NV40 benchmarks

some benchmarks for the nV40 have been released, look...

-38K marks in 3DMark 2001SE

-Halo ( no AA, 8x AF ) 1600x1200 51.1fps

-Far Cry ( no AA, no AF ) 1600x1200 53.4fps

-UT2004 ( 4xAA, 4xAF ) 1600x1200 71.9fps - 1280x960 82.9fps ( both botmach )

-Star Wars Kotor ( 4xAA, 8xAF ) 1600x1200 49.3fps

and these are with non final drivers.

also another quote..

"Even if the R420 is a 12-piped card, it would have to be clocked at 633Mhz to have the same fillrate as that."

And that isnt going to happen.
So it looks like nvidia is going to be ontop again (yay).

http://www.theinquirer.net/?article=14827

http://www.beyond3d.com/forum/viewtopic.php?t=10946

Thanks for the links MM, I love looking at the latest technologies, even though I won't be buying a leading-edge video card any time soon.    I can't wait for the pci express standard to come out. The mountain of used AGP cards on ebay will be good bargain hunting.  After all, yesterdays tech is good enough for me!  

I'll always be an ATI grrl!  Far as I'm concerned they can't be beat.

Nvidia isnt going to make the same mistake twice. Im betting the NV40 is going to be very good. Its totally different from the FXs (hence why they have dropped the FX from the name, i.e. Geforce 6800).

I'm betting that those 3DMark scores are due more to driver hacks than to the merits of the NV40. I highly doubt that ATi and nVidia's new GPUs are going to vary massively in performance. As it is now, the 9800XT and GFFX5950 are fairly close in performance in DX8.1 and lower games (with the NV card being slightly faster), and the ATi cards offer as much as a 50% better framerate in DX9 games utilizing pixel shader 2.0. I suspect this next generation will be the same where one company's card will be good at one thing and the other company's at something else. Preliminary reports of NV40 scoring 12k in 3DMark03 are most likely not accurate, or are due to nV's infamous driver hacking. Never in the history of video cards has the next generation offered a 100% raw performance increase over the previuos generation. I say 100% increase because with non-dodgy drivers, the GFFX 5950 gets around 6k-6.5k in 3DMark03. A score of 12k would imply a near 100% increase. Even a default score of 10k in 3DMark03 is a bit hard to swallow. 8-9k seems more probable. 38K in 3DMark01 seems very unlikely as well.

Well the NV40 NDA has been lifted, and there are no 38K default 3DMark01 benches. I was half surprised to see 12k scores in 3DM03, but I'll wait until a reputable review site analyzes nV's drivers- remember the NV30/35 was able to pull 8K default benches using cheat drivers. Several of the reviews have already noticed low IQ anomalies in certain games (UT2004, Far Cry), and it wouldn't be surprising if something flaky was going on in pre-set benchmarks like Aquamark and 3DMark. Really, 3DM03 and Call of Duty are the only apps where the NV40 has a near 100% lead at default. Otherwise, it takes very high resolutions for the NV40 to pull way ahead, which is good for futureproofing. The card looks good, but not the price. I think anyone would be foolish to outright buy this card now when we'll be able to see how it compares to ATi's offering in a month.

There are now some comparisons between the X800 series cards and 6800s. It seems that the two top end models of the cards (X800 XT Platinum and 6800 Ultra Extreme) are more or less identical in terms of performance and the image quality is also very similar. There are a couple of things to note about he cards tho.

1. 6800s support PS3.0, X800s only support 2.0 with a couple of other features like displacement mapping.

2. the ATI card is clocked much higher than the geforce equivilent indicating the nvidia archtecture is more efficient. The highest clock speed they can get out of either card is 600Mhz due to the 16nm manufactureing process. This gives the Geforce more headroom for overclocking and also later card revisions.

3. The ATI drivers are more mature than the Geforce ones meaning the geforce is more likley to get performance gains due to newer/better drivers.  

Reviews:
http://www.hardocp.com/article.html?art=NjEx
http://www.hardwareanalysis.com/content/article/1710/
http://www20.tomshardware.com/graphic/20040504/index.html
http://www.hothardware.com/viewarticle.cfm?articleid=517&cid=2

Interesing editorial:
http://www.hardwareanalysis.com/content/article/1711/