RE: [fpu] Rounding of Floating Point Numbers

["Rudolf Usselmann" <russelmann@hotmail.com>]

>-----Original Message-----
>From: owner-fpu@opencores.org [mailto:owner-fpu@opencores.org]On Behalf
>Of Jamil Khatib
>
>
>Try to read a paper called "Lecture Notes on teh
>status of IEEE standard 754 for Binary Floating-Point
>Arithmetic" by Prof. W. Kahan.
>
>There is a link for it in the FPU pages
>
>There is a section about rounding,
>
>There are three round to nearest approaches 1) Round to nearest even
>2) round half-integers away from 0 3) truncate to integers towards 0
>
>I suggest that the rounding mode should be selected by
>a control word in the FPU

I don't know what this "Lecture Notes" are about, and don't understand
what a "half-integer" is. Let me reiterate again what the IEEE-754 spec.
says about rounding (again this is from the actual specification from
IEEE-754 that lies in from of me right now):

1) Round to Nearest [DEFAULT]
2) Directed Rounding [USER SELECTABLE]
   a) Round to +INF
   b) Round to -INF
   c) Round to Zero

Per specification, the default has to be rounding to the nearest, and the
other three user selectable.

rudi

________________________________________________________________________
Get Your Private, Free E-mail from MSN Hotmail at http://www.hotmail.com