• We’re currently investigating an issue related to the forum theme and styling that is impacting page layout and visual formatting. The problem has been identified, and we are actively working on a resolution. There is no impact to user data or functionality, this is strictly a front-end display issue. We’ll post an update once the fix has been deployed. Thanks for your patience while we get this sorted.

3dmark05 rigged to favor nvidia?

Originally posted by: Czar
friend of mine pointed this out

http://www.beyond3d.com/forum/...rder=asc&start=100

http://www.hexus.net/content/r...04NzgmdXJsX3BhZ2U9Nw==

Seems that 3dmark uses a feature that is not a part of directx and is only available on nvidia cards, and enabling this feature gives nvidia a boost of 10-16% and actualy sacrifices image quality

Been there before..

Honestly I think 3dMark should no longer be used to benchmark or do what most of the better review sites do and use 01 just for comparison and newer programs for the newer features.. Programs that don't have anyone interest at heart.
 
I hope that anand disables this nvidia only feature in his benches, since its actualy an option and thats a good thing 🙂
 
the performance isnt really the issue its that it is optimised to give nvidia cards a boost and not to give ati cards a boost and does so by going outside of the dx9 frame
 
Originally posted by: Czar
the performance isnt really the issue its that it is optimised to give nvidia cards a boost and not to give ati cards a boost and does so by going outside of the dx9 frame

Certainly appears so.
 
Originally posted by: Zebo
How can you say it's favor when Nvidia needs to use two of thier highest end cards to hit 7229 but Ati can hit 7080 with one card?

my bet is that it is cpu limited once you stick those 2 cards on there, i bet with a more powerfull proccessor the distance would increase.
 
Easily testable i suppose. Apart from 2 cards cost more than most peoples PC's (exaggeration perhaps but you know where im coming from)
 
Originally posted by: Zebo
Originally posted by: Czar
the performance isnt really the issue its that it is optimised to give nvidia cards a boost and not to give ati cards a boost and does so by going outside of the dx9 frame

Certainly appears so.

Even that is not the full issue.
The bigger issue is that Futuremark did NOT include 3Dc extensions BECAUSE they were not part of the DX9 spec, and that was the reason, but then included nVidia specific optimisations that are not part of the DX9 spec either, and those extensions lower image quality (according to Hexus.net's article) but increase speed.
 
Survey Says....Who Cares!?!

This was done last time, I dont think anyone expect anything different. It has been established that 3dmark isnt a great benchmark anyway. I really think that HardOCP has the right idea w/ benchmarks.
 
It's worth noting that Nvidia's feature may not be part of the DX9 spec, but it is used by games, most notably Doom 3(listed as Nvidia_shaadow_acceleration, or something like that).
 
Originally posted by: ViRGE
It's worth noting that Nvidia's feature may not be part of the DX9 spec, but it is used by games, most notably Doom 3(listed as Nvidia_shaadow_acceleration, or something like that).

true but doom3=opengl while 3dmark05=directx9+nvidiastuff
 
Originally posted by: Czar
Originally posted by: ViRGE
It's worth noting that Nvidia's feature may not be part of the DX9 spec, but it is used by games, most notably Doom 3(listed as Nvidia_shaadow_acceleration, or something like that).

true but doom3=opengl while 3dmark05=directx9+nvidiastuff
Yea, but the API isn't really important, 3DMark would still be just as useful if it were OpenGL. That said, I'm glad they did put it in, since now we have a very concrete example of the technology, and its IQ/performance effects.
 
Originally posted by: ViRGE
Originally posted by: Czar
Originally posted by: ViRGE
It's worth noting that Nvidia's feature may not be part of the DX9 spec, but it is used by games, most notably Doom 3(listed as Nvidia_shaadow_acceleration, or something like that).

true but doom3=opengl while 3dmark05=directx9+nvidiastuff
Yea, but the API isn't really important, 3DMark would still be just as useful if it were OpenGL. That said, I'm glad they did put it in, since now we have a very concrete example of the technology, and its IQ/performance effects.

3dmark has always been a directx benchmark from the first version to this version and they say so themselves that it is a directx9 benchramk. Dx9 is the arena and they give nvidia a bigger club
 
So...ONE card has an option the other cars has NOT.

Game/Benchmark USES this option and gets beter FPS ---> people scream MURDER 🙂


This is the most ridiculous thing i ever heard.

Of course, developers in a perfect world WILL use (even specific) features available on a certain card - and there's NOTHING to say against it. IMHO they should use ALL and ANY features as long as it improves performance and image quality 🙂

Benchmarking, DirectX standard:

This DST 'feature' is not in the directx standard - ehrm, ok...neither is 3dc....nor does the current ATI h/w supprt full SM3.0 compatibility.

If one game-engine (and even benchmark, for that matter) utilizes these features..i think its GOOD.

The question is NOT really whether how card X performs compared to card Y UNDER THE EXACT SAME CIRCUMSTANCES - because the cards ARE different BY DESIGN already.
An ATI card does not do EXACTLY what an NV card does (on hardware level). ..its different chip-designs and different driver code.

Seeing people 'complaining' about unfairness because one's card feature is used is ridiculous. If ONE card would achieve what ANIOTHER card does - but 10x faster because it has features the other card does NOT have - well blame the game-engine developers for using these features ???????? Come on 🙂 Get real...
 
Wouldn't this mean that the Nvidia card is better? Because it has a function that it uses in games that would make it faster, isn't that what what everybody wants?
 
Back
Top