Futuremark Preps DirectX 11 Exclusive Benchmark

I know, all I did was set OC to manual and change the BLCK from 133 to 175...if I find something that pushes it, I go higher

 

same here with a PhII X3.. stock runs at 2.8 .. bump the "fsb" to 230 give me 3.5 (3504 iirc) and if I really need more I up the V to 1.475 and bump the multi until I get 3.8Ghz with all 4 cores unlocked.. not bad for a $105 CPU on an $80 mobo ... for that little coin Intel can't touch it now if NV's opena a can of whoop ass on ATI's pricing I can be happy

 

No, their success is based on them showing stuff that's not really available in a given time period. The succesful 3DMarks did just that. You could only run Nature on a GF3 for a while, for example. Vantage, OTOH, outside of being uselessly obfuscated and unnecessarily mysterious about its inner workings, didn't do that great (otherwise Futuremark as a company may be in a better spot than it currently is), mainly because it came pretty late in the game.

If they actually launch at the end of the year, you'll have a fair chunk of DX11 content out, so they'll need to do something special to make it interesting, or just produce another Vantage, and I doubt there are enough paying overclockers around the world to sustain their business model long term.

 

http://www.xtremesystems.org/forums/showthread.php?t=252322

 

Another futuremark benchmark to favor nvidia ? Yay .....

These benchmarks have never been fair or valid. I don't see why there is evne a thread here.

 

You already forget how future mark tended to paint the HD 2900 XT in much better light than it performed in games? Doing nearly as well as the 8800 GTX at times?

The last version is the only one I've seen with an obvious, although I think unintentional, bias due to the inclusion of PhysX.

This time around they have access to both Direct Compute and OCL. Chances are it'll once again favor one vendor or the other depending on which card they use for coding, but at least it won't be like the PhysX fiasco.

Regards,
SB

 

Considering many games favor one architecture over another I don't see why that's an issue as long as it's not excessive as with PhysX in Vantage. It's pretty much impossible to write completely fair software when the hardware it runs on is so diverse. Besides, how do you tell if a benchmark favors one party? We don't have a de-facto reference implementation of a "fair benchmark" to use as a baseline.

 

Yeah, how can anyone discern that NVidia is being favoured?

Aside from that, Vantage appears to very heavily lean on Z fillrate. Games actually worked out much the same way. NVidia decided this was important enough. ATI made R600 with a stunningly slow Z fillrate. So the myth that Vantage deliberately favours NVidia was born.

 

I hope they will somewhat go back to roots and produce a benchmark with a build in DEMO! This way it can be used to stress your system and show pretty pictures + music.
I've ran Vantage on my computer only when changing GFX cards.
I've ran 3DM2001/03/05/06 hundreds of times even on the same card.

It needs to look stunning and give representative score otherwise I'm not buying it anymore.
I only hope that Vantage was born ill because of the development put towards Shattered Horizon. Actually for me S.H. looks at least 10x better than Vantage tests.


EDIT:
I agree with AlexV, timing is suspicious, but not because of nV and rather because AMD/nV! AMD will be close to or releasing RV9xx and nV will be close to or releasing GF100 done right (in theory).

 

Well the timing is appropriate no? Since both parties now have DX11 cards available?

 

Did I get this wrong or is there merely a prerelease demo with their preferred technology partner this time around? When's it gonna be released?

 

That dx11 demo from futuremark just looks like this.
http://www.tauchfahrtzurtitanic.de/

 

I remember pixel shader 1.4 tests not counting towards final scores. I remember the months of fcheating they allowed nvidia to have. I remember physx which nvidia owns being a huge part of the last one.

 

I don't remember feature tests to ever have counted towards the final score (at least since 3DMark 2000). What would make 1.4 special in that regard, especially taken into consideration that it was only introduced in the SE-Version of 3DMark 2001?

 

C'mon stop whining and I dont even own an Nvidia card. Whatever pushes the envelope is fine.

 

I reckon it is because prior versions allowed PhysX to alter the scores?

 

I just hope the next version doesn't bring machines to single digit fps while looking worse than mainstream games. Doing work because work must be done is pointless.

 

Its the fact that it doesn't push the envelope thats the problem. It seems to me that when nvidia is in the lead with new tech , 3dmark takes advantage of it , when Ati is in the lead it does not.

 

In Vantage for example they've been moving everything to FP16 with R600 being designed around single-cycle FP16 throughput whereas G80 wasn't.