..one of the guys from forum.3dfx.cz has just finished it,you can find it here: http://havli-3dfx.ic.cz/benchmark-2-1.php

very nice complete comparison, in my opinion only thing that's missing is the 6000, but I got the impression that it was just for cards that were officially released.

But there are also a V4-4800 and a V3-3500SE in there, those cards were never released.

ultima wrote on 10.03.11 at 13:04:21:

only thing that's missing is the 6000

that is kind of my fault,as I was supposed to provide the results for the 6000 (I have almost an identical test system),but I have way too much stuff on my hands right now.let's hope I can make it 'till summer.

Nice results, was suprised by the Kyro 2 doing so well.

I wasn't so much suprised by the Kyro II, back in those days the Kyro II had something nVidia still didn't use and that was tile based rendering, with hidden surface removel.

Result of that was, that even tho the Kyro had waaaayyy less throughput and bandwidth, it used it for almost 100%, in contrast to nvidia, who was still in the "throw a lotta raw power against it" and did a lotta unneccessary rendering, of parts you'd never see on screen.

It was well known that the Kyro II was in a lot of cases faster then the Geforce2 GTS and Pro, only thing it lacked was hardware T&L, which made it crap out badly in tests that supported that.
In those it was often slower then even the Geforce2 MX.