Path: utzoo!utgpu!jarvis.csri.toronto.edu!mailrus!tut.cis.ohio-state.edu!rutgers!bellcore!texbell!killer!elg
From: elg@killer.Dallas.TX.US (Eric Green)
Newsgroups: comp.sys.cbm
Subject: Re: C128,C128D, and C128C??? Questions
Message-ID: <7606@killer.Dallas.TX.US>
Date: 22 Mar 89 01:42:42 GMT
References: <89Mar20.194245est.2737@godzilla.eecg.toronto.edu>
Organization: The Unix(R) Connection, Dallas, Texas
Lines: 49

in article <89Mar20.194245est.2737@godzilla.eecg.toronto.edu>, leblanc@eecg.toronto.edu (Marcel LeBlanc) says:
> 
> 2Mhz 6502 is (almost) 4 times faster than a 2Mhz Z80.  This is because
> memory (or IO) accesses on a Z80 take 3-4 cycles, whereas the same operation
> on a 6502 takes only one cycle (yes, the 6502 was an early RISC :-) ).  So
> even if you could get the Z80 to run flat out at 4MHz, it would still be
> only a little more than half the speed of the 8502 at 2MHz.  But then, I
> guess CP/M can use all the speed it can get...

Actually, it's not that simple. Yes, memory access takes 4 cycles on
the Z-80, making actual data fetch of a 4mhz Z-80 about the same as a
1mhz 6502. But the Z-80 has 16-bit registers, and has a LOT of
registers (at least, when compared to having a whole 3 8-bit
registers!) This means that you don't have to go to memory as often,
and can more easily do 16-bit arithmetic and do 16-bit indirect
addressing.  This is one reason you'll find better, e.g., Pascal
compilers on a CP/M machine, than on a C-64. I've used Turbo Pascal on
a CP/M machine, and it makes any Pascal on the 64 look sad, insofar as
code size & speed is concerned. Same with BDS "C" vs. Power C.

I don't have a Z-80 manual lying around (I'm not ancient enough ;-),
but I also suspect that a 4mhz Z-80 takes less time for internal
processing than a 1mhz 6502. e.g. moving B to A may require less time
than moving .X to .A on 6502. 

I explored the 128's CP/M mode some time ago, right after getting my
1750 (it wasn't worth it with a 1571 and a 1541 as the only two of my
drives that it'd access). There are text editors under CP/M that make
anything under 128 or 64 mode look pitiful. Not to mention that you
can probably still get pirate copies of DBASE and WORDSTAR ;-).
Unfortunately, screen update is MUCH too slow for it to be useful.
There seems to have been a design decision, somewhere in the software
cycle, to switch into 6502 mode and use the Kernal for all I/O. I
suspect this is responsible for the drastic slow-down... if the Z-80
was handling all its own screen and modem i/o, that'd solve the two
biggest problems of the 128's CP/M mode.  (and, BTW, there is a
program available to turn off the 40 column screen -- it does speed up
compute-bound programs, but does little to speed up slow I/O
operations).

I solved the screen i/o problem by buying an Amiga ;-). But if anybody
out there has any CP/M experience, I'd be quite interested in seeing
the results of a hacked CP/M that did its own screen i/o instead of
punting to the 6502....

--
|    // Eric Lee Green              P.O. Box 92191, Lafayette, LA 70509     |
|   //  ..!{ames,decwrl,mit-eddie,osu-cis}!killer!elg     (318)989-9849     |
| \X/            Amiga.  The homestation for the blessed of us.             |