Originally Posted by
supreet
GeForce 6600 GT
500 MHz
4000 MTexels/s
128-bits
GDDR3
500 MHz
16.00 GB/s
VS & PS = 3.0 (now thats an advantage)
Pixel Pipelines- 8
Vertex Pipelines- 3
Vs
Radeon 9800 Pro 128MB
380 MHz
3040 MTexels/s
256-bits
DDR
340 MHz
21.76 GB/s
PS & VS = 2.0
Pixel Pipelines - 8
Vertex Pipelines - 4
I am little surprised though that 6600GT beated 9800pro/xt in every departments. And thanks chaos for telling me..i am using 9800pro and for god sake never thought that it got 4 vertex units, i always thought its 2 ....and funny thing is i daily used AIDA32 to see voltage values never bothered clickin on GPU....lol..thanks anyway.
Well the big reason why the 6600 is faster is cos the NV4X architecture is very efficient. Clock to clock, the 400MHz NV4X matches a 500MHz R4XX(which is nearly the same arch as R3XX). So a 500MHz 8 pipe NV4X part will obviously own a 380MHz 8 pipe R3XX. But then the 6600GT is not a true 8x1 part. When doing raster ops, it behaves more like a 4x2. Hence its not that much faster. Also it has much more pixel shader power than the 9800 and hence will run newer games better. The 6600GT is rather bandwidth starved though. I'm amazed at what NV is pulling off with such meagre amount of available memory bandwidth.
Athlon 64 3000+, Asus K8N-E Deluxe, 1GB DDR400, BBA 9800 Pro, WD Raptor 36GB, 160 GB 'Cuda SATA, NEC ND-3500A, SB Audigy LS, Altec Lansing 251.
Housed in a Wavemaster Silver.
Bookmarks