Would the model development team like to double check these numbers?

[Patrick LEE]

Originally Posted by mchandler

It speeds up the GCM dramatically, but we are getting spurious results with the GCM, which are related to excess cloud creation. The problem appears to be related to a problem with the RANDU function and the underscore (and the compiler options that should allow us to ignore the underscores).

Hi Mark,

I forgot to say that the RANDU function is only for the cumulus cloud cover. You need to consider the large-scale clouds as well. Hence, I am NOT so sure that RANDU is the problem. May be you can post your results, so that others can have a better idea on the problem.

Patrick.