Was anyone else disappointed with the A1200?

[AestheticDebris]

Originally Posted by Thomas Richter

You buffer 4 pixels horizontally, obviously, not 32. Actually, that's not at all different from the c2p engine, except that your c2p engine in the middle will become a simple memcpy if you want to pre-render in fast. It is still less work.

Four pixels at a time isn't enough to avoid having to write byte at a time to a planar bitmap. You end up having to do C2P even if you're interleaving both operations rather than rendering the entire thing then converting as you blit. To optimize for planar you'd have to do 32 columns at a time and probably repeat the logic for each bitplane.

All of which is just because the kind of rendering needed for a Doom-like engine is just inherently pixel (and column) based, which is why the Amiga struggles with them in the first place.

[Thomas Richter]

Originally Posted by AestheticDebris

Four pixels at a time isn't enough to avoid having to write byte at a time to a planar bitmap.

I'm talking about *chunky* here. Not planar. Chunky pixels for a chunky graphics systems. 4 pixels = 4x8 bit = 32 bit = 1 long word.

[Bruce Abbott]

Originally Posted by Thomas Richter

Hardly. It just means the chip RAM bus bandwidth is too narrow. Another missed opportunity of course.

Changing the goalposts? We were talking about whether the A1200 suffered from not having a chunky graphics mode. ChipRAM bandwidth is another matter entirely.

AGA was already twice as fast as OCS. Increasing it more would involve extensive changes to the chipset and many compatibility issues. That's what they were trying to do with AAA, but it turned out to be a lot harder than they thought. This doesn't faze Amiga fans though, who think Commodore could have magicked up anything in no time at zero cost if only Gould and Ali weren't in charge.

[lionagony]

Originally Posted by Bruce Abbott

AGA was already twice as fast as OCS. Increasing it more would involve extensive changes to the chipset and many compatibility issues. That's what they were trying to do with AAA, but it turned out to be a lot harder than they thought. This doesn't faze Amiga fans though, who think Commodore could have magicked up anything in no time at zero cost if only Gould and Ali weren't in charge.

What do you think could have been created and on what timeline if the entire orginal Amiga team had stayed on and been well funded the whole time?

[Thomas Richter]

Interestingly, other vendors could "magicked" things. They just bought PC VGA chips and added the software to support them.

[Bruce Abbott]

Originally Posted by TCD

It's kinda funny to see this thread drone on about the technical side when the Amiga was commercially dead in the US by the time the A1200 was released.

Why is it funny?

"I was so disappointed with the A1200. It's commercially dead in the US you know. Well no I don't live in the US - but hey, if I did...".

The Amstrad CPC6128 was initially launched in the US in 1985. Total disaster. All those people in other countries who bought one then bitterly complained about how disappointed they were that it failed in the US - not!

[Bruce Abbott]

Originally Posted by Thomas Richter

Interestingly, other vendors could "magicked" things. They just bought PC VGA chips and added the software to support them.

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?

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

[oscar_ates]

Why they could not change/improve anything on AGA chipset in 2 years when they released cd32 after A1200 and only could shoehorn a cd rom drive on top of the existing AGA?

[Bruce Abbott]

Originally Posted by oscar_ates

Why they could not change/improve anything on AGA chipset in 2 years when they released cd32 after A1200 and only could shoehorn a cd rom drive on top of the existing AGA?

They were working on the CD32 even before AGA was released. Not changing/improving it for the CD32 was a deliberate decision made to maintain compatibility and not slow down development.

Some quotes from the book Commodore the Final Years by Brian Bagnall:-

The obvious difficulty Commodore had selling its CDTV units gave Mehdi Ali cold feet... On July 6 1992 he ordered Jeff Porter not to order any more parts for... CDTV units, promising that they would revisit it in August once Porter found the required cost reductions.

Porter was able to to cost-reduce his CDTV-CR to $281.29... He also produced a cost estimate for an AA version for $325.99 called the CDTV-2... says Guy Wright "The CDTV-2 has different chips that never appeared in any other Commodore product up to this point. They really did improve the Agnus and Daphne and whatnot."

No details on exactly what those improvements were, perhaps it was just AGA with CDROM support. But it was still a CDTV so...

On Monday. August 3, 1992 Porter met with Ali and presented his plans. The basic CDTV-CR without mpeg was deemed inferior to the Philips CDi while the AA version was too costly... ALi laid off the CDTV staff... Gail Wellington... would join Philips on the CDi project as a marketing manager.

For their part, Commodore's engineers were sorry to see the cancellation of CDTV. "That was a pretty good product," says Greg Berlin. "... but things weren't going so well at that point. And it cost money to get into production. As I recall, that was the key issue."

Years later, most Commodore engineers look back on the CDTV as a missed opportunity... "It was a beautiful idea...", says Michael SInz, "... a great idea ahead of its time"...

CDi ended up dominating over CDTV, mostly due to its movie playing abilities... However, it only sold around one million units and was a commercial failure, losing Philips approximately a billion dollars in the process.

In light of that result, perhaps Mehdi Ali made a wise decision...

No kidding. Commodore couldn't afford another boondoggle.

So if CDTV was dead how did the CD32 come about?

At the exact same time Mehdi Ali closed down the CDTV division, another CD-ROM product rose from the ashes. This time the company would make sure the product made sense commercially right from the start...

...the executives did something Commodore UK's David Pleasance thought they should have done long ago. They asked the UK's Kelly Sumner what product the UK developers wanted. "So in mid to late summer, we began to talk to several of the key [games] developers in the UK, saying "we have this technology, what do you want?" The consensus came back - stick with games, it's gotta be CD-based, it's gotta be 32-bit, but the key point was the price. We had to have a price that was competitive with 16-bit technology today," says Lew Eggelbrecht...

On June 27 , 1992.... Jeff porter made a first attempt to cost out a stripped down CD Game System, which he called CD stripper. The new system would have no keyboard, no disk drive, and limited expansion... a whopping 8MB of RAM... the total bill of materials for the AGA system came to $232.67.

On September 4, Jeff Porter began writing a formal product overview for the system, which he dubbed the Amigo...

In order to speed up development on the system, much of the technology... would be inherited from George Robbins' A1200. "The chief engineer on each project had a lot of discretion to build what made sense," says Porter. "So for the CD32, we took an even cheaper CD mechanism... and wrapped an A1200 around it"....

With the preliminary spec ready, Porter and software engineer Allan Havemose traveled to the UK's leading game houses: Ocean, Psygnosis and Gremlin... "it's probably the first time we went to the developers and end users and we really tried to find out what they wanted..." says Eggelbrecht.

One of the biggest changes came from the expensive... 8 MB of memory. When the developers found out it added over $50 to the retail price, they settled for 2 MB.

Commodore's software engineers set the software strategy for the Amigo. They wanted all games and apps to work across all Amiga platforms with CD-ROM. Amigo titles would also work with any AGA Amiga...

"By late October, we were off and running," says EggelBrecht.

[grond]

Originally Posted by Bruce Abbott

AGA was already twice as fast as OCS.

But not faster than ECS. And using your style of argumentation: ECS was so much better than AGA because ECS allowed fewer bitplanes which means that the chipmem was relatively faster!

[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?

[TEG]

Originally Posted by Bruce Abbott

One of the biggest changes came from the expensive... 8 MB of memory. When the developers found out it added over $50 to the retail price, they settled for 2 MB.

So it was all or nothing? Why did they tough in B&W?

The PS1 will be 2 MB RAM + 1 MB VRAM. Why did they not put the equivalent in the CD32, let's say 2 MB of FAST + 2 MB of CHIP?

It was a console so all the problems we discussed previously in this thread about how putting FAST on board of the A1200 would cause compatibility headache to expansion boards, is not pertinent here. Decisions as incomprehensible as ever.

[MortenSkov]

I agree it would have made sense to add 2mb of fast to the CD32 at launch. It would have add a little cost but Commodore would also get license fees from games sold on the system. I know they needed the money, but it would have made for a much more competitive system.

It also makes sense when looking at the CD1200 which needed a trap door card anyway. I believe it would have been a very popular product in Europe - I certainly would have bought it - and it could come with 2 mb fast as standard.

That way Commodore would point directly to an interesting upgrade path? Want to play great CD32-games on your A1200? Buy the CD1200, play the games, play music cd's and make your A1200 twice as fast.

[oscar_ates]

Originally Posted by TEG

So it was all or nothing? Why did they tough in B&W?

The PS1 will be 2 MB RAM + 1 MB VRAM. Why did they not put the equivalent in the CD32, let's say 2 MB of FAST + 2 MB of CHIP?

It was a console so all the problems we discussed previously in this thread about how putting FAST on board of the A1200 would cause compatibility headache to expansion boards, is not pertinent here. Decisions as incomprehensible as ever.

There is additional 512kb RAM for the sound in PS1, so total memory is 3.5MB.

Also, CPU is 33MIPS, 3d polygon chip 66MIPS, MDEC chip 80MIPS. 3-4MIPs of the A1200/CD32 is nowhere near, even 68060 CPU.

[TEG]

Originally Posted by oscar_ates

There is additional 512kb RAM for the sound in PS1, so total memory is 3.5MB.

Also, CPU is 33MIPS, 3d polygon chip 66MIPS, MDEC chip 80MIPS. 3-4MIPs of the A1200/CD32 is nowhere near, even 68060 CPU.

Yeah, it was only to have a point of comparison about the memory.

[Amigajay]

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?

[TEG]

Originally Posted by oscar_ates

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

This one was obviously no longer available for Commodore.

[Bruce Abbott]

Originally Posted by grond

But not faster than ECS. And using your style of argumentation: ECS was so much better than AGA because ECS allowed fewer bitplanes which means that the chipmem was relatively faster!

Only on the A3000 and only with 4 bitplanes or less in lo-res and 2 bitplanes or less in hi-res, not that exciting.

The doubled bandwidth between CPU and ChipRAM was just one of the improvements in AGA. With 2x and 4x fetch modes the blitter had more slots free and the CPU wasn't slowed down at all even with 8 biplanes in hires. ECS in 64 colors (EHB) would have been good for Doom if it wasn't for the performance hit due to nearly half the CPU slots being blocked. That's why it's slower than 8 bitplanes in AGA.

One would would think this stuff was noncontroversial, but leave it to an Amiga fan to call the facts 'your style of argumentation'. The argument that AGA was 'not faster than ECS' is silly, but not surprising when the game is 'bash Commodore at every opportunity'. Commodore were the bad guys, getting between us and the Amiga we should have had. Or so the narrative goes.

This sense of entitlement seems to be unique to Amiga fans. The numerous deficiencies of other platforms are just accepted without pointing fingers, but not the Amiga. Why? 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.

[khayoz]

Originally Posted by gilbert

was anyone else disappointed with the a1200?

no.

[Bruce Abbott]

Originally Posted by TEG

So it was all or nothing? Why did they tough in B&W?

The PS1 will be 2 MB RAM + 1 MB VRAM. Why did they not put the equivalent in the CD32, let's say 2 MB of FAST + 2 MB of CHIP?

I don't know what the configuration of that '8 MB RAM' was, but I'm inclined to think it was all ChipRAM - not some Chip and some Fast. My reasons for thinking this are:-

1. 8 MB is a strange total if 2 MB of it was ChipRAM.

2. The A4000 has a jumper to select 8 MB ChipRAM. This suggests they were thinking about it.

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

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.

I had experience developing for both the CDTV and CD32. We needed to play sampled music in the background while loading images from the CD. That meant we couldn't use CD Audio, so the sample had to be kept in RAM which took up a lot of space. But the HAM images only took a second or so to load, and I cached them so while one image was being viewed the next one was already loading. Memory was tight on the CDTV, but on the CD32 we had heaps. It loaded and displayed the images twice as fast even without running the CD-drive at 2x speed - such was the power of the CD32. It was amazing to see our CDTV title working so much better on the CD32 with no changes whatsoever!

Our CD32 title was converted from the OCS Amiga version which came on 5 floppy disks. For the CD32 we changed from synthesized speech to recorded samples, holding enough samples in memory for each lesson. Even with this and AGA graphics there was plenty of memory free. We didn't need more than 2 MB.