Was anyone else disappointed with the A1200?

[Bruce Abbott]

Originally Posted by oscar_ates

Why put that 5$ 1984 68020 CPU and no fast-ram for machine released in 1994?

Did not they even follow what was happening on PC and consoles side like games Doom-II, tie fighter and need for speed?

The CD32 was released in September 1993, not 1994. So back at you.

As my documentation shows, the CD32 design was started in July 1992 and firmed up in October 1992, at around the time the A1200 was announced. Date on the first engineering prototype schematic is October 6, 1992.

Doom II was released on October 10, 1994. When the CD32 was being developed even the original shareware Doom didn't exist, and that officially needed a fast 486 which was way outside the reasonably expected capabilities of a cheap games console.

Tie Fighter was also released in 1994. Minimum spec was an 80386DX with 2MB RAM. It is quite similar to Wing Commander which was ported to the CD32 in 1993. Some think Wing Commander was a better game than Tie Fighter. It looks and runs pretty good on the CD32, much better than the earlier A500 version!

Had Commodore not gone bankrupt, in Fall 1994 they planned to release an enhanced CD32 with the AGA+ chipset running at 57MHz (providing 2x the performance of AGA), and a 28MHz 68EC030 CPU. NVRAM would be increased to 32k. I can't find any detailed information on this AGA+ chipset, but I presume it would run at a bus speed of 14MHz, doubling the CPU to ChipRAM bandwidth and blitter speed. The A1200+ would receive this stuff too.

Unfortunately Commodore just didn't have the money to survive that long. I'm not sad about it though. A similar thing happened to all the other home computers, a good thing IMO because it maintained the essence of those platforms. So what if the Amiga fell behind the latest PCs in hardware capabilities? It's now 30 years later and we are still enjoying using it. Having slightly more powerful hardware wouldn't have made much difference - 2 years later it would still be behind. But AGA still impresses me today when I use my A1200. Somehow this 30 year old technology manages to be more responsive and reliable than my 'modern' PC - and a lot more fun.

[Promilus]

Had Commodore not gone bankrupt, in Fall 1994 they planned to release an enhanced CD32 with the AGA+ chipset running at 57MHz (providing 2x the performance of AGA), and a 28MHz 68EC030 CPU

Yes, they did plan AAA and Hombre as well. Hadn't they went down maybe you'd get that incompatible shitty Amiga console.

Unfortunately Commodore just didn't have the money to survive that long.

From your perspective as hardcore 68k Amiga fan it's actually better they didn't. Open to discussion is where would they have made new chips and at what price. AGA itself already proved to be demanding in that regard.

[AestheticDebris]

Originally Posted by Bruce Abbott

My theory is that owning an Amiga made them feel superior, and when this was eroded their egos took a hit. Someone was to blame for that and it wasn't them - it was Commodore, obviously.

My theory is that most people are able to look back at their favourite platforms and see the weaknesses that, at the time, they may have overlooked. And can look critically at the decisions Commodore made which ultimately lead to bankruptcy and failure and see that not all of them were necessarily good ones. Almost like it's ok to like the Amiga without needing to be a sycophant.

YMMV of course.

[Karlos]

I'd go further. It's possible to recognise that your favourite machine had issues even at the time, but you forgive those because it allows you to carry on using your favourite software. I have owned 3 A1200's, two of which are still (technically) working. One of them was a Frankenstein tower (though internally more modest, mostly for the space, cooling and power - no PCI board, just BV/BPPC). I got another that was refit fist with my older 040 but then down to a basic fast ram only config. Perfect for running OctaMED as a basic sequencer and Paula 4ch sampler. The tower machine was for more serious productivity use (esp coding) and both were great for gaming (whdload), with the more expanded machine being for the few higher end titles.

I didn't buy them thinking they were an answer to the 386 or 486. I just didn't think about those machines at all. Other than having to use them at university occasionally, they were just a non entity.

[TEG]

Originally Posted by Bruce Abbott

So the main concern of Commodore when they made the CD32 was its compatibility with the existing Amiga library !??

We saw you enjoyed it:

Originally Posted by Bruce Abbott

It was amazing to see our CDTV title working so much better on the CD32 with no changes whatsoever!

But what was the point !?? To be able to run the wonderful CDTV software library and seduce all the public that this represented? For existing Amiga users to be able to run their floppy games collection?? For editors to just have to put theirs games on a CD without improvements?

Really, it does not make sens for this machine. It was supposed to be the first of a new line. They should have put at least a 6030 inside and others enhancements the techies would be able to provide without thinking in term of compatibility.

I suspect it was just an excuse to finish to definitively sink the brand and Medhi Ali a submarine to carry out the mission. Perhaps it was because they can't do otherwise due to the plan being very late but it's the same. No excuse.

[TEG]

Originally Posted by Bruce Abbott

3. Having any FastRAM would raise the cost more because it would require a separate memory controller and a significantly larger more complex PCB.

Argument rejected. We were in 1993, no more in 1985. Akiko was produced by VLSI, so I guess, to have a better lithography than CSG was able to output. And so adding the memory controller in this chip must not have been very expensive.

Akiko is the CD32's all-purpose 'glue' chip and forms part of the AGA chipset used in that system. Akiko is responsible for implementing system glue logic that in previous Amiga models were found in the discrete chips Budgie, Gayle and the two CIAs.
https://en.wikipedia.org/wiki/Amiga_custom_chips#Akiko

Originally Posted by Bruce Abbott

4. If it was FastRAM the developers might have been more keen on it.

The A1200 only had 2 MB ChipRAM, so developers would want to target this size anyway. Having a CD-ROM drive removed some of the pressure for more RAM because CD loading times were much faster than floppy, so you could get away with holding less stuff in memory and/or not have to decompress it before use.

Yeah, for all the graphic and sound elements but having FAST completely free the bus for the CPU computation. So you have less compromises to do of style "if I add two sprites it will slow down all because more memory slots access are used", ie intricate programming to get the juice of a machine which was precisely what publishers didn't want any more. And so an important point for the success of your platform.

The CD32 architecture add this "secret" weapon thought 8 years before and when all the fire power was needed, it was not used.

[Bruce Abbott]

Originally Posted by TEG

So the main concern of Commodore when they made the CD32 was its compatibility with the existing Amiga library !??

No - the concern was compatibility with other Amigas, specifically the A1200. The idea was that they would produce a CDROM drive for the A1200 that would allow it to play CD32 titles. That couldn't happen if the CD32 had some specific hardware that couldn't be emulated. It would also mean that A1200 users who bought a CD-ROM drive from elsewhere would be able to run CD32 titles.

I'm glad they thought of that. I developed our CD32 title on the A1200, and only produced a 'gold' disc when the product was (hopefully) complete. This meant I could start developing for the CD32 long before getting my hands on one. It was also a lot more convenient and cheaper (discs were costing us $500 each to burn!).

But what was the point !?? To be able to run the wonderful CDTV software library and seduce all the public that this represented? For existing Amiga users to be able to run their floppy games collection?? For editors to just have to put theirs games on a CD without improvements?

And what's wrong with that? Imagine if you bought a new PC and none of your existing software would run on it. You'd be pissed. And it wasn't just CDTV software, many A1200 titles would be well worth porting to it too.

Remember that people who bought a CD32 might not have any other Amiga, and pirating wasn't an option. This was a great opportunity to bundle some of those classic titles that would work even better on the CD32. Even former A500 users might be interested in getting more playable versions of older games.

Really, it does not make sens for this machine. It was supposed to be the first of a new line.

Er, no it wasn't. It was supposed to be a cheaper A1200 for people who just wanted to play games.

I suspect it was just an excuse to finish to definitively sink the brand and Medhi Ali a submarine to carry out the mission. Perhaps it was because they can't do otherwise due to the plan being very late but it's the same. No excuse.

Commodore was effectively sunk before the CD32 was released. They couldn't even make enough A1200s to meet demand because parts suppliers were insisting on cash up-front. The idea that an even more expensive console would have saved them is ludicrous. Their best bet was to offer the existing Amiga architecture at a lower price point that was closer to other console systems - specifically the Sega Mega CD.

[Bruce Abbott]

Originally Posted by TEG

Yeah, for all the graphic and sound elements but having FAST completely free the bus for the CPU computation. So you have less compromises to do of style "if I add two sprites it will slow down all because more memory slots access are used", ie intricate programming to get the juice of a machine which was precisely what publishers didn't want any more. And so an important point for the success of your platform.

Funny I never had that problem - perhaps because I was quite happy with a machine with enough bandwidth that the CPU ran more than twice as fast as the CDTV/A500. And perhaps I was also happy with double the RAM, 4 times more colors and 4 times wider sprites.

I'm not decrying anyone who who pushed the CD32 to its limits, but every system has limits - and every coder knows you can easily bang up against them if you try. That doesn't mean limits are bad. Many great games got their character from having to work within the limits of the platform. Modern games leave me cold because the slick graphics don't add anything. I bought several of the 'enhanced' Crystal Dynamics Tomb Raider games for the PlayStation 2, thinking the improved graphics would make them better. I was wrong. The greater 'realism' didn't improve the atmosphere and made them harder to play.

Tons of great games could have been produced for the CD32 without trying to push everything to the max - and then complaining that you had to 'compromise'.

[Gorf]

Originally Posted by TEG

Argument rejected. We were in 1993, no more in 1985. Akiko was produced by VLSI, so I guess, to have a better lithography than CSG was able to output. And so adding the memory controller in this chip must not have been very expensive.

The real deal would have been to "reverse" the memory design of the Amiga:

Only FastRAM with a tailored memory controller for the 68020/68030 and changing Alice into a programmable DMA-controller.

ChipRAM would then be simply any regions (more than one) that is cleared for DMA access. For Alice this would still look like a continuous region located at the usual spot in the memory map - but the CPU could have (almost) full speed access to e.g. a bitmap buffer in FastRAM, write all data and declare that region ChipRAM afterwards - in exchange for the old bitmap buffer - during the vertical blank.

That way the the ChipRAM speed bottleneck would be largely reduced.

(for compatibility reasons the first 1 MB could be hardwired, otherwise a 030 with MMU would be handy ...)

[Promilus]

Originally Posted by Bruce Abbott

No - the concern was compatibility with other Amigas, specifically the A1200. The idea was that they would produce a CDROM drive for the A1200 that would allow it to play CD32 titles. That couldn't happen if the CD32 had some specific hardware that couldn't be emulated. It would also mean that A1200 users who bought a CD-ROM drive from elsewhere would be able to run CD32 titles. (...)

And yet CD32 has such hardware. IIRC Akiko, parts of CD-ROM controller and obviously FMV add-on. You're just fooling yourself. They made it based on AGA because it was cheap option to let some cd rom game collections be made along with new games developed in parallel for A1200 and CD32 so without cost increased. It also means as a gaming console it didn't offer anything relevant than CD-ROM equipped A1200. And obviously FMV add-on didn't really help either. Lack of easy way to hook up floppy was kind of pita.
CDTV game library wasn't vast. Compatibility with that wasn't priority, priority was to establish position in gaming console market. And it kind of did, initially, in UK alone iirc. But either way it would not survive the next wave of gaming consoles so what's the point anyway? Even if they didn't get a ban to US, even if they were able to produce significant amount of gaming consoles it would still have been a dead end. Just like it was with 32X which actually has better specification and is capable of much more (which fairly recent Doom 32x: resurrection proves), same can be said about Jaguar which did kick the bucket anyway as well. So why the claim CD32 was built well? It was cheap AGA implementation which they though CD presence alone would let it fly off the shelves with no actual plan whatsoever to what's next. And the AA+ or CD64... that's the same kind of BS as AAA and Hombre.

[TEG]

Originally Posted by Gorf

The real deal would have been to "reverse" the memory design of the Amiga:

Only FastRAM with a tailored memory controller for the 68020/68030 and changing Alice into a programmable DMA-controller.

ChipRAM would then be simply any regions (more than one) that is cleared for DMA access. For Alice this would still look like a continuous region located at the usual spot in the memory map - but the CPU could have (almost) full speed access to e.g. a bitmap buffer in FastRAM, write all data and declare that region ChipRAM afterwards - in exchange for the old bitmap buffer - during the vertical blank.

That way the the ChipRAM speed bottleneck would be largely reduced.

(for compatibility reasons the first 1 MB could be hardwired, otherwise a 030 with MMU would be handy ...)

Is it a kind of architecture already used in some platform you know? And what would be the advantage? More flexibility?

[TEG]

Originally Posted by Bruce Abbott

I'm not decrying anyone who who pushed the CD32 to its limits, but every system has limits - and every coder knows you can easily bang up against them if you try. That doesn't mean limits are bad. Many great games got their character from having to work within the limits of the platform. Modern games leave me cold because the slick graphics don't add anything. I bought several of the 'enhanced' Crystal Dynamics Tomb Raider games for the PlayStation 2, thinking the improved graphics would make them better. I was wrong. The greater 'realism' didn't improve the atmosphere and made them harder to play.

Tons of great games could have been produced for the CD32 without trying to push everything to the max - and then complaining that you had to 'compromise'.

OK but even with a better platform there was still margins by large to push to the limits. More muscled and dynamic 2D games on the Amiga, world would have been very welcome.

"Primitive" 3D games like Star Fox on the Nintendo (1993), too. And perhaps we would have witnessed the emergence of others styles we don't know about thanks to creativity of the developers specific to the Amiga, proved with Another world, Apidya, SOTB and so on.

[Gorf]

Originally Posted by TEG

Is it a kind of architecture already used in some platform you know? And what would be the advantage? More flexibility?

Most unified memory architectures used to have a variant of this method.
Now in the 2020s everything is multichannel .-..

The RasPi (1-3) is still close but different again as the VideoChip is the memory controller and the CPU is getting DMA access ...

Advantages in my idea:

Commodore knew already how to make fast memory controllers (Ramsey) and how to make DMA controllers (SuperDMAC) ... so most of the puzzle pieces are already there.

And yes: flexibility would be the big advantage here!
Your ChipRAM could utilize more memory if you have a big FastRAM expansion, by mapping in and out different areas ... so no more copying between those two regions.

It this memory mapping is fine grained enough one could also puzzle together different blocks on the screen and move them around - a whole different set of gfx-effects could be achieved this way.

[Bren McGuire]

this is boring lol

[Gorf]

Originally Posted by Bren McGuire

this is boring lol

Is someone forcing you to read this thread?

[Bruce Abbott]

Originally Posted by Promilus

And yet CD32 has such hardware. IIRC Akiko, parts of CD-ROM controller and obviously FMV add-on.

Hardly. The CD-ROM controller was only necessary because they were using a cheap mechanism with no internal controller (a good idea for this class of machine). Other Amigas would use standard SCSI or IDE drives. The other main job of Akiko was to integrate just the essential parts of the CIA chips etc. to get the cost down. No new functionality there. C2P was a last minute addon that would be emulated on other Amigas.

FMV was an addon that could have been provided for other Amigas too. Big box Amigas could already play FMV from 1992 with Peggy, and any Amiga with a SCSI port could do it from 1996 with Squirrel MPEG.

You're just fooling yourself. They made it based on AGA because it was cheap option to let some cd rom game collections be made along with new games developed in parallel for A1200 and CD32 so without cost increased.

Sure it helped to keep costs down, but more importantly it expanded the user base and prevented existing users from feeling left out. Why would this be considered a bad thing?

What they really wanted was to bring CD-ROM to all Amigas, making it the standard distribution format (same as PCs were starting to do - for good reason). How do we know this? First off Commodore produced the A570 CD-ROM drive for the A500. It was originally intended to be released at the same time as the CDTV so that A500 users could run CDTV titles. The CD32 was designed so that CD32 discs could run on any AGA machine with CD-ROM drive (or even just a hard drive if it had enough space).

But why should you believe me on this? I could just be another Amiga fan making stuff up. Best go to the primary source:-

Brian Bagnall -

Commodore's software engineers set the software strategy for the Amigo [CD32]. They wanted all CD games and apps to work across all Amiga platforms with CD-ROM. New Amiga titles were not expected to work on the old CDTV, but Andy Finkel would attempt to achieve compatibility of existing CDTV titles with the Amigo. Furthermore, Amigo titles would work with any AGA Amiga (A1200 or A4000) with CD-ROM.

It also means as a gaming console it didn't offer anything relevant than CD-ROM equipped A1200.

...except price and convenience. This was by design.

And obviously FMV add-on didn't really help either.

FMV was an addon because it had limited appeal. Its functionality could easily be duplicated with a plug-in card, but this was not the main focus of the CD32. Its main purpose was to provide a lower cost way to play the latest Amiga games, strengthening the Amiga line (a lost cause perhaps, but that was the plan).

Lack of easy way to hook up floppy was kind of pita.

Perhaps, but adding a floppy drive port would raise the cost unnecessarily since all CD32 titles came on CD.

However Commodore did design the CD32 to be expandable to full computer status. A floppy drive interface could be made quite cheaply. I made one with a few TTL chips, and built the floppy drive into the side of the CD32. A Polish outfit called TOMS produced an addon with disk drive and RGB video ports.

it would not survive the next wave of gaming consoles so what's the point anyway? Even if they didn't get a ban to US, even if they were able to produce significant amount of gaming consoles it would still have been a dead end.

By the same logic the entire Amiga line was a dead end, so why bother doing anything?

I for one am glad that they managed to squeeze one more Amiga out before the end, and I'm glad that it wasn't something radically different. It meant that CD32 owners didn't have an orphan.

It was cheap AGA implementation which they though CD presence alone would let it fly off the shelves with no actual plan whatsoever to what's next. And the AA+ or CD64... that's the same kind of BS as AAA and Hombre.

Wrong. Commodore was constantly planning for 'what's next'. The problem wasn't planning, it was implementation. However by 1992 they had systems in place that greatly reduced chip development time. Had Commodore not been going down the tubes in 1994 I am fairly confident that they could have produced AA+ in that year and AAA in 1995, as they planned.

Commodore also knew that 68k was a dead end, which is why they planned to switch to PA-RISC with Hombre. That was a big step, but a necessary one if they were to survive beyond 1995.

It's funny how Commodore gets excoriated for having 'no actual plan whatsoever', when that certainly applied to the vast majority of PC makers. They simply shoved in whatever was available at the time, with the sole goal being to sell something now. Who cares what might be coming up in the future? That strategy worked very well for them.

[Thomas Richter]

Originally Posted by Bruce Abbott

So you think Commodore should have replaced OCS with a VGA chip? Which 32 bit VGA chip would they choose in 1989, and what would they do about the lack of compatibility with earlier Amigas?

The A1200 is from 1992, not 1989. For VGA chips, make your pick:


https://www.dosdays.co.uk/topics/Man...rrus_logic.php


Just to name one family.

Originally Posted by Bruce Abbott

Or do you think they should have added it on top at no extra cost? Yes, magic is what you are talking about.

Certainly at a lower cost compared to developing chips in house.

[Thomas Richter]

Originally Posted by Bruce Abbott

Wrong. Commodore was constantly planning for 'what's next'.

So much that they killed chip development. For example new chips for the A3000 - instead they sold the same shit, just more expensive, and fixed it with chewing gum and hotglue, speak the flicker fixer, instead of providing a chipset that could prodice a VGA signal natively.

[Promilus]

By the same logic the entire Amiga line was a dead end, so why bother doing anything?

Being a dead end was apparent in 92 when AIM came to life. Anything beyond that holding onto 68k was dead end. Anything which came before was active development.

C2P was a last minute addon that would be emulated on other Amigas.

And a handful of games did use that but emulating it does slow things down - in case of bare 1200 considerably due to increasing ratio of c2p routines overhead vs actual rendering tasks.

except price and convenience. This was by design

Most CD platforms at that time already tried to bump up sales with soundtrack and fmv... CD32 developers in general didn't even do that with few exceptions. Also AFAIK CD32 uses memory mapped NVM (EEPROM) for savegames which other Amiga models does not have so either way that's one another functionality which would have to be somehow emulated (e.g. battery backed up RAM). And afaik not all cd32 games work (properly) on cd-rom equipped A1200/4000 which means it really wasn't nearly as compatible across platform as you claim it to be. And if you need some software compatibility layer... then hell, why try so hard to maintain hw compatibility?

Perhaps, but adding a floppy drive port would raise the cost unnecessarily since all CD32 titles came on CD.

You've just wrote it was couple of TTL chips... And no, it is not "unnecessarily" as it's "to provide as ample software base as possible after launch".

Commodore was constantly planning for 'what's next'. The problem wasn't planning, it was implementation.

No. It was planning. Unrealistic goals and timetables. Again... who's responsible for couple of thousand scrapped C65 prototypes? Oh, it wasn't planned? Or what? Was C64GS planned or not? Aaaa...

Commodore also knew that 68k was a dead end

Yep, that's why Haynie claims he begun work on new 64b RISC gaming console in 91 and later on learned that Ed Hepler did independently work on something similar as well and should they cooperate that might actually finalize before Commodore financial troubles became irreversible and with compute power ample enough to keep up with Saturn and PS1. OFC everything works just great on paper and in limited simulation environment. And I did prove how delusional Hepler was with nearly cache-less PA-RISC implementation and it should be pretty obvious how delusional Haynie was claiming it could've been compatible with 68k amiga through software emulation... The one thing Dave was right is that with gaming consoles of that era backward compatibility wasn't actually that much of an issue.

That was a big step, but a necessary one if they were to survive beyond 1995.

Sure, but betting on architecture mostly related to workstations and servers for cheap home computers and consoles doesn't seem like the right choice either.

when that certainly applied to the vast majority of PC makers.

Well of course it does. And nobody is claiming that Acorn did all good with their tech or Atari, or Tandy, or Gateway, or Amstrad... what pains most of us is that Commodore went down having such a terrific piece of hardware since mid 80s. And it does feel they didn't push their advantage when they had one. Neither did they enough to keep the distance between Amiga and competitors. How could they... alone. It was impossible and plainly stupid to try anyway.

[Amiga Lifestyle]

How many computer platforms made it more than 10 years?

Sweet mother of Jesus this thread started in 2017. How many computer models lasted longer than this thread.