Bookmark and Share

Notice: On April 23, 2014, Statalist moved from an email list to a forum, based at statalist.org.


[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

st: RE: Use of stata and maximisation of one's computer capabilities


From   "Mak, Timothy" <[email protected]>
To   "'[email protected]'" <[email protected]>
Subject   st: RE: Use of stata and maximisation of one's computer capabilities
Date   Thu, 6 Oct 2011 15:10:48 +0000

Hi Francisco, 

You said you're using Stata/SE. If you're not using Stata/MP, I don't suppose it can take advantage of the multiple cores that you have in your computers. Stata's report http://www.stata.com/statamp/statamp.pdf suggests the command -mlogit- is 75% parrallelizable. Apparently, with 4 cores, you can expect a 2.3 times increase in speed and with 8 cores, a 2.9 times increase in speed. 

Tim

-----Original Message-----
From: [email protected] [mailto:[email protected]] On Behalf Of Francisco Rowe
Sent: 06 October 2011 08:10
To: [email protected]
Subject: st: Use of stata and maximisation of one's computer capabilities

Hi,

I am running a set of models in a multinomial logit model framework, using 4,000,000 observations, 3 choices (plus the reference category) and 54 variables. Stata takes 40-60 mins, running each the model.

I ran my code in two different machines:

1) A Win 2008 R2 terminal server (2x quad core/16HT cores, Xeon E5540 @2.53GHz, 32GB RAM, 500+GB free disk space) -using Stata/SE 10.0

2) A Mac 2.66 GHz Intel Core i7, 8 GB 1067 MHz -using Stata/SE 10.1

Despite of these machines' capabilities, my models take too long. I have set the memory usage to:

On the Win server:

                   current                                 memory usage
   settable          value     description                 (1M = 1024k)
   --------------------------------------------------------------------
   set maxvar        32767     max. variables allowed          12.501M
   set memory         1024M    max. data space              1,024.000M
   set matsize       4500     max. RHS vars in models        154.873M
                                                           -----------
                                                            1,800.169M

On mac:
                   current                                 memory usage
   settable          value     description                 (1M = 1024k)
   --------------------------------------------------------------------
   set maxvar        30000     max. variables allowed          12.390M
   set memory         1024M    max. data space              1,024.000M
   set matsize       10000     max. RHS vars in models        763.779M
                                                           -----------
                                                            1,800.169M 

Is there any way to take advantage of the full capabilities of these machines? Do anyone have any suggestions on an "more" efficient way (in time terms) running my models using Stata?

Regards,

Francisco.
*
*   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/


© Copyright 1996–2018 StataCorp LLC   |   Terms of use   |   Privacy   |   Contact us   |   Site index