Amiga 1200 versus Atari Falcon?

[Hattig]

They both had advantages and disadvantages, they both had poor design decisions, and they were both the final products in the 'all-in-keyboard' style of computers popular since the late 70s.

The Falcon was crippled by the memory bus. The A1200 was crippled by the lack of Fast RAM. One of these was fixable for little cost (£100 for a 2MB RAM expansion in 1993) - and hence the A1200 was better at the time considering it cost a lot less than the Falcon when purchased).

The Falcon might have had better vertical resolution than the video-optimised Amiga chipset, but you would need a dedicated monitor to make use of that. But if you had that, it was better than the poor interlaced modes the A1200 offered - how many people ended up using 800x300 Super72 mode on their Multisync monitors? Productivity mode didn't seem to fill the screen - not for me anyway. Let's not talk about higher colour depths either...

Let's face it, at launch the A1200 needed an A1400 for £599 as the 'sell-up' option, with a 28MHz 68020, 4MB RAM (2+2) and a HD floppy.

[Hattig]

And the last post is being far to modest. A 16 bit color display on a Falcon at 800x600 is going to look better than a 256 color display on an Amiga at any resolution. 256 ciolor modes were outdated when AGA was introduced.

256 colour modes were the standard for years after AGA came out. What was outdated was the lack of higher resolution modes (60Hz+ 800x600 non-interlaced, 1024x768 non-interlaced, etc).

Did the Falcon allow 800x600x16bit? That's 55MB/s of video bandwidth... pretty high for a home computer of that era.

[Hattig]

I'm with MarkTime - I don't think x86-64 is all that bad because it does fix a lot of the problems with x86, especially the number of registers and non-orthogonal instructions. And the implementations run fast and get a very high instructions-per-clock too.

Back in the 80s though, the 68k was king, the custom chipset in the A1000 was astounding, and it was all good. The A500 was all good too, getting the technology into a far cheaper product. But we had 1985 hardware until 1992, with no upgrades apart from ECS! AGA would have been great in 1988, and expected in 1990 - and that's an AGA that had a full 32-bit blitter and ran faster internally. It did come, eventually, and was not what it should have been, and that is Commodore's fault. At least it came out.

A better hardware release plan would have been something like:

1985: OCS
1988: OCS + ECS + 32-bit Blitter + 64 palette entries + 8 bitplane support (incl. HAM-8) and 18-bit colour palette.
1992: Above + 16-bit sound, 8 sound channels, 256 palette entries, more sprites, byteplanes (chunky graphics), alpha transparency in sprites and multiple playfields, etc...

Instead it seems that the engineers got distracted by new chipsets and gave up on the incremental upgrades - probably because that wasn't the done thing in the 80s.

Things like Natami are for hobbyists to get a possible dream 1996 Amiga, with the hardware that could have been.

[Hattig]

Oh definitely, ARM is sneaking up on Intel just as the market is shifting towards tablets and mobile devices. Marvell has quad-core ARMs on the market soon, 2GHz dual-core A9s will be along soon enough, A15 is coming in the future... and they will all be beyond the current 'good enough' performance level.