3dmark05, the 'drivers' situation

[schuey74]

I don't get it....

go read this article... http://www.megagames.com/news/html/hardware/atinewdriversget3dmark05crown.shtml
which is called "ATI New Drivers Get 3DMark05 Crown", in megagames.com
and tell me if you find it stupid....

since they say that X800 xt pe, gets higher score then SLI of nvidia....

but all of the sudden the say that X800 xt pe is overclocked....
now i don't get how this related to the driver?

if it does relate,
so does this mean that ATI once again is cheating using the drivers?

or probelly the driver brings the Picture quality down.. and performance high...

which for nvidia... it doesn't change!
The high score is a combination of the new drivers and the extreme overclock. How is that difficult to understand?
"so does this mean that ATI once again is cheating using the drivers?" - is this a joke?
Nice assumption on picture quality. Actually, neither Nvidia or ATI are identical to the reference image. And that's what you're suppose to compare them to. Comparing an ATI image to an Nvidia one will tell you absolutely nothing unless the difference is blatantly obvious, like banding for example.

[OWA]

Actually, neither Nvidia or ATI are identical to the reference image. And that's what you're suppose to compare them to. Comparing an ATI image to an Nvidia one will tell you absolutely nothing unless the difference is blatantly obvious, like banding for example.

From a comparison I was just trying to do for something else it looks like ATI is quite a bit different from the reference image in the section I was testing. I'm currently trying to get someone else to test it to make sure I haven't done something wrong. To me, watching it run, it looks like ATI isn't showing all the mist that the creature causes in the Canyon Flight game test so the image ends up looking quite a bit different. You can look at the link below to see the images I'm talking about. Bear in mind, I'm still trying to have someone double-check me to make sure I haven't screwed something up.

http://www.nvnews.net/vbulletin/showthread.php?p=440677#post440677

Edit: The 4.9s showed the same image so ATI and Nvidia are both consistent between their drivers. I don't know. Maybe I'm not doing something correctly.

[neliz]

Bit-tech allready did this for you so no need to do it!

Okay, here is the proof ATi is not cheating with their drivers:

http://www.bit-tech.net/review/359/5

and a quote:
Running the default test, the image quality is noticeably worse than our User Defined test (below) - you instantly see that the water is not quite as clear; it simply does not look as realistic as it did before. This is one of our major gripes with this new version of 3DMark: pixel-processing speed is useless if image quality is poor. We remember back to 3DMark03, where NVIDIA added some optimisations to their drivers in order to return a higher score - these gave optimisation the bad name that it has today. We would have expected image quality to be a big part of this release of 3DMark, but there are many occasions where shimmering, moiré and generally poor image quality are experienced during the default test.

The scenes are rendered fairly similarly, but there are some differences between the way that the water is rendered - this is likely to be down to the fact that the two drivers are rendering with different pixel shader paths, with the ATI driver using PS2.0b and the NVIDIA driver making use of PS3.0. Deciding which looks better can be a little subjective, but the lighting and shadows that are cast on to the airship by the NVIDIA driver would appear to be a little crisper, while the water appears to have slightly more ripple detail when rendered by ATI's driver.

[Morrow]

Bit-tech allready did this for you so no need to do it!

Okay, here is the proof ATi is not cheating with their drivers:

http://www.bit-tech.net/review/359/5

and a quote:

Na, I don't think they cheated but nevertheless don't you find it strange that although ATI had access to 3dmark05 beta program since the beginning of the developping cycle that they were not able to find and fix this bug much earlier?

Why is this "memory allocation" bug suddenly fixed in a swift move after the final 3dmark05 has been released and not some months ago? They most certainly knew about it all the time...

[neliz]

Because of the cycles Catalyst drivers are issued in.
If it's not a showstopper bug then it goes in the normal 10 week catalyst cycle and it ends up just in time.

Keep remembering that it is a bug that affects all AGP cards, if one would just fix it without the proper testing period, you could end up with disastrous results.

not that I mind though, nV got their behinds wooped

[OWA]

Bit-tech allready did this for you so no need to do it!

Okay, here is the proof ATi is not cheating with their drivers:

http://www.bit-tech.net/review/359/5

and a quote:

I'm not sure I agree with their results though. There are some more than minor differences and it doesn't seem to be because of 3.0, DST or whatever. When comparing the mist, ATI is clearly not rendering most of it. Because it was mentioned on another site about the possibility that 2.0b and DST could be a difference, I made another comparison disabling those and had the same results. Nvidia matches the reference image pretty closely but ATI doesn't. Bear in mind, I'm only talking about the mist here.

ati..............................reference image...........nvidia 2.0b, DST off
http://img26.exs.cx/img26/2586/ati_807_gt3_1410.th.jpg (http://img26.exs.cx/my.php?loc=img26&image=ati_807_gt3_1410.jpg) http://img26.exs.cx/img26/9236/ref_image_gt3_1410.th.jpg (http://img26.exs.cx/my.php?loc=img26&image=ref_image_gt3_1410.jpg) http://img66.exs.cx/img66/1492/nv_6681_DST_off_20b_gt3_1410.th.jpg (http://img66.exs.cx/my.php?loc=img66&image=nv_6681_DST_off_20b_gt3_1410.jpg)

[neliz]

You can see by the water that the mist is rendered, it's not visualized properly.
It's an issue with lightsources, in the bit tech article, check the return to proxycon image.
A small lightsource is rendered birght on the nv card and dull on the ati card
http://www.bit-tech.net/images/review/359/IQ-returntoproxycon-large.png
You can see that, with AA/AF on the lights on the top left are rendered differently (much less glow on the ati card) and so are the buttons behind the containers (zoom in once (normal size) to see the difference)

This happens with every IQ comparison, like the roster behind the soldiers, it's rendered dull on the ATi card and bright and fluffy on the nv card.

It might be an option to look into color/gamma settings..

[SlyBoots]

JMO, but the ATi image looks more realistic to me. Even though the nvidia image is closer to the referance they both look 'overbright'...??