A comparison between the GeForceFX and the R9700

[bearxor]

This should be intresting.. Tom's Hardware has some benchmark numbers that nvidia provided for a few things... Let's search and see what we find for the 9700 Pro:

UT 2003 Asbestos HQ 1280x1024-32 4xFSAA + 8x Aniso
GeForceFX: 108.4fps
R300*: 71.6fps
Geforce4 TI4600: 39.3fps

3D Mark 2001 Nature 1280x1024-32 4xFSAA + 8x Aniso
GeForceFX: 40.6fps
R300*: 31.6fps
Geforce4 TI4600: 16.2fps

Doom III 1280x1024-32 High Quality
GeForceFX: 49.8fps
R300: ?
Geforce4 TI4600: 20.9fps

*= The R300 chip benchmarks are run here with 4XAA+16X Ansio
?= I don't know if theres actually a way for us to benchmark Doom3 with the leaked version, and I couldn't find any hard data on it. If someone can point me somewhere, I'll update it.

The 3DMark score for the R300 was very hard to find. I just ended up searching the ORB for someone who had AA enabled, since the high scores do not, obviously.

Tom's said the GFfx scores were done by nvidia on a P4 3ghz. The specs of the ati 'puter I found was a P4 2543 with 1GB RAM. Not much else I could find. If anyone can point me to a detailed review of the 9700 with the numbers, I'll update that as well. I threw the GF4 numbers in because they were listed as comparison to the GFfx.

The UT2003 come straight from Anand: Anandtech review of 9700 Pro
The GFfx Benches from nVidia: Toms Hardware
Name of user 3dMark2001SE score came from: Eidolon

But, it looks more like the GFfx is falling in line with the 'about 20% faster' thinking. The UT2003 scores are a little over 30% faster, but we might want to wait for some benches coming from someplace other than nvidia before we decide. Doom3 Benchmark numbers would probably really help here.

 

[Deeko]

Benchmarks now are pointless. Being conducted on non-final hardware with non-final drivers. The scores of both cards could, and in the NV30's case will, change by release. btw, those Doom III scores were conducted on nvtest3 or something like that, so that will certainly run much better on the GeForceFX.

 

[Slappy00]

Yay another ATI vs Nvidia thread... Boy i bet this will be the pinnicle of civilized discussion

 

[Lonyo]

Might the 16x rather than 8X anis have an effect on speed?

 

[NicColt]

>those Doom III scores were conducted on nvtest3 or something like that,

they were ran on a special optomized version of NVDOOM made by nVidia and for which I understand is a stand alone version of a demo. Not even the game.

All this is to be taken with a grain of salt until you see it to believe it.

It was tested on a card no one even has yet, on a processor no one could buy with a game that not even officially released. and now to make things worst were getting paper scores. All these BS paper launches have got to stop.

 

[mrman3k]

I don't care what any marketing benchmarks are shown, what I want to see are real benchmarks from all of the hardware sites and actual customers after they have bought this card!