Was anyone else disappointed with the A1200?

[Bruce Abbott]

Originally Posted by Gorf

The A1200 came just a few month later - so these machines were clearly designed in tandem at more or less the same time.

If only that were true. A600 first engineering prototype was designed in August 1991. A1200 first engineering prototype was designed in March 1992, 7 months later. (dates from schematic drawings).

Originally Posted by oscar_ates

A600 opened the door wide for commodore's downfall

Commodore was going down no matter what they produced, but the A600 was closer to what they should have been producing than the A3000.

and here above people say it was a great machine. I find this weird.

The A1000 was a great machine, yes? And so was the A500... and the A600. If, like me, you didn't possess an A1000 and never had an A500, it was more than great.

Nobody thought about the cpu was still 7mhz chipset same etc after 6 years? Look to the German amiga engineers and marketing opinion about it:

I had an A3000. The A600 was way more portable, had a nice styling, and was much more compatible. Having the same chipset and CPU as the A500 is what I wanted!

Bill Sydnes head was cut off by the other inept commodore executives Gould and Ali after 1 year in his position because of the A600 flop.

But he managed to get the A600 out before getting fired, so it's all good!

[grelbfarlk]

[Gorf]

Originally Posted by Bruce Abbott

But he managed to get the A600 out before getting fired, so it's all good!

Aren’t we all glad he managed to produce the first nail in the coffin …

What a nonsense

[Gorf]

Originally Posted by Bruce Abbott

That is not what ARM did. In fact Acorn's ARM based computers were a flop, and licensing ARM did nothing for them. I presume that's not the outcome we wanted for the Amiga - ergo it was not 'done right'.

For Arm as a company it worked out very good obviously.
Of course not for Acorn, because Acorn was at this point already a subsidiary of Olivetti, and could not profit from the spinoff of Arm.
The new investors of the spinoff did profit - amongst them Apple, which already invested in Arm in 1990: 43% for just $3M.

MSX was the result of combined efforts by Microsoft and ASCII Corporation (a magazine publisher who sold Microsoft BASIC in Japan) to develop a home computer standard. Of course it would include Microsoft BASIC. ASCII Corp then expanded into making software and peripherals for gaming consoles. Neither company had a computer line of their own to promote - a totally different situation to Commodore.

Of course it is different - that is the whole point.
Commodore would have to leave production of computers, if it were to license clones. It can produce reference designs, evaluation boards and so on, but the end customer product would be made by clone makers …
Same goes for the production of chips:
Commodore would design them, others would produce them.

Not saying Commodore should have done this, but it might have been one way to survive longer. But not the only way.

I would note that Apple's drop in profitability in 1993 was greater than Commodore's, but comparing Apple to Commodore is a red herring. Between 1990 and 1995 Apple's net income dropped. One would think an ecosystem that was 'healthy and growing' would improve net income, not reduce it.

The drop in 1993 was in large part because of the Newton: new product, huge marketing budget, meager sales…
But nevertheless Apple stayed profitable that year.
They had at this point billions in cash and assets and wanted to increase marketshare - so they lowered the prices and widened the product lineup of Macs and increase the numbers of Macs sold.
Earning still halve a billion in 1995, but selling 3x more machines than in 1990.

Even at its best years, at the height of the C64 sales, Commodore had not more than 100M in profit.

[Promilus]

@Gorf

Commodore would design them, others would produce them.

Not saying Commodore should have done this, but it might have been one way to survive longer. But not the only way.

Of course. That's just an option, not something I recommend as well. But something worth exploring. In downsized form Commodore might still earn healthy money with little risk for themselves and with enough money landing in r&d department - which obviously be the sole focus. It could go either way - as fabless, or using own factories (which would need investment of course but might provide some variety of chips for 3rd parties like e.g. WDC?) Should Commodore survive mid 90s it could even release PCI based card with amiga chipset for PCs and next gen Amigas to provide partially software but mostly hardware compatibility with ECS (and maybe AGA) to still get money from largely obsolete technology at that point.

[Bruce Abbott]

Originally Posted by Gorf

For Arm as a company it worked out very good obviously.
Of course not for Acorn, because Acorn was at this point already a subsidiary of Olivetti, and could not profit from the spinoff of Arm.

I'm not talking about Acorn or ARM, but the Archimedes computer they produced. If you are interested in maintaining the platform (Archimedes or Amiga) what Acorn did with ARM is not a good example.

Even at its best years, at the height of the C64 sales, Commodore had not more than 100M in profit.

That's right, and some other famous home computer manufactuers did even worse. People here are talking as if Commodore was in the same league as much larger companies or even the entire PC industry. They managed to sell a lot of C64s by cutting prices to the bone and being that little bit better than the even more flawed competition, at a time when the market was still evolving and small players were still able to enter it.

That was an amazing time because there was so much quirky innovation at a techcological level that hobbyists could interact with, which is why that era is so popular for retro computing today. But it wouldn't last forever. By the end of the 90's PCs were becoming ever more complex and costing huge amounts to develop, only possible due to the 'team effort' of many players in the industry.

[Promilus]

Why would one maintain obsolete platform when new one was introduced? Amiga users have biased opinion because Commodore didn't last long enough to introduce new and largely incompatible Amiga. So they assume platform must be maintained until the last user dies (or company itself). And no. Team effort doesn't really apply to PC that much. Sure there are things like USB IF, JEDEC, PCI Sig, Khronos Group etc. But those aren't restricted to IBM PC clones and Commodore would have been most likely be a part of those as well.

On the other hand technologies like new GPUs came from a fierce competition (not inter-companies team effort) which basically left 2 dominant companies (nvidia and ati -> amd). Same with processors and although Intel and AMD have cross-license agreement it doesn't cover all technologies. E.g. AMD had 3DNow! and intel was never interested (for obvious reasons, it was just a stopgap solution), AMD have IMC since K8 and Intel only introduced that in Nehalem several years later. Same goes to HyperTransport vs QPI. Even now there's hardly any chipset for intel processors other than intel chipset and hardly any chipset for amd processors other than amd chipset. So much for "team effort". Commodore knew 68k is going EOL. Commodore knew PCI is a way to go. And sooner or later incompatible platform would emerge. And what amiga fans did? Well they maintained the last introduced platform themselves eventually bringing both PCI and USB in fairly limited form to Amiga world as well. And entirely new CPU (which is already bigger change than between A3000 or A5000 and A7000 right?)

[Gorf]

Depends on how you define "incompatible platform".
Over the decades: sure.
But each new model has to offer at least large backwards compatibility.
In PC-land it took forever to sneak out old DOS 8086 compatibility and while Apple was a little bis faster it always provided full compatibility after each architecture change for a couple of years.

But there is nothing that would have stopped Commodore from doing the same, if it had survived…

[Gorf]

Originally Posted by Bruce Abbott

I'm not talking about Acorn or ARM, but the Archimedes computer they produced. If you are interested in maintaining the platform (Archimedes or Amiga) what Acorn did with ARM is not a good example.

But it was!
It was the best decision Olivetti could have made - that they simply had no long lasting interest in the Archimedes does not make the other part a bad decision.

And now: today you can run RiscOS on an Pi and can use the speed of the CPU directly without costly emulation!

https://www.riscosopen.org/content/d...s/raspberry-pi

That's right, and some other famous home computer manufactuers did even worse. People here are talking as if Commodore was in the same league as much larger companies or even the entire PC industry. They managed to sell a lot of C64s by cutting prices to the bone and being that little bit better than the even more flawed competition, at a time when the market was still evolving and small players were still able to enter it.

Having 40% marketshare of ALL computers sold is really not just "a little bit better" - in 1984 this was more than the entire PC industry.

For a short time Commodore was a giant. But they could not capitalize on it.
For one because of Jack’s price strategy and also because Irvin did not want to issue new stock, because that would dilute his share of the company.
Commodore could have collected hundreds of millions of new capital during its hight of success and invest in modernizing its chip plant and R&D.

[Promilus]

@Gorf - even modern PC can run software for old DOS, either through DOSBox (or similar software) or even directly (although it's getting harder since UEFI). This only shows there's no big reason to keep everything backward compatible at hardware level. There already are plenty emulation layers built in into BIOS/UEFI - mouse can be treated as dos compliant device even if it's USB HID device. Same with keyboard. So - as you see - there's always some workaround which let us get new and improved hardware without actually having to support old solutions IN said hardware directly.

Apple did use totally software emulation integrated in OS (68k@ppc and ppc@x86), now they do use hybrid approach as part of x86 emulation is done inside ARM CPU and the rest is handled by software in OS.

One way or another NG Amiga would have to drop both 68k and OCS/ECS/AGA compatibility in hardware. And most of those A500 and 1200 users which are pretty "traditional" now would likely join the revolution in the past should new platform look appealing (for both users and developers) and some decent solution for backward compatibility could be devised. There are hardcore 68k traditionalists now because Commodore died and PPC NG lead nowhere with most current solutions being ridiculously expensive while still significantly lagging behind mainstream. All current turbo solutions offload much of the functionality to turbo itself (with rtg, audio, network and storage, even controllers). It already is some kind of NG amiga on top of classic amiga.

[Gorf]

@Promilus
All I am saying is: you really can not make a hard cut - it has to be a gradual process.
The PPC based NG-Amigas (of AROS on x86 for that matter) did cut off chipset support too early, at a point in time when uae was not good and fast enough yet.

At some point Commodore or a functional Amiga company would have needed to switch architecture - sure.
I guess AGA on a PCI card would have been the best decision for this transition. The base platform even could have been a PC.

Imagine Amithlon with a AGA-on-PCI* next to modern RTG and 3D in the late 90s
(Something like a reverse PiStorm in some sense)

*) or AAA or Hombre on PCI … whatever would have been the last iteration…

[TEG]

Originally Posted by Gorf

For one because of Jack’s price strategy and also because Irvin did not want to issue new stock, because that would dilute his share of the company.
Commodore could have collected hundreds of millions of new capital during its hight of success and invest in modernizing its chip plant and R&D.

I cannot find it back at the moment but I read that effectively, Tramiel wanted to go to shares and Gould to debt and it was part of the clash between the two.

For memory:



Talking about compatibility, they should have sold the A1000 bundled with a C64 emulator to keep their base, at least to show they were taking care because it was prejudicial for the A1000 itself. Why this new platform will not be replaced again by a new one in 3 or 4 years? Instead of that, it was all about PC compatibility obsession.

I remember, magazines were asking the question in articles dedicated to the future incredible machine of Commodore.

[Promilus]

Well, yes, but then what's the point of Commodore 128? C64 compatibility (either embedded in stock config or through optional hardware and software - as bare 7MHz 68000 would most likely fail to handle that task) is a nice idea but it either that or C128. Not both. And C128 was pretty successful as well.

[TEG]

There was A64 on the Amiga but I don't know about the speed. Emulating the 6510 at full speed does not seem to me an unfeasible task for the 68000 and their would have be no problem for the sprites part.

You're right about the C128, ideally should have been included too. But I don't know if there would be a legal problem distributing the Basic V7 in the Amiga. And I think nobody would have complain if the CP/M part was missing

[Promilus]

afaik A64 was only good for BASIC. Games were often distorted and demos didn't run properly either. What was good - there weren't high requirements but still it didn't run full speed 6510 on 7MHz 68000, only higher clocked 020/030. I wouldn't count on any good software emulation of C64 below 040. Both speed and compatibility wise. That's why hardware solution for C64 compatibility INSTEAD of release of C128 might have been exactly what A1000 needed to make even more astonishing debut. But the biggest problem of that solution was - would they make enough A1000 & add-ons to satisfy the market?

[Gorf]

Originally Posted by Promilus

afaik A64 was only good for BASIC.

Yes. It has its own Kernal- and BASIC-emulation wich takes up around 200k, but manages to be quite fast on a vanilla Amiga ...
(But not for games.)

A pure C64-BASIC mode and being able to use the 1541 might have been enough. Maybe even catch the usual peeks and pokes. That is what the Achimedes later did: providing a BASIC mode that was compatible with the Beeb, but no full emulation.

A full emulation is probably impossible on a 7MHz 68000

Darek Mihocka managed to reach 40% speed of an Atari800 with his XFormer on a 8MHz ST.
And he used about every trick in the book:
https://www.atarimagazines.com/st-lo...XFORMER_II.php

Well not every Trick: Peter McGavin (or Ulf Nordquist?) invented Pseudo-JIT to emulate a z80 and the Amiga - the same method Renee Cousins uses now for the Buffee project.

So maybe something like a C64-XFormer combined with PJIT and enough RAM might pull this miracle off ... but more likely you would still need a 14MHz 68k

Here is the original Author of XFormer confronted with the newest iteration of his emulator (XFormer 10) ... emulating hundreds of 8bit Ataris in parallel on a modern PC:

https://fb.watch/j4Oz2geBKJ/

(skip to minute 1:50)

That's why hardware solution for C64 compatibility INSTEAD of release of C128 might have been exactly what A1000 needed to make even more astonishing debut. But the biggest problem of that solution was - would they make enough A1000 & add-ons to satisfy the market?

the C128 ... Frankenstein's Bil Herd's Monster, literally build from dead parts

It had of course the worst timing!
A striped down version, without Z80 and without VDC, but still with 128kB and 2MHz in early 1984 instead of the disastrous 264 line, would have been perfect.

People wanted a better but still compatible C64 and no Plus/4

[Cris1997XX]

Originally Posted by Gorf

Yes. It has its own Kernal- and BASIC-emulation wich takes up around 200k, but manages to be quite fast on a vanilla Amiga ...
(But not for games.)

A pure C64-BASIC mode and being able to use the 1541 might have been enough. Maybe even catch the usual peeks and pokes. That is what the Achimedes later did: providing a BASIC mode that was compatible with the Beeb, but no full emulation.

A full emulation is probably impossible on a 7MHz 68000

Darek Mihocka managed to reach 40% speed of an Atari800 with his XFormer on a 8MHz ST.
And he used about every trick in the book:
https://www.atarimagazines.com/st-lo...XFORMER_II.php

Well not every Trick: Peter McGavin (or Ulf Nordquist?) invented Pseudo-JIT to emulate a z80 and the Amiga - the same method Renee Cousins uses now for the Buffee project.

So maybe something like a C64-XFormer combined with PJIT and enough RAM might pull this miracle off ... but more likely you would still need a 14MHz 68k

Here is the original Author of XFormer confronted with the newest iteration of his emulator (XFormer 10) ... emulating hundreds of 8bit Ataris in parallel on a modern PC:

https://fb.watch/j4Oz2geBKJ/

(skip to minute 1:50)




the C128 ... Frankenstein's Bil Herd's Monster, literally build from dead parts

It had of course the worst timing!
A striped down version, without Z80 and without VDC, but still with 128kB and 2MHz in early 1984 instead of the disastrous 264 line, would have been perfect.

People wanted a better but still compatible C64 and no Plus/4

I think a 14MHz 68020 with enough fastRAM would run Commodore 64 games no problem

[TEG]

So the 6502/6510 8 bits @1 MHz was that powerful? OK

I see here that 6502 was sold $7.45 so it was certainly produced for something like $3. I wonder if it would have been possible to made an architecture which include it as a slave of the 68000 for emulation purpose. After all their had the control over Agnus for logic interfacing. Eventually I imagine you can add some parts of the VIC/VIC-II in Denise to help if needed.

Or they may found some tricks to help the software emulation. Perhaps a dedicated chip with some parts of the 6502/651/8502. They had an equip of engineers and all plans to hand.

But it would implied a vision since the day they bought the Amiga tech for sure and perhaps as Gorf said, fundraising when they was at the top to pay for R&D.

[Promilus]

So the 6502/6510 8 bits @1 MHz was that powerful? OK

You think of 6502 emulation as something trivial. It's not about computing performance but emulating logic behind CPU. The same can be said about 68k emulation. To create cycle accurate 68000 7MHz few MHz x86 is needed. Not because MC68000 is so powerful but because making it accurate takes many, many instructions. JIT doesn't care about it that's why it can run code very fast but code written for specific speed will always go bad. And many C64 games do not work correctly on anything else than ~1MHz 6510. And there's yet another part to consider - VICII, his timeslots and text mode for graphics (which many platform games did use essentially using character map as tileset - to conserve memory of course and improve speed). It's not like Amiga chipset does handle that accurately in hardware. It doesn't even have such text mode afaik.

And on top of that 6502 has unique ISA with small amount of registers . So you have to emulate 8bit CPU flags along with instruction intended operation result (like bit shifting, increment, decrement and so on) and all of that being as close to original speed as possible. It wouldn't matter if there was 64bit 7MHz RISC there... it would still have a hard time emulating 6502 accurately enough.

Some hardware would have to be created to support C64 compatibility mode on bare A1000/2000/500. One can only wonder what A1000 with C64 compatibility mode could accomplish.

[Locutus]

I suspect a C64-SoC-on-a-card would have been the way to go, do something smart to allow it to render into a pub screen using a analogue video overlay or some such.

It's essentially what Apple did with their Apple 2 compatibility cards for the NuBus Mac's, and those sold in reasonably respectable numbers.

Again, CBM never invested properly in their semiconductor foundry so a single chip C64 SoC would never have happened.