Unreleased Amiga chipset better than anything today?

[sokolovic]

Originally Posted by Bruce Abbott

No logic at all?

A2000 to A4000/040 to A4000/030 and A4000T is logical
CDTV to CD32 is logical
A500 to A500+ to A600 and A1200 is logical.

The only illogic is A2000 to A3000. Don't make your flagship model smaller than the previous one! Can't fit a Toaster in the A3000!

Nope it isn't logical.
Logic would have been making the AGA équivalent of thé A1000 in 1990 and release thé home wedge version in 1992 with a desktop version alongside. And you don't release 2 CD system one year and a half apart with a classy and expensive one based on the old chipset and a cheaper console oriented
one based on the new chipset.

[TEG]

Originally Posted by sokolovic

And you don't release 2 CD system one year and a half apart with a classy and expensive one based on the old chipset and a cheaper console oriented
one based on the new chipset.

Why not ? The first system wasn't selling. So you stop it and re-use the technology packaged in a better system. What was wrong, was the initial move, the CDTV idea.

[AestheticDebris]

Originally Posted by TEG

Why not ? The first system wasn't selling. So you stop it and re-use the technology packaged in a better system. What was wrong, was the initial move, the CDTV idea.

I mean it "makes sense" in that regards, as almost all things do. But what didn't really make sense was Commodore's general roadmap, which seemed often to be a case of throw everything out there and hope something sticks. I can seem some logic behind the CDTV and a better designed CD32 might have made sense too. But stuff like the A600 was literally all over the place (unsurprisingly given nobody could seem to decide if the project was a cost down version of the A500 or an upgrade option).

[grelbfarlk]

Does this make pretty good sense?

[CCCP alert]

The A2000 from Commodore GMBH project was chosen instead of what Jay Miner was working on, the 128 colour 68020 based Ranger. That's what I remember about that story.

Commodore freely admitted the A3000 was essentially a 'bundle machine' nothing new they just incorporated some popular add-ons being used with the A2000 and made a new machine out of it. If the A3000 came with 24bit graphics and a 12 or 16bit sound card on the Zorro slots at least it would have been an improvement rather than a streamlined A2500 etc.

[Promilus]

Originally Posted by TEG

Why not ? The first system wasn't selling. So you stop it and re-use the technology packaged in a better system. What was wrong, was the initial move, the CDTV idea.

Yeah... CDTV failed because it wasn't a console, it wasn't Amiga (there was no logo anywhere), it did support external floppy directly though. But for a CD - hardly any developer made some good game for CDTV. Now... what's the difference with CD32? They cut off floppy drive port on the back, stick Amiga on front and sold it as gaming console. Still without investment for either CD exclusive editions or for improvements coming from CD media of games already released on floppies. So what was the incentive to buy that piece of hardware? Big library of existing games? Yeah, but you need floppy for that or wait till someone releases CD games set. It obviously makes A1200 better choice still.

[Bruce Abbott]

Originally Posted by TCD

Is it? I still fail to see the need for the two middle steps.

The A500+ had the latest chipset and OS from the A3000. It came with 1MB standard, expandable to 2MB via the trapdoor slot - 4 times more ChipRAM than an OCS Amiga. That meant more graphics, more sound, bigger bitmaps etc., for better games and more sophisticated apps. Why would you not want that?

WB2 was big step up on WB1.3, with major improvements across the board both for developers and users. More and more applications coming out were WB2 only because developers were using its new features. If you wanted to run those apps you would need it, and you would want it anyway because it was better. Sooner or later WB1.x would have to be retired, so why not now?

The A600 was the first in a new range of more compact Amigas, made possible through surface mount technology and stuff from PC laptops (PCMCIA, 2.5" IDE). It would replace the A500+, with the added advantages of smaller size and better ergonomics. Gone was the modulator sticking out the back, and perhaps a hard drive hanging off the side. Now you had a machine with all that built into a unit not much bigger than a console and just about as portable.

Perhaps they went a little overboard by removing the numeric keypad, but it didn't bother me as I don't use those keys anyway (my current PC keyboard doesn't have them). I would rather have the smaller size.

The A600 played a role similar to the IBM PS/2 Model 30 or Sony PS One, keeping the 'legacy' platform alive while there was still demand. This policy worked well for Sony (over 28 million PS Ones sold).

Perhaps you think they should have skipped those models and just produced the A1200, and only made OCS A500s with WB1.3 until then. But that would result in even more complaints that Commodore was just sitting on its hands and not innovating (this time justified). So now they were wrong for bringing out new models too often? Damned if you do, damned if you don't.

[Bruce Abbott]

Originally Posted by Promilus

So what was the incentive to buy that piece of hardware? Big library of existing games? Yeah, but you need floppy for that or wait till someone releases CD games set.

You mean big library of pirated games, right?

Because if you were buying them the idea was that CD versions of them would soon become availabe. Many of those games would get enhancements, or play better simply due to the more powerful hardware (faster CPU, wider bandwidth graphics, quicker loading from CD etc.) which would make a huge difference to many games - well worth waiting for the CD32 version.

CDs were also much more reliable and durable than floppy disks, so you didn't have to worry about your copy-protected original disks going bad with use.

Most games had a lifecycle that in some cases wasn't long. Then they would be re-released as 'budget' titles, perhaps by a different publisher. The same would be done on CD. So the 'big library of existing games' is the IP, not actual games in boxes sitting on dealers' shelves.

It obviously makes A1200 better choice still.

No, it doesn't - for all the reasons above. Nobody was releasing A1200 versions of old games like they would for the CD32. As soon as developers heard of the CD32 they were onto it.

[Bruce Abbott]

Originally Posted by CCCP alert

The A2000 from Commodore GMBH project was chosen instead of what Jay Miner was working on, the 128 colour 68020 based Ranger. That's what I remember about that story.

Ranger wasn't great. As you say it maxed out at 7 bitplanes, and didn't have much else worth talking about. The case design they were working on was awful, with big square Zorro cards that plugged in horizontally.

The A2000 was a sensible development of the A1000, with conventional slots and a case and power supply big enough to put anything you wanted in there. It also replaced the expensive unwieldy Sidecar with a simple 'bridgeboard' card. The A2000 was a very practical design which was just what professional users wanted. Didn't have the rumored Ranger graphics, but that wasn't a bad thing because Ranger didn't offer that much - if it even existed at all except on paper.

Commodore freely admitted the A3000 was essentially a 'bundle machine' nothing new they just incorporated some popular add-ons being used with the A2000 and made a new machine out of it.

Demonstrably false.

If the A3000 came with 24bit graphics and a 12 or 16bit sound card on the Zorro slots at least it would have been an improvement rather than a streamlined A2500 etc.

24 bit graphics and 16 bit sound on Zorro? It was too expensive already. But if you wanted those things you could plug cards into it, and Zorro III was much faster than Zorro II which made the A3000 the ideal choice for this.

[Promilus]

So... out of every game released for CD32 how many actually DID get enhancements OVER regular AGA versions (because ... hint - A1200 came first!) You keep talking about enhancements, put actual number here. CDTV could enhance games over OCS the same way CD32 does AGA. So no, you are dead wrong. It did not happen widely. There was no great number of improved titles and only fraction of the great software library was available for CD32 during it's debut. So - again - what's the incentive to the USER? Aaa, nothing worth mentioning. And for developers? "you can write for basically the same hardware as A1200"? LoL. There were hardly any really decent upgraded games for AGA at that time as well as it was relatively new piece of hardware.

CDs were also much more reliable and durable than floppy disks

So what when you end up with just few CDs at launch and can't get even simply connect external floppy to play your A500 games.

Nobody was releasing A1200 versions of old games like they would for the CD32

There hardly was anyone releasing enhanced version for CD32 in the first place. So I don't know which developers you are referring to. Nobody got CD32 devkit at least half of an year earlier and agreed to make jawdropping CD enhanced game for the launch of this console. And nothing like that happened afterwards. So where the hell do you see a success here? Or even hypothetical one?

[hammer]

Originally Posted by Bruce Abbott

Bill Sydnes wasn't hired until 1991, so anything that happened before that wasn't on him. Pity you can't put the blame for those 'split R&D resources' on our favorite villain, eh?

Refer to Mehdi Ali. Obtaining the information for splitting R&D is difficult.

Originally Posted by Bruce Abbott

The C65 was mistake. Don't get me wrong I would loved to have owned one, but the priority should have been the Amiga, particularly the A500. Can you imagine how embarrassing it would have been if they had released the C65 while the A500 still had ECS? Even the frickin' C64 is getting 256 colors and the Amiga isn't?

Yes, the A3000 had a prototype of the AGA chipset in it in 1991. It was not production ready. Dave Haynie talking about the Buster chip for the A3000+/A4000:-

A3000 was released with buggy Buster chips.

A1200 doesn't need DMA Zorro 3 and Budgie is effectively a cutdown Buster.
A fully functional A4091 SCSI card wouldn't sustain Commodore.

AA3000+ clone was later released.

AGA's Lisa has a 16 million color palette with 256 color registers while C65 has 4096 color palette with 256 color registers.

Bill Sydnes wasted more than 6 months on "A1000 Jr".

Alice and ECS Agnus are very similar with the major difference being with AGA Lisa and ECS Denise.

Originally Posted by Bruce Abbott

The A3000+/A4000 wasn't ready to ship in 1991. Bil Sydnes figured that in the mean time they could quickly push out a cost reduced model based on the original A3000. But like the rest of Commodore he wasn't listening to the community, which was consumed by PC envy and already disappointed that the A3000 didn't have something better than VGA. So nobody wanted the new models and they had to rush out AGA machines quicker than intended. Pity they hadn't sensed the urgency sooner, say in 1989 or 1990. But at that time the A500 was selling well and nobody could see what was coming.

Amiga Hombre's memo was mentioned around February 1991.

Unlike the Japan Inc competition, Commodore doesn't have experience with hardware-accelerated 3D and related RISC CPU IP despite owning the MOS 65xx CPU technology. HP's PA-RISC with MAX SIMD technology was selected as the basis for the 3D accelerator.

Originally Posted by Bruce Abbott

It was around this time (1991/2) that the chip development cycle started getting shorter due to new chip design technology that could do in a month what previously took 6 months to a year. Previously a design had to be 'finished' (ie. ready for prototype testing) up to a year before production began. That means they needed to start on the 'next big thing' earlier, with a more realistic spec that could be produced on time.

The A3000 was also a mistake. Over-engineered in some respects, outdated in others, and way too expensive for what it was. They should have stuck to 'big box' models at the high end. Slimline models should have covered the mid-range, not top end. Bill Sydnes got that right, he just didn't understand that we weren't interested in more of the same. We wanted AGA yesterday!

The problem with engineers is they tend to get big ideas that need to be reigned in to produce practical products. Commodore's problem was they let their engineers work on their pet projects without any input from marketing, which itself wasn't doing a great job of reading the market.

Dave Haynie blames Bill Sydnes the 'human bus error' for cancelling the A3000+ etc., but if he had presented an AGA machine worth producing I bet Sydnes would have run with it. You could say it's not his fault that he couldn't get it done sooner, but that wasn't Sydnes' fault. The C65 was 'completed' in 1990? It never should have been started. The engineers working on it should have been helping Haynie and Co produce the AGA replacement for the A500 (A1200?) for release in 1990/91. The high end could be taken care of by 3rd parties. Just keep churning out those A2000's so we stuff them full of cards.

"AA3000+" with buggy Buster is better than ECS Amiga 3000 with buggy Buster. Amiga 3000's ECS wasted the faster 32-bit Chip RAM.

Amiga 3000's Amber has extra video frame buffer memory.

[Promilus]

@hammer

despite owning the MOS 65xx CPU technology.

They bought that technology and didn't really continue to develop it further. And it still lives through Western Digital Center - both pure hardware and FPGA IP. And it was WDC which made 16bit variant. So it's obvious who just tried to milk the cow...

As for using decent CPU for 3D acceleration - actually it was proven as viable solution in Verite from Rendition. Obviously quirky drivers and proprietary API didn't help but that was basically how CAD was accelerated before. With a basically on-GFX coprocessor doing all kinds of vertices calculation and line drawing instead of CPU. It was briefly replaced by fixed function blocks but at some point it became obvious developers need some more control over rendering pipeline. And that's how history made a circle and GPU itself controls all vertices, clone them, move them around, apply textures and shader operations... CPU basically controls the data flow to the CPU and supervise the whole process.

[TCD]

Originally Posted by Promilus

But for a CD - hardly any developer made some good game for CDTV.

That's partially because it was marketed as a multimedia device (think CD player+) and while it could play games it was more of an afterthought. So while you could get an encyclopedia on a CD, there was not much incentive for people to get the device when all they wanted is to play music from a CD. The buzz around CD and multimedia was strong in the early 90s, but the actual implementation and software support was not.

[TCD]

Originally Posted by Bruce Abbott

The A500+ had the latest chipset and OS from the A3000. It came with 1MB standard, expandable to 2MB via the trapdoor slot - 4 times more ChipRAM than an OCS Amiga. That meant more graphics, more sound, bigger bitmaps etc., for better games and more sophisticated apps. Why would you not want that?

And then you got the machine (mainly for awesome new games that would use those awesome improvements) and the software (mainly games) was written for the older model and hardly any games used those new features. So why did developers not jump on the new machine as a standard? Ah yes, because people didn't upgrade. Because there was no real reason if you just wanted to play games. Like TEG mentioned earlier Commodore should have promoted the machine with some software (not just the OS) to show off what it can do that the A500 couldn't. Give people a reason to upgrade.

[Promilus]

@TCD - so... should they sell it as CD console it suddenly WOULD be a good idea? Because that's the general difference between CDTV and CD32 in reference to existing OCS and AGA machines respectively. And Bruce is attempting to prove that despite CDTV being a flop CD32 wasn't because... arguments you MIGHT apply to CDTV as well.

[TCD]

Originally Posted by Promilus

@TCD - so... should they sell it as CD console it suddenly WOULD be a good idea? Because that's the general difference between CDTV and CD32 in reference to existing OCS and AGA machines respectively. And Bruce is attempting to prove that despite CDTV being a flop CD32 wasn't because... arguments you MIGHT apply to CDTV as well.

The CDTV was a A500 with a CD drive. So yes, sell it as a CD based games console that also can run multimedia stuff and not the other way around

Edit: If anybody can find a price for a standalone CD player from 1991 that would be good to have.

[StevenJGore]

Just adding my own pointless thoughts to this pointless thread.

I remember the crushing disappointment back in the day when I bought my CD32 with FMV module and then slowly realised that 90% of CD32 games I could have just played on my A1200.

There were basically 3 categories of CD32 games:

1. Shovelware with no enhancements, and some weren't even AGA.
2. A1200 AGA games with a token rendered intro and a few CD audio tracks. Maybe a few enhancements over the AGA floppy version if you were lucky.
3. Games that showcased the CD32 and were designed from the ground-up for the CD format.

Games in category 3 were basically Microcosm, and a whole raft of promised games that never made it to release, like Inferno. I bought my CD32 for this category.

My next console after the CD32 was the PlayStation. It wasn't a case of Sony just being in the right place at the right time, they actually knew what they were doing in terms of development, marketing, games studios, etc.

[TCD]

Originally Posted by StevenJGore

Games in category 3 were basically Microcosm, and a whole raft of promised games that never made it to release, like Inferno. I bought my CD32 for this category.

I think that's a very good summary of the main problem of the CD32

[TEG]

Originally Posted by StevenJGore

Just adding my own pointless thoughts to this pointless thread.

I remember the crushing disappointment back in the day when I bought my CD32 with FMV module and then slowly realised that 90% of CD32 games I could have just played on my A1200.

There were basically 3 categories of CD32 games:

1. Shovelware with no enhancements, and some weren't even AGA.
2. A1200 AGA games with a token rendered intro and a few CD audio tracks. Maybe a few enhancements over the AGA floppy version if you were lucky.
3. Games that showcased the CD32 and were designed from the ground-up for the CD format.

Games in category 3 were basically Microcosm, and a whole raft of promised games that never made it to release, like Inferno. I bought my CD32 for this category.

My next console after the CD32 was the PlayStation. It wasn't a case of Sony just being in the right place at the right time, they actually knew what they were doing in terms of development, marketing, games studios, etc.

As already discussed somewhere in the forum, without the Xor patent troll, the CD32 would have probably found its own public made of younger players who did not experienced the Amiga as a computer.

The CD media was exploding the 880K floppy disk storage limit which was clearly too small for AGA need, cool adventure or platforms games with graphism creativity in the Amiga style could have been developed and flourish.
And it seems the CD32 sale were OK at launch.

So yeah, for Amiga geeks it was not something to buy but if it would be enough to Commodore to survive, perhaps more attractive models would allow to continue the race. Someone posted the CD32 expected roadmap somewhere in the forum but I can't put my hand on.

[TEG]

I found back the CD32 roadmap and did some OCR:

Fall '93
Relative Performance: 1x
CD32
- 68EC020 @14 MHz
- AA, 4 Chips
- 2 MB DRAM
- 1 MB ROM
- 1 KB NVRAM
- 2x Speed CD

Fall '94
Relative Performance: 3x
CD32+
- 68030 @28 MHz
- AA, 4 Chips
- 2 MB DRAM
- 1 MB ROM
- 32 KB NVRAM
- 2x Speed CD

Fall '95
Relative Performance: 20x
CD64-3D
- PA RISC @50 MHz
- RISC +3D, 2 Chips
- 2 MB VRAM
- 2 MB ROM
- 32 KB NVRAM
- 4x Speed CD
- FMV In Software