Acanthus, you call 55fps in a single player game on a $350 board "insane?" That's a respectable score at that setting, it just so happens that ATi scores better. If the 5950U was getting 30fps, then I'd question what Brent was thinking benching at that setting.
Ben, OK, but that last review I linked showed a 9600XT 25% faster than a 5700U at 12x10 no AF. I haven't seen nV take a smaller hit with AF yet (THG's VGA Charts III shows Radeon far in front of FXs with AF), so I'm going to assume ATi maintains (and possibly even lengthens) its lead with AF. I'm just surprised by PCT's claim that his 5950U not just looks but plays better than his 9800P, given the evidence against it.
Again, the IQ improvement may just be down to higher-quality textures/filtering. Brent's first two screenshots do show some differences in the reflections. The car reflection near Max is sort of brighter in the first shot, and the pipes are shinier/use higher-res textures in the first two shots. It's almost as if the 9800P is AA'ing or otherwise filtering those reflections and the 5950U isn't, though it's hard to tell from a static screenshot. It's probably more likely, however, that ATi's AF algorithm is using lower-res MIP-maps for the pipes (because of their "odd" angles). I don't see any differences in texture quality on his coat, though.
PCT's hiccups are also strange, given that he wasn't using AA, and MP2's highest texture setting appears to be just "High," not something like "Extra High." The better IQ is probably down to better AF, and possibly due to more texture memory, but the faster performance and lack of "hiccups" seems strange. Oh well, place this one in the unsolved bin for me. I don't doubt that MP2 looks and plays better for PCT on his 5950U, I just think that maybe there was something amiss with his 9800P to cause it to perform slower than his 5950U.