Continuing this here (don't want to ruin the thread): http://www.xtremesystems.org/forums/...5&postcount=32. The 500FPS cap limit is worthless if you 'just' recalculate the results; YES, the effect of Nature is gone, but there are still 6 other tests (3 with 10x, 3 with 20x) that are influenced by the graphics card. The only difference recalculating makes is that the difference between low-end and high-end is smaller. I'll give you an example: 1) http://hwbot.org/result.do?resultId=760984 2) http://hwbot.org/result.do?resultId=766493 Cpu speed still is the same, only difference is the power of the graphics cards. If we recalculate the second score (nature > 500), you'll find this: 1) 56939 2) 64208 Reason: the 95GT is consistently slower in each of the six other tests (+ doesn't even hit 500FPS in nature), thus it'll always have a lower score. If you use a limitation that cannot be crossed, you will have to handicap you graphics card for ALL tests. (both systems were almost untweaked - I reckon both had lod on, though) I just realised that 3DM01 allows different vga clocks in each test. THAT makes my solution unworkable for the 3DM01 benchmark. The reasoning still is valid, though, high-end cards will give you an advantage; a cap limit with recalculation just makes the difference less apparent
