BSD for AT?

[Paul De Bra]

In article <1445 at leah.Albany.Edu> rds95 at leah.Albany.Edu (Robert Seals) writes:
>The Sun 386i with SunOS 4.0 is an example of an Intel architecture
>running a BSD-derived system*. Unfortunately, it's somewhat proprietary
>(the call of capitalism has won you, Bill joy (;). There are certain
>problems running Unix on chips <= 80286, which have been detailed elsewhere.
>But now that I have this nice 20MHz 386 box on my desk, I want 4.3!!!
>
Sun could easily come out with SunOS 4.0 for a basic 386 box, since they
did some of the development for the Sun 386i on Compacs (mostly because
they didn't have the Sun 386i yet while they were writing the software).
But they don't want to do that of course, because they want to sell the
386i.

Now, I ran a couple of tests on a 16Mhz 6386 (not exactly a fast box one
could say, no cache or anything...) with plain AT&T Unix and on the
25Mhz Sun 386i (with cache and everything). The following small table
will make it very clear that although the Sun hardware is  indeed about
twice as fast as thhe  6386, the  Unix is SLOW.

test                Sun 386i/250      AT&T 6386

LOW-LEVEL
250000 getpid()       16.8             32.5
2500000 func-calls     5.7             10.5
sieve                  3.6              6.5
100000 sines           3.1              8.6
loop 100000000         9.9             11.2

HIGH-LEVEL
pipe 5Mbytes          13.8             14.3
some shell script      2.5              2.9
8 scripts &           18.7             16.7
15 scripts &          35.4             31.3

I think I'll wait until Sun gets more out of the 386 before trying to
switch.

Paul.
-- 
------------------------------------------------------
|debra at research.att.com   | uunet!research!debra     |
------------------------------------------------------