Improved AA/AF, Video Decoders, Memory Interface Future GPUS

[jb]

I DON'T think T-AA is the future, or at least not for a long time. it's nice to have for older games, but considering you have to have vsync on, it's not ideal by any means.

Oh yea I was just throwing that out there as an example to show that new AA methods are not dead just yet.

ChrisRay,

I dont think we need better AA quality right now. I think we need better compression and performance values. Even now 4x MSAA performance isnt what I'd call ideal on either hardware. (On performance factors)

Well again if you lock your self into todays AA methods. But there could be other methods out that that give better AA and proformance. Think out side of the box

 

[Tim Murray]

Utty, are you going to give your "Quincunx is actually a good idea" lecture?

 

[ChrisRay]

Hey quincunx wasnt such a bad idea back when 32/64 meg cards were standard. ;p

 

[Ailuros]

Well again if you lock your self into todays AA methods. But there could be other methods out that that give better AA and proformance. Think out side of the box

That's exactly what I did yet I sadly didn't get an answer or prediction at least.

Right now there's just not enough bandwidth available and transistor counts are too critical for single cycle 4xRGMS (or sparced sampled eventually). Both ATI/NVIDIA's accelerators are capable of 2xMSAA per cycle; now double the amount of samples per cycle and it would logically mean 4xAA at today's 2xAA performance or 8xAA (via a second loop) probably at today's 4xAA performance. It would increase both performance and antialiasing quality since theoretically with 8xRGMS/sparced you get an 8*8 EER.

R3xx/4xx's accomplish 6x sparced MSAA by looping 3x times 2xMSAA.

Memory footprint shouldn't be much of a consideration as 512MB ram equipped cards should appear soon enough. The bandwidth and/or transistor counts remain for now probably a headache.