Tweet
I have had no problems at all getting FSAA to work on my Radeon8500 - just turn it on and set the wanted level, that's all.. Besides personally I find the Voodoos solution to look a bit "overfiltered". The Matrox solution sounds (and, in the press release papers also looks) much better IMO as it does exactly what Antialiasing is supposed to do: smooth edges, all the stupid "full screen" AA algorithms are just "smoothing" the whole image out as if it was bilinear/bicubic filtered.
And the R8500s FSAA while really looking nice at the 4x quality setting and up, is just too slow to be usable even on an AthlonXP2000+ with a R8500@300/300. Besides I think anisotropic filtering / high LOD settings are much more important than AA, cause you can kind of simulate the latter with ultra-high resolutions, but this won't help NVidia with it's sometime very blurry textures.
To those NVidiots/FanATIcs over there: this is just the common "what I have in my computer now is simply the best, and has ever been and will ever be" syndrome - this is also true for some Matrox supporters, hell even Creative seems to have a handful of such "supporters".
It's absolutely irrelevant if a card gets 85fps or 85000fps, you simply can't see this. What IS important, however, is that the card can deliver about 50-60fps minimum framerate at the highest-quality settingswhich the Parhelia seems to do but the GF4 fails miserably. At best the frame-rate should be a constant 60fps regardless of the quality settings, this is much better than 150fps average with maybe 300fps max but only 29 fps minimum.
Still I'm sure we will find some certain writers for THG that are gonna bench the Parhelia in 320x240@2Bits to obtain a victory for the GF4. And no wonder since they're running a NVidia site since the old TNT times. How's that for being "unbiased" and "neutral"...
P.S.: Another word to the oh so important "culling": There were quite some driver revisions out for both the old Radeon and the 8500 that had major parts of HyperZII broken and thus disabled. Guess what: the performance delta was only a few percents.... The culling technologies didn't really bring a major breakthrough for the GF3, either, just a few percent. What did improve speed a lot however was the optimized crossbar controller on the GF4 and the Parhelia seems to have a similar if not better implementation.
Hellbinder doesn't have a clue anyway, he's just another man without head (and the important parts inside, as it seems). Who has a clue however, is Humus and his words are the only correct ones: It's useless to do any speculations until there are any real-card benchmarks out.
And the R8500s FSAA while really looking nice at the 4x quality setting and up, is just too slow to be usable even on an AthlonXP2000+ with a R8500@300/300. Besides I think anisotropic filtering / high LOD settings are much more important than AA, cause you can kind of simulate the latter with ultra-high resolutions, but this won't help NVidia with it's sometime very blurry textures.
To those NVidiots/FanATIcs over there: this is just the common "what I have in my computer now is simply the best, and has ever been and will ever be" syndrome - this is also true for some Matrox supporters, hell even Creative seems to have a handful of such "supporters".
It's absolutely irrelevant if a card gets 85fps or 85000fps, you simply can't see this. What IS important, however, is that the card can deliver about 50-60fps minimum framerate at the highest-quality settingswhich the Parhelia seems to do but the GF4 fails miserably. At best the frame-rate should be a constant 60fps regardless of the quality settings, this is much better than 150fps average with maybe 300fps max but only 29 fps minimum.
Still I'm sure we will find some certain writers for THG that are gonna bench the Parhelia in 320x240@2Bits to obtain a victory for the GF4. And no wonder since they're running a NVidia site since the old TNT times. How's that for being "unbiased" and "neutral"...
P.S.: Another word to the oh so important "culling": There were quite some driver revisions out for both the old Radeon and the 8500 that had major parts of HyperZII broken and thus disabled. Guess what: the performance delta was only a few percents.... The culling technologies didn't really bring a major breakthrough for the GF3, either, just a few percent. What did improve speed a lot however was the optimized crossbar controller on the GF4 and the Parhelia seems to have a similar if not better implementation.
Hellbinder doesn't have a clue anyway, he's just another man without head (and the important parts inside, as it seems). Who has a clue however, is Humus and his words are the only correct ones: It's useless to do any speculations until there are any real-card benchmarks out.
)
Of course that level could be an extreme worst case scenario for the kyro2(all the grass and trees and transparent textures and almost no overdraw), but i still think its impressive that a G400max can run it(well it doesn´t exactly fly, but considering its age, its still impressive).
Comment