> Hello,
> thanks for your answer;
> the problem I see with your approach is that on machines with just 2
> different kinds one could face problems, when one tries to compile
> subroutines for the extended precisions.
> So one has to introduce some kind of conditional compilation.
Not at all. An ordinary compilation will do.
> Best regards
> Alois
>
> PS.: I gave up on trying to think about the problems mentioned by Pierre
> Hugonnet. From his latest posts it seems he wants intermediate
> precisions.
ÿ
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
|