Concerning the initially claimed 16x sample MSAA capabilities for the DeltaChrome I have the feeling that S3 somewhat has overestimated the bandwidth requirements for such a high sample number. Why they still haven't opted for say 2x or 4x MSAA instead of 2x SSAA is beyond me.
Irrelevant of IHVs I'd say we might see one or two generations with maybe 8x MSAA or around that and then they might start moving to more exotic algorithms like fragment AA or whatever allows high sample AA with minimal bandwidth costs.
Irrelevant of IHVs I'd say we might see one or two generations with maybe 8x MSAA or around that and then they might start moving to more exotic algorithms like fragment AA or whatever allows high sample AA with minimal bandwidth costs.