Path: utzoo!utgpu!jarvis.csri.toronto.edu!cs.utexas.edu!tut.cis.ohio-state.edu!ucsd!sdcc6!sdcc4!ln63wkp
From: ln63wkp@sdcc4.ucsd.edu (Viet Ho)
Newsgroups: comp.sys.amiga.tech
Subject: Floating point format
Keywords: ffp, amiga floats format
Message-ID: <8590@sdcc6.ucsd.edu>
Date: 11 Mar 90 22:14:54 GMT
Sender: news@sdcc6.ucsd.edu
Lines: 18


I need to know the binary storage format of amiga floating
point numbers (single precision 32bits) for a program I'm
trying to simulate for class.   According to Jake Commander's
AMIGA Assembly Language Programming, the layout is :

   8bits for the exponent stored in two's complement
   24 bits for the mantissa w/ MSB as the sign bit.
    
Is this correct?   Is there a format like that of the PDP-11
or AT&T 3B2's where the MSB of a 32 bit word is the sign
of the mantissa, followed by 8 bits bias exponent, and 23 bits
for a normalized mantissa with the MSB (always a 1) not stored.
???

          Email responses to    vho@ucsd.edu 
  thanks!
                                            -Viet Ho