wrong dude...
with T&L (like 3dmark 2000) it beats an T&L card !!
why ?
T&L is not a true feature!
we can have T&L via cpu it's nothing that kyro can't do !
T&L on geforce (the same for radeon) is processed on the same bandwidth of other tasks, fill rate will drop, if a game is bandwidth limited (most are in decent definitions), kyro will smile at this because the T&L is processed outside the card so there is no competition on kyro's bandwidth, thus kyro fill rate stays the same.
for a good cpu I even recommend kyro 2.
If we enable (hardware) T&L on a geforce (or radeon) on this bandwidth limited games the FPS will drop accordingly.
you could say that a geforce can process 10 million polygones, yap that is a theoretical maximum, geforce in this T&L maximum would not be able to do anything more!
Nvidia/ATI implementation of T&L is not that great. they should have used a different bandwidth path for this "marvellous" T&L.
their's T&L implementation on todays geforces are only for marketing purposes!
on a lower cpu this T&L could even help, but not on a decent cpu.
you could say that The GeForce/radeon boards can use the AGP bus to transfer vertex data without using any local bandwith. But that will kill the bandwidth avaiable from the cpu to the video card choking the speed!
there are many other factors in the future that will help kyro 2...
games with many textures layers , games with more complexity (more overdraw). these games will need more fill rate and kyro 2 have lots of it, unlike geforce and radeon!
if present games with T&L are representatives of future games (within a year from now, I am sure they are) then kyro will beat T&L card in T&L games with decent cpu's ....
I think we will need more fill rate than T&L in the near future . when true T&L games arrive we can get a kyro 3 or whatever !!!
VErtex shaders not Hardwire T&L will be the future...
T&L will be the MMX of videocards... vertex shaders will replace that T&L like 3dnow and SSE did for MMX !!
<P ID="edit"><FONT SIZE=-1><EM>Edited by powervr2 on 05/05/01 03:22 AM.</EM></FONT></P>