812)|Voicemail is pretty sketchy/use Email
UW:   Th: 206-543-1044 (fax=3286)|Change last 4 digits of phone to FAX
(my tuesday/wednesday/friday locations are completely unpredictable.)



*
*   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/



© Copyright 1996–2014 StataCorp LP   |   Terms of use   |   Privacy   |   Contact us   |   What's new   |   Site index