🚀 go-pugleaf

in article <922@aber-cs.UUCP>, pcg@aber-cs.UUCP (Piercarlo Grandi) says:
> Summary: 68030 == 68020+10/15%

> You get me a 68030 and I tell you :->... From various papers and benchmarks
> on the few 68030 machines around (Next, Macs, Suns), it seems that one
> statement I have read (68030 == 68020+10/15%) is quite reasonable overall.

Apparently, if you do nothing more than drop a 68030 into an existing 68020
system, this is almost exactly what you get.  The Mac II vs. IIx is a good
example of this situation.  However, once you design a real 68030 system,
it's much more like (68030 = 68020+50/100%).  Especially in systems that
use the Morotola MMU.  A 68030 system can access memory exactly twice as fast
as the 68020/68851 combo, at the same clock speed.  It also has a cache line
"burst" fetch mode that allows much more efficient operation between an '030
and a medium speed memory system.  NeXT, for instance, uses this burst mode,
and gets decent performance out of slow memory.  They could be accessing
memory around 50% faster than they are, though -- a machine in the price range
of the NeXT should have an external cache.  Some decent examples of tight
'030 systems are the Apollo 4500 and the Hp 9000/370.  They ain't cheap, but
they're examples of what the CPU can do.  Now only if the bean counters would
let us '030 PC designers build 'em like the Workstation guys get to....

> Piercarlo "Peter" Grandi           | ARPA: pcg%cs.aber.ac.uk@nsfnet-relay.ac.uk
> Dept of CS, UCW Aberystwyth        | UUCP: ...!mcvax!ukc!aber-cs!pcg
> Penglais, Aberystwyth SY23 3BZ, UK | INET: pcg@cs.aber.ac.uk
-- 
Dave Haynie  "The 32 Bit Guy"     Commodore-Amiga  "The Crew That Never Rests"
   {uunet|pyramid|rutgers}!cbmvax!daveh      PLINK: D-DAVE H     BIX: hazy
              Amiga -- It's not just a job, it's an obsession