Re: [fpu] Rounding of Floating Point Numbers

[Jamil Khatib <jamilkhatib75@yahoo.com>]

as I know there are 4 types of rounding in the IEEE
std,
Round toward nearest, Round to zero, Round to +inf,
Round to -inf

All of them must be implemented the last three for
overflow, under flow results, zero in integer
conversion "As I think".
But round to nearest used in normal numbers.

In fact I am not sure exactly what is the meaning of
round to nearest even, but I think this should be
described in teh standard, else each FPU will produce
different result. "I do not have the std and did not
find any information how to round.


Regards
Jamil Khatib

--- Rudolf Usselmann <russelmann@hotmail.com> wrote:
> 
> Does anyone have a description or can explain
> to me how to round properly and efficiently ?
> This would be for IEEE-754 single precision
> floating point numbers (actually intermediate
> results from adder of multiplier, before nor-
> malizing ...). Also suggestions as to which
> rounding mode is most universal[better] (I'm
> tending sofar towards nearest even ...).
> 
> Thanks,
> rudi
>
________________________________________________________________________
> Get Your Private, Free E-mail from MSN Hotmail at
> http://www.hotmail.com
> 


__________________________________________________
Do You Yahoo!?
Kick off your party with Yahoo! Invites.
http://invites.yahoo.com/