Anyone have an update on PowerVR and Kyro III or IV?

[Arun]

So, well, it sounds like it's the new ATI&nVidia alliance VS PowerVR now. Woah, Kristof, you guys must have AMAZING tech if they went that far

http://www.theinquirer.net/?article=10952

Could I have been right?
Heck, that story kinda makes sense. But what'd make more sense though would be that they'd say "We won't do more than this in that timeframe in relation to manufacturing costs" - so we don't get another 107M 0.15u monster, and even less a 200M 0.13u one...

Uttar

 

[Ailuros]

Why would 200M be a presupposition for a PS/VS3.0 design? Granted I've no exact idea where those would range, but I really don't think that from shaders2.0 to 3.0, transistors will have to double.

Needless to say that the above assumption doesn't make much sense anyway. It would rather be ImgTec's partner that would set out for a contest in the graphics market. There it rather depends how much the partner is willing to invest and how aggressively he wants to enter the market.

And before anyone asks who that might be, it's the toughest secret in the whole story and probably kept that way for a reason. I wonder if we finally find out about that one, upon announcement. It's one little detail I've been dying to know...

By the way if the Inquirer story should be true, then it's high time both put an end to the constant idiotic refreshes of refreshed parts and shift those wasted resources into next generation products. So where's the opposition to the alliance anyway? Or do you really expect that we'll see refreshes of Series5 every couple of months? Nothing is impossible of course, but in that sense I rather see an alliance of the three in relative terms, since same limitations and drawbacks affect all the market not just those two.

 

[Entropy]

So, well, it sounds like it's the new ATI&nVidia alliance VS PowerVR now. Woah, Kristof, you guys must have AMAZING tech if they went that far

http://www.theinquirer.net/?article=10952

Heh.
Actually, if ATI and nVidia is making agreements, I'd sooner suspect that it has to do with the shift to finer lithography. If the reports of TSMC being ready for 0.09um volume production towards the end of this year is correct, and the rumours of the NV40 and R420 being based on 0.13um are also correct, then for a 18 month product cycle, the new process has to go unused for just under a couple of years. If they move to 0.09 by next autumn for instance, which could reasonably be expected otherwise, their next generation of high-end parts have had life-times of just over half a year. Ouch. So Mike Magee could be right on this one. (Although it would be myopic of them IMHO.)

Though personally I am not sure either of them could be trusted to stick to any agreement, if under sufficient pressure in the marketplace. nVidia and ATI work in largely the same markets, and they have natural reasons anyway to sit down and have a chat about the future of the land they are both trying to control. Where Intel and to some extent Microsoft is heading with various platforms, and how that could be influenced to mutual benefit would be natural topics for instance.

Entropy

 

[jb]

For us games it would probably be a better return on our money we spent if they did slow down a bit. Besides it would be nice to see a "new" feature actually used on my card before the next gen comes along and I upgrade. At least IMHO