Sacking of Los Gatos engineers point of no return?

[psxphill]

The only reason the A500 was in West Chester was because the Los Gatos team rejected the "Fat" architecture as unworkable. They were wrong. And you clearly have not the slightest idea of how custom chip development is done, or how long it takes.

I've been playing with an A2000 recently and one thing I've been pondering is why the FAT agnus wasn't always 1mb chip ram, rather than 512k chip + c00000.
 
Was the FAT agnus a copy and paste of the thin agnus (plus stuff) and it just took time to extend the dma with extra address lines?

[psxphill]

Yes it is sad, Commodore UK could have made the Amiga a success and it would probably still be around today,

Commodore UK had more marketing sense than CBM, which isn't hard. They also had stacks of money from selling A500's by the truckload, to people who mostly used them for playing pirate games.
 
If Commodore UK had continued the way they were going then it would only have been a temporary stay of execution. As Commodore collapsed the games developers were already leaving for the new CD consoles (hoping for less piracy).
 
There wasn't much room for serious software either, there were a few niche markets that did ok for a while but you have to be pretty serious to drop a couple of grand for a Toaster/TBC/etc.

[psxphill]

but C= UK NOT getting their bid in is one of the few things that went right in the past 20 years of Amiga.

I don't know if I'd go that far, I think they'd have done a better job than escom did. Although it was nice to be able to walk into a high street and buy an amiga monitor, lugging it across town was less fun .
 
I think commodore uk could have done a better job in charge of R&D than cbm did as well.
 
They didn't have enough money to get the Amiga through 1995 when PS1 & Windows 95 owned the market.
 
I personally think that ECS was a waste of resources, FAT agnus should have allowed 2mb of chip ram from day one and then they should have moved straight onto something that allowed chunky pixels and up to 8mb chip ram. I believe it would have been possible if it hadn't been for all the bad management decisions.

[psxphill]

It would be interesting to know whether Commodore was taking about Chunky designs at all seriously before that.

Probably not and that is why they failed.

[psxphill]

What pre-1994 games would have benefited from a chunky mode?
 
By the time Doom is out, Commodore is basically a done deal.
Unless we are figuring that Team 17 is going to beat ID software to the whole "first person shooter" thing, its purely theoretical.

It is theoretical, but the only way that commodore could have succeeded would have been to head off the move to PC and console gaming.
 
The only way they could have done that would have been if they'd produced a platform that was good for texture mapped 3d games.
 
By not realising this, they kept plugging away at 2d hardware because it was selling well. Something like hombre should have been out instead of AGA. However Commodore were used to it being cheap to produce and didn't invest.
 
They needed to produce something game changing, like the Amiga was compared to the c64. Nobody in control had the appetite or vision.

[psxphill]

With a decent blitter and some VRAM, all of them

The original blitter would have worked more or less with chunky pixels, it would be just like having a 2 colour display. You'd have to be careful not to set a stupid shift when blitting, although that would be an interesting effect. Moving the blitter to 32bit would have made it more efficient.

 
AGA chip ram bandwidth would probably have been good enough, although it would be better if you could have enough bandwidth for dual playfield 8 bit or single playfield 16 bit without saturating the bus.

 
020ec would be good enough if a simple texture mapper was implemented in the blitter, you could use line draw hardware as the basis for the edge tracing and you just need a small texture cache and a line mapper.

 
The extra bandwidth would allow for high res productivity screens for the big box amigas & the 3d texture mapping would be useful for fast test previews in modelling applications.

 
Doom on the PC probably wouldn't have happened. It was just an expanded tech demo to show off what you could do with a fast processor and chunky pixels anyway.

 
If CBM had started this after the A500 was launched then they should have been able to ship it by 1991. It would have been expensive to make at first, but the money from the a500's would have made up for this. By positioning the cheap version as a cd based console that could have a floppy/keyboard added and sold it at the price of a megadrive + 32x + megacd then it would have made it more attractive to the masses.

 
So basically no CDTV/A500+/A600/A3000 & all the other projects that got canned.
 
Apart from new hardware there should only have been cost reduction changes made. So switch to surface mount etc, but keep all the expansions the same. This would have allowed processor & scsi cards etc to get very cheap.

 
The only other thing they would have needed was good marketing & not wasting money.

[psxphill]

pre '94 games improved by chunky graphics"... Without some VRAM the improvements would have been minimal...

Doom wasn't the first game on the pc to take advantage of chunky graphics. Wing commander came out in 1990, commanche in 1992, x-wing in 1993.
 
You needed lots happening on screen to make a 2d game look good, an immersive 3d games can get away with less.
 
Doubling the chip ram bus by fitting double the number of chips as on AGA would have probably been enough bandwidth ( i.e. 8 x OCS bandwidth ). Also a dedicated texture mapper would be much quicker than using a CPU, so the requirement of huge bandwidth is lessened.
 
VRAM wouldn't help with effects like dual playfield.

[psxphill]

Chunky allows in a single memory write what takes O(biplanes) number of write in planar, ie a 256 colour pixel write on planar takes 8 memory writes... But only 1 on chunky. This is why I feel AGA was really a poor upgrade

Worse still, while it's common for byte writes to 32bit ram to be handled by the bus using address strobes. To write to a planar pixel requires you to read all 8 bitplanes, modify them and then write them back.
 
It wouldn't have been completely insane for the original amiga chipset to use chunky graphics either. 32 colour mode would have been 8 pixels in 5 bytes, the blitter could be slightly modified to deal with that.
 
As commodore never had the appetite to move from planar graphics, it was basically los gatos that sealed the amiga's fate.

[psxphill]

@desiv
 
Commodore (engineers at least) knew that chunky was needed. Both Hombre and AAA had it.
But they were not given the funds to revision the silicon.
 
Faster CPU may trump Chunky, but at one hell of a cost.
.....
 
I don't think they needed to fix that, just needed provide hardware far enough ahead of the curve to give programmers something to dig into, and that would not look to dated by the end of the cycle.

The Amiga's strength was always that you could achieve more because of the custom hardware & a modest CPU. By making the clever part yourself you can keep the cost down.
 
Commodore management didn't realise they needed to push the engineers to design something like this, they were too arrogant & greedy. Engineers can't fix the company from the bottom.
 
To have survived the playstation, you would need to have gotten reasonable hardware out to developers before the initial Playstation dev kits got out. So we're talking 1992 at the latest.
 
Commodore probably couldn't have stopped the playstation & PC from being huge, but there would definately have been a viable market for games. More viable than the Sega Saturn for instance.
 
Between 1987 & 1992 it would have been possible to design hardware that could have survived. Even as early as 1990 it was obvious what was going to happen but it was too late by then.
 
Whether they used vram or not is the least of the problem. Thats just a small implementation detail which you'd decide based on price/performance vs other alternatives.
 
Amiga was the hare and the PC was the tortoise. Commodore didn't know what to do with what they'd bought. They knew how to make things cheap, but had lost their hunger to dominate when Jack left.

[psxphill]

i'm with you on that desiv. my ideal? 020/28, chunky and some fast. backed up with hd floppy. we could have got away with that...skin of our teeth

I'd probably trade the fast memory for a texture mapper in the blitter & CD drive as standard with HD floppy & RAM as optional extras. Unless the chip ram bus was too saturated at low res, but that would be very bad news.