AT&T 3B1 Unix PC

Eric J Korpela korpela at ssl.berkeley.edu
Thu Jun 19 19:37:05 CDT 2008


On Thu, Jun 19, 2008 at 3:50 PM, Eric Smith <eric at brouhaha.com> wrote:

> Eric wrote about the AT&T Unix PC (7300):
> > Biggest problem was how slow it was at floating
> > point, probably due to a crappy math library.
>
> Was it significantly slower than other contemporary 68K machines
> that used software floating point?
>

What I had to compare it to at the time was a Sun 100u and a Tandy 16.  It
was comparable to the tandy 16 and significantly slower than the Sun 100u,
especially at transendentals.  It was also significantly slower than a 7.16
MHz V30 (also with software floating point).  Of couse the PC world was a
far more competitive place for C compilers with everyone saying theirs was
the fastest.  I've often wondered if the AT&T and Tandy compilers used
floating point routines provided by Motorola rather than developing their
own.

Eric



More information about the cctalk mailing list