Problems in Modernizing a IIgs, con't

[bjblackbear@ibm.net (B.J. Major)]

Scott G. said:

>I see.  Well, if you regard a chip as fast as the Pentium in the
>same light, then it is not an insult to the GS.  :)

Mind you the 65816 has traveled a MUCH shorter distance from the 6502 
than even the 80286 traveled from the 8086. The difference between 65816
and 6502 is more like the difference between 8080 and 8086.

Also, consider that one of the WORST features of the P6 is its backwards
compatibility. That's backwards compatibility spelled "Backwards. 
Compatibility?". You would really want to lumber yourself with a similar
abomination?

Intel's future is either die (at least as far as major CPU vendor status
goes), or quietly go RISC and provide emulation software/firmware as the
only token at backwards compatibility. Intel inside instantly equates to
"low-end" when workstations are under discussion.

>>The guy who wrote the original message [to B.J.] said it was pin
>>compatible! I can only work with the information I'm given, you
>>know. It wasn't an assumption, it was information received.
>
>That guy was just quoting rumors, then.  After all, we would not

He did claim to have rare schematics and other technical information,
but anyway that's not the point.

>>>but the 65C816 is much more efficient at handling memory.  It can
>>>access memory in a single cycle whereas the 68000 needed 4 cycles. 
>>as the entire meat of the program? That's the only situation where
>>your observation will have a real-world effect. I stand by what I
>>said.

>What I forgot to say was that the 65C816 can also complete most
>instructions in a single cycle.  The 68000 could not.  That's where
>the 65C816 would come out ahead.  :)

Nope, not even vaguely (have you ever programmed the 680x0?). Consider 
the addressing modes offered on the two CPUs and consider how many 
discrete instructions you would need to use on the 65xxx to perform the
equivalent of, say, move.l (a0,d0.w),(a1,d2.l). Or even movem.l 
d0-d7/a0-a6,-(sp). The 680x0's thruput is substantially higher, clock
for clock. The 680x0's instruction set is also far more orthogonal than
the 65xxx's, which makes it easier to write optimizing compilers.

Also consider that the 65xxx series uses zero page as basically an extra
register store, and all indirect addressing modes go through pointers in
ZP. This is torture for a multitasking OS, having to juggle pointers 
through ZP and maintain ZP contexts for each process.

>For the record, I'm saying this about the 68000.  Obviously, with
>all the money Apple poured into supporting Macs and therefore
>Motorolla, by the time the '030 came out, this fault was fixed.

Absolutely NOT. Apple was never anything more than a minor consumer of 
680x0 chips. 68000s and 68010s in particular are used widely in embedded
systems and non-Mac workstations, for example: video arcade games and 
home systems (Sega Genesis, Commodore-Amiga 1000, 500, 600, Sinclair QL
(actually a 68008), Atari ST/STFM/STe, and more arcade boards than you 
can count on the fingers of a fairly large crowd, from Data East, Japan
Capsule Computers [Capcom], Taito, Atari, Konami, Sega and more), laser
printers, intelligent serial interfaces, multiplexers, buffered SCSI 
controllers, modems (ZyXEL is one vendor that used a 68000, NetComm even
today uses the MC68302 which is a 68020 derivative, and Motorola's own 
modems use the same family of MCUs).

If there was any large consumer market driving the engineering of the 
680x0 range, it sure wasn't Apple machines. Probably about half of all 
coin-operated video arcade game boards made between 1988-1992 used one
or two 68000s, and that's an incredibly huge market.

Even now, Apple is only "a" major consumer of PowerPCs. It is probably 
the second or third largest consumer of workstation-grade components,
but workstation CPUs are only a small part of the PowerPC family. The
largest consumers are people selling by-the-carat black boxes based
around the PowerPC variants intended for embedded applications.

>Then again, had Apple used the same money to support the GS and
>therefore Western Design Center, the derived 65C816 chip would
>continue to have the advantage.

What you're saying here is that if the product had been developed, it 
would have been developed, but since it wasn't, it hasn't been. Yes, one
does equal one.

>was based on the 65C02.  The Pentium did it, so can a derived
>65C816 _had Western Design Center had more support_ from big
>computer manufacturers such as Apple.

Look: To be blunt, who cares what CPU was in the final result? You're 
chasing your tail with a semantic argument here. EVERY time a 
manufacturer chooses to switch parts, somebody gets alienated. I'm sure
there are people who complain bitterly about unleaded gasoline. Even the
IIgs is not absolutely one hundred per cent compatible with an Apple II.
Heck, there is even incompatibility between different ROM and
motherboard revisions amongst the "traditional" (non-GS) II family. The
incompatibility between Mac and GS is just more extreme.

>the market.  Now you might point out that modern PowerPC based
>Apples can emulate even the GS.  That's true, but the early Macs
>such as the SE could not run GS software!  Apple alienated most

Apple continued to support the II in its educational niches, support 
didn't die just like THAT. There were relatively few GSs sold, anyway. 
And there were/are II emulation hardware modules for LC-PDS machines.

>Like you said, it _was_ possible for Apple to upgrade the GS and
>Western Design Center to upgrade the 65C816.  At least to the

And like I didn't say, but as Apple clearly recognized, it was
pointless, because instead of spending millions over-developing a
maverick chip which nobody else except a couple of console vendors was
using, Apple could simply go to Motorola and pick up a ready-made
industry-standard accepted MPU with a large variety of support chips,
extremely high volume (=low price), superior performance to current
65xxx state-of-the-art, and a guaranteed upgrade path. And that is what
they did.

>or the GS.  Apple got greedy and thought "if we dump the GS and
>make the incompatible Mac more glamorous, then the schools would
>need to switch.  If we supported the GS, schools would feel

No way. The Macs, up until (and probably after) the first color Macs, 
were absolutely not intended for educational use in the same way as the
Apple II. They were monochrome DTP engines.

>That makes sense, since IBM and Apple are cooperating.  Even
>Micro$oft makes (made?) a PowerPC version of Windows NT.

Discontinued due to lack of a coherent standard for PowerPC hardware.
For which I do blame Apple, I must say.



=== From the Apple PowerBook of ======================
Lewin A.R.W. Edwards [Team OS/2] <http://www.zws.com/>
Programmer + Hardware Eng     Certified Apple Engineer
Need a programmer?    <http://www.zws.com/resume.html>
Tel  +61-412809805                   Fax +61-392281335
"No non-trivial programming task is atomic" - LARWE 97
=== (C) Copyright 1997 by Lewin A.R.W. Edwards =======