[fpu] More rounding questions...

[snemiro@fi.uba.ar]

Hi.. As a Final Project for University , I'm implementing a FPU core (IEEE 
754 compatible) in VHDL, and I don't know the following:

1. If a result is a subnormal number, must I turn both Underflow and 
Inexact bit on??

2. If a result is a subnormal, must I round it? I dont have the 
standard...I assume that I will loss precision if I round it...

I was taknig a look at some results into this vector tests, but I can't get 
the same results...



ftp://win-
ftp.uia.ac.be/pub/cant/IeeeCC754/Logfiles/Hardware/AMD/multd

TIA

sergio
--
To unsubscribe from fpu mailing list please visit http://www.opencores.org/mailinglists.shtml