On Thursday, September 19, 2002, at 04:06 PM, Richard Fateman wrote:
> It is entirely possible that the 68020 etc uses "extended"
> precision (using 64 bit fractions instead of 53 bits) for
> floating point registers. It would be surprising though if
> this explained how answers differed in the last 8 decimal places.
In addition to the standard 32-bit and 64-bit types, the 68020/030/040
all have an extended precision floating format of 96 bits. 16 bit for the
exponent, 16 bits of unused space, and then 64-bits for the mantissa. It
is quite possible that the double format on his machine is in fact the
extended format. A number of old Macintosh compilers in fact did this. By
extrapolation, I would guess that the 68060 has a similar extended
floating point format.
----------------------------------------------------------------------
Michael W. Martin Phone: (281) 333-2177
Draper Laboratory FAX: (281) 333-5276
2200 Space Park Dr. EMail: martin@simba.jsc.draper.com
Houston, TX 77058 WWW: http://www.jsc.draper.com/
USA Mail Code: EG/Draper
----------------------------------------------------------------------