st: efficiency (or lack thereof) of the colon operator

[Kit Baum <baum@bc.edu>]

The speed differential has been reported for Intel and AMD. It is not  
limited to that platform:

(1) on a fairly old PowerPC G4 chip,

. mata: x=uniform(10000,1)
r; t=0.04 17:48:27

. mata: for(i=1;i<=10000;i++) y=x*1
r; t=5.75 17:48:52

. mata: for(i=1;i<=10000;i++) y= x:*1
r; t=10.32 17:49:25


(2) On a newer PowerPC G5,

. mata: x=uniform(10000,1)
r; t=0.01 17:51:55

. mata: for(i=1;i<=10000;i++) y=x*1
r; t=2.33 17:52:06

. mata: for(i=1;i<=10000;i++) y= x:*1
r; t=5.89 17:52:18


and
(3) on a quite old Sun SPARC processor, the colon operator is  
realllllly slow--

. mata: x=uniform(10000,1)
r; t=0.08 17:50:28

. mata: for(i=1;i<=10000;i++) y=x*1
r; t=6.49 17:50:39

. mata: for(i=1;i<=10000;i++) y= x:*1
r; t=26.44 17:51:10



Kit Baum, Boston College Economics
http://ideas.repec.org/e/pba1.html


*
*   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/