st: Setting same seed, getting different random numbers.

[Patrick Roland <patrick.rolande@gmail.com>]

I encountered a strange problem. I tried running the same code on two
different versions of Stata - one Stata/SE 12.0 on a server, and one
Stata/SE 11.2 on my own computer. Both generate normal random numbers
in mata with the same seed. But after a while the random numbers they
generate differ! The code is as follows

set seed 2
mata
f = rnormal(1000000,1,0,1)

On the two machines, f[1,1] = .7005116833. But one one f[1000000,1] =
-.5182349644, and on the other it is .7083750352.

This strikes me as bizarre and impedes replication of my code across
computers. Any ideas as to what's going on here?

Thanks.
*
*   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/