Talk:Amiga Advanced Graphics Architecture

AA versus AGA?

Do you have a source for saying that AA was a different version of AGA? I've always heard them as being different names for the same chipset - e.g., see [1]. Mdwh 02:45, 12 March 2006 (UTC)

It's the same chipset. AA and AGA are just different acronyms for it.

Source for AA/AGA

To all those who want to revert my creation of 'AA chipset' as a separate article from AGA:

AA is different to AGA. The CD32 had the AGA chipset. The A1200, A4000 and A4000/030 had AA. I quote from Amiga Format Annual 1993, page 29: "For starters, we would expect the 'double-A' chip set, which has just appeared in Commodore's new Amiga 4000, to filter down to the base model... The new chip set is, in theory, just as cheap as the current one to fit."

From Amiga Format Annual '94 page 74: "The Future Entertainment Show in November 1992 sees the UK launch of the new 32-bit Amiga A1200. It looks like an A600 with the numeric keypad restored. Inside it's a different story altogether. It's faster, more colourful and marks the move into a new era of Amiga technology with the Double-A chipset... The graphics have been considerably advanced with the inclusion of the new Double-A chipset. These add a 256-colour mode and a Super-HAM mode with over a quarter of a million colours on screen. The palette of available colours has moved up from 4,096 to an amazing 16.7 million."

People who say that the A1200 had an 'AGA' chipset are correct if they use the acronym 'AGA' to refer to certain extra capabilities, which represented a clear cutoff between 'old school' and 'new school' around 1993. AA included the AGA capabilities. But it was *not* the same silicon as found in the A600, it had extra modes, and it had twice the bandwidth. If you want to put 'AA chipset' under 'AGA', then I can accept that, but don't just revert it - Richardcavell 05:10, 12 March 2006 (UTC)

I'm not sure if there is a typo in what you are saying about the A600, but the A600 had nothing to do with the AA/AGA chipset, it used ECS. As Mdwh mentioned above, AA was renamed AGA to avoid references to other famous "AAs". — Pixel8 11:42, 12 March 2006 (UTC)

I agree with Pixel8. In the first quote, it's not clear what they mean by "base model" - if this was just when the A4000 had been released, it may have been the A600. In the second quote, it clearly says that the A1200 had AA. I'm confused as to what you're saying - firstly you said that A1200 had AGA not AA, now you say it has AA and not AGA.

As for the CD32, it did have a slightly different chipset, but this was more advanced, not less, due to the addition of a new chip (Akiko) for chunky to planar conversion. So not only was this distinction not marked by AA versus AGA, but also it is incorrect to say that the CD32's chipset was less advanced.

The timescales are all wrong here too. The A1200 was released in 1992, not 1994 as you say in AA chipset. The CD32 came out in 1993, so it's rather confusing if a "4th generation" chipset came out before the 3rd generation! Mdwh 16:08, 12 March 2006 (UTC)

Just another piece of information: In german 'AA' is a colloquial term for stool.... The name was most probably changed to 'AGA' to avoid ridicule on the german market.--Qdr 08:46, 12 August 2007 (UTC)

I'm wrong

Hi, guys.

I'm wrong, you're right. I did some further research. The Amiga Format magazines were using the terms interchangeably and I got confused. Someone give me a 'moron' barnstar. Someone else please undo all the separation of AA/AGA that I've been up to. (But keep the technical details of AA, that's all correct). - Richardcavell 02:39, 13 March 2006 (UTC)

NPOV

The problem with this article is that it does not take a neutral tone in this paragraph:

Unfortunately many opportunities to add further improvements, which would have made the chipset more competitive, were missed. Apart from the graphics data fetches the chipset still operates on 16-bit data only, meaning that a lot of bandwidth is wasted during register accesses and copper and blitter operations. Also the lack of a chunky graphics mode was a speed impediment to graphics operations not tailored for planar modes. Over all the AGA chipset was a basic evolutionary upgrade that could not compensate for years of technical progress that the Commodore Amiga took little part of.

Certainly, the above opinion is true, but it's the fact that it's written as an opinion that is the problem. It needs to be worded neutrally. As it stands, the above is not factual. If you can find a significant person saying the above as their opinion, that could be cited.

Technical leadership

It's a fact that the Amiga lost its technical leadership (obviously this happened at some point, since clearly no Amiga compares to today's computers, for example). I was saying it's POV/speculation to say that this happened when AGA was released, or due to its release (even if AGA was revolutionary for its time, the Amiga would still have ultimately lost its technical leadership due to Commodore going bust and the end of Amiga development).

Also on this note, I think it would be worthwhile to say how AGA compared to PC graphics of its time, rather than just saying it was or wasn't as good? (Was SVGA out in 1992? If so was it standard in all PCs or was VGA still common? Did even VGA have advantages over AGA?) Mdwh 00:21, 28 August 2006 (UTC)

Why didn't commodore implements feature XX?

I removed several sentences which tried to explain why commodore did not implement chunky modes or 32bit blitter in the aga chipset, because it looked very much like a POV.

Just to be clear: Adding 32 bit blitter/copper operations could have been done in a purely transparent fashion just as it was done with the graphic fetches. There is no argument not to do this due to ECS compatibility. It would have have added two additional weeks of development time. Same goes to the chunky modes. --Qdr 08:32, 12 August 2007 (UTC)

Backwards compatibility with Amiga 500?

I have the vague impression that the AGA chipset meant that some Amiga 500 games did not work on the Amiga 1200. (Is this why some games are "AGA fixed"?) I came looking for information about that in this article, but found none -- could someone add a word about that backwards compatibility? -- Mecandes (talk) 14:55, 11 February 2008 (UTC)

AGA is backwards compatible with whatever's in the A500, but that does not necessarily mean a game will run on a A1200. The A1200 has a different OS/firmware/CPU than a A500. Using utilities like the "Degrader" one can hide some of these differences, so most A500 games should run. --Anss123 (talk) 15:10, 11 February 2008 (UTC)

And more significantly, this was during an era when 'hitting the hardware' and making use of tricks and undocumented features was standard practice for games programmers. If a game was programmed that made certain assumptions about the hardware it was running on, which were no longer true when running on an AGA machine, then it would fail due to those assumptions. The AGA chipset is one of differences between the A500 KS1.3 (the most popular model before AGA), the others being RAM size (and chip ram size), RAM type (chip vs fast vs slow), kickstart and CPU. Anyone of those factors can cause software to fail. Take for example the RAM. It was not uncommon for the assumption that if there was 1Mb RAM, then it would be 512Kb slow and 512kb chip (the most common arrangement in the A500), this assumption failed even with the small change in some models to 1Mb Chip, so some software would not run. 'AGA fixed' is often used loosely to encompass all the differences between the A500 KS 1.3 and the A1200. However, I agree it would be interesting to have some discussion of the particular differences between the chipsets which might cause problems (clock and timing issues, register sizes, maybe?). 60.240.207.146 (talk) 12:58, 30 March 2011 (UTC)