i see strong results in 3dm00 with v56k, but mediocre in 3dm01. people and systems are same. looks like cpu bottlenecking
Hm... OK but i like to see a 6K on a 3.3V AGP Board hitting over 5000 in 3dmark01. AGP2PCI Adapter for using a IntelCore Or something else with a old Voodoo is like cheating.
My (Till today) unbreakable Wall is hitting 20.000 Points in 3dmark99, 10.000 Point in 3dmark00 and 4800Points in 3dmark01 with a regular 3.3V AGP Mainboard. And without Driver Cheating. (LOD- Very Ugly) And i don't know a 6K that scales beyond a 55K with such Setup. The Limit is Not the Voodoo, But the Processor.
The problem using a KT333 board with 6000s is the CPU using 3DMark2001 since T&L must be done via SW for such reason the scores between a 6000 and 5500 is very similar but if you use an agp2pci adapter in same platform with a powerfull C2D or betteer, for example an ASUS WS with PCIx device and test ast same setting a V5-5500 and V5-6000- 6000s is arround 1750 point better than V5-5500 on 3Dmark2001.
Is there any limitation? Can I use a modded V5 5500 card? For example a V5 5500 moded AGP4X, 210MHz and 128Mb? I can test it on 3DMark2001... WoW, a lot of years since at FalconFly Forum we 'battle' for getting the best score- TM30 had great scores too and later we started the battles with the V5-6000 and you were to able to check online the scores of everybody...
OK, the CPU is the bottleneck for Software using T&L.
It's not only a Problem for a 6K.
I played BlackMesa2012 on 3dfx and saw a fps increase by switching from Northwood P4 3.06HT to a Prescott 2M 670 using a Voodoo4. In that Game, the Voodoo4 could reach the Level of a 55K, but i cant tell you why? Sure Both run into vram limitations in some scenes.