Fortran - Re: double precision difference between matlab and fortran 90

You are viewing an archived Text-only version of the thread. To view this thread in it's original format and/or if you want to reply to this thread please [click here]

Author

Re: double precision difference between matlab and fortran 90

louisrom1@gmail.com

2008-03-17, 8:18 am

On Mar 14, 5:08 pm, glen herrmannsfeldt <g...@ugcs.caltech.edu> wrote:
> James Van Buskirk wrote:
>
> There are a number of ways they can come out different.
> One is that constant expressions are often done at compile
> time, and sometimes with a different result than the same
> expression at run time. Another is the rounding mode of
> the floating point processor.
>
> Fortran doesn't say much about rounding of floating point,
> but the IEEE standard does. Also possible is the
> difference between 53 bit double and 64 bit temporary
> real for x87.
>
> It seems that you are getting the two possible rounded
> results from the subtraction, rounded up or down.
>
> -- glen

Is there a way to control the way Matlab and Compaq visual fortran
round their numbers?

Sponsored Links