(j3.2006) (SC22WG5.4104) [ukfortran] Standard intrinsics and coarrays
Thu Oct 22 00:33:20 EDT 2009
On 10/21/2009 10:57 AM, N.M. Maclaren wrote:
> It is processor dependent whether all images use a common
> generator or whether each image uses a separate one.
What exactly does this mean? I get the feeling, but I believe we need to
be more specific. The description of RANDOM_SEED is pretty specific
about what is guaranteed, what is processor dependent, etc. For example,
"When the values differ, the use of either value as the PUT argument in
a subsequent call to RANDOM SEED shall result in the same sequence of
pseudorandom numbers being generated."
For the case of "separate" generators, it is clear, for the most part,
since it follows the existing rules independently on each image. But we
need to say more. What if all images supply the same PUT argument to
RANDOM_SEED. Are they guaranteed to get the same sequence if they call
RANDOM_NUMBER in unison? Or not? If the generators are truly
independent, why should there be any restriction on ordering between
RANDOM_SEED and RANDOM_NUMBER?
In the case of a "shared" generator, I am not sure I understand exactly
what is meant, guaranteed, or processor dependent. Is it obvious to
others and I am just being dense?
Aleksandar Donev, Ph.D.
Luis W. Alvarez Postdoctoral Fellow
Center for Computational Sciences and Engineering (https://ccse.lbl.gov)
Lawrence Berkeley National Laboratory (http://www.lbl.gov)
E-mail: adonev at lbl.gov
Phone: (510) 486-5782 Fax: (510) 486-6900
Address: MS 50A-1148, LBL, 1 Cyclotron Rd., Berkeley, CA 94720
More information about the J3