🚀 go-pugleaf

in article <747@cps3xx.UUCP>, usenet@cps3xx.UUCP (Usenet file owner) says:
> In-reply-to: 128a-3aj@e260-3b.berkeley.edu's message of 30 Sep 88 22:24:09 GMT

>   A 68020 (arguably) does more work each cycle than an 80286.  Also,
> something else to watch is the memory access time.  With high-speed
> memory prices going up, putting a 30-MHz 68030 in your machine will
> cost a LOT--not just the processor and main memory, but putting a
> reasonable cache (say, 32-64K) of really high-speed memory in to avoid
> wait states (disclaimer: I haven't seen a 68030 hardware manual, so I
> don't know its memory rate/cycle time ratio).

The '030 can run sustained memory cycles at two clocks per cycle, and burst
memory cycles at one clock per cycle for up to three cycles.  Apparently none
of these features are used in the IIx, but from a chips alone point of view,
an '030 can more than double the speed of the '020-'851 combination at the
same clock rate.

Another thing to consider when comparing 680x0 system to 80x86 systems is ".L".
Or, put another way, WE get longword instructions, they don't under MS-DOS or
OS/2.  So a '386 machine running at 25MHz or so is still running only 16 bit
instructions.  That 32 bit bus isn't going to help all that much.  All your
68000 code is already full of 32 bit wide instructions, so all 680x0 system get
an immediate benefit from 32 bit wide memory.

-- 
Dave Haynie  "The 32 Bit Guy"     Commodore-Amiga  "The Crew That Never Rests"
   {ihnp4|uunet|rutgers}!cbmvax!daveh      PLINK: D-DAVE H     BIX: hazy
		"I can't relax, 'cause I'm a Boinger!"