[MITgcm-devel] portable random number generator

[Andrea Molod]

On Sun, 18 Jul 2004, Ed Hill wrote:
> 
> Would it make sense to add full check-pointing so that both the
> "identical initial conditions" and the "pickup" cases can be handled
> gracefully?  And perhaps we could make it a new package named "random"? 

hi ed,
sorry for not being clear maybe. as i see it, there is only one 'case' and
no conflict between repeating what is already there and handling the
restart from a pickup. the change i suggest adds to this is the ability
to handle the second 'case', and i was recommending that further changes
(to flt) may be needed to get it to be reproducible.

do you have an objection to the change i am suggesting for the meantime?

as for the other random number generator, you told me once about it and
i am sure that eventually i would like to go and get it and use it. for
now, however, i would prefer to just use (for the fizhi development) what
is there now - for expediency, nothing else.

thanks,

andrea

------------------------------------------------------------------
Andrea M. Molod, PhD.
Department of Earth, Atmospheric & Planetary Sciences 
MIT
Tel: (617) 253-5458, Email: amolod at mit.edu 
WWW:http://paoc.mit.edu/paoc/people/person.asp?position=Postdoc&who=molod
----------------------------------------------------------------