Disaster! Futuremark won't approve GeForce 180.48 WHQL drivers

[Sampsa]

I have some sad news. According to my source, Futuremark is not going to approve NVIDIA's latest GeForce 180.48 WHQL drivers.

I heard the reason is the PhysX support which is now native in drivers and enabled by default.

This would mean all current 3DMark Vantage scores on X58 motherboard with SLI or 3-Way SLI are not going to be valid and presented in Hall of Fame.

source: Muropaketti.com (Finnish)
http://plaza.fi/muropaketti/futurema...impia-ajureita

What do you guys think?

My opinion:

[T_Flight]

I'd love it if they'd just fix the program so that user's could run it at their native res. I was on the forums lastnight trying to figure out a cause for 3DM06 not running in Vista and it seems like 3DM has really went downhill badly since I was into last. I got so aggravted with the whole mess I uninstalled both of them from the computer and loaded up Furmark and couldn't have been happier. I may never reload them. It seems like it's one string of problems after another anymore.

[Praz]

It's a mess for sure. I understand the reasoning behind disallowing PhysX support as it pertains to the CPU test especially if a single card is used. But this is a completely different situation. FutureMark should code Vantage so that the CPU test will not run if PhysX support is enabled or it runs as an unapproved score. One wouldn't think that would not be too difficult to implement.

[dinos22]

seriously are they stupid or something
surely there is at least one person in that company that can make a filter surely to not have to piss off half the freaking world with their ridiculous actions

its ridiculously easy to make out Physx based runs but Futuremark is playing politics with Nvidia here and nothing more. They are trying to teach them a lesson

i am really ty with this dumbass company

now that ATI has finally developed their own physx-type technology they should let nvidia/ati use and develop physics based performance through GPUs

common Futuremark do something right for a change

[dread77]

It's a mess for sure. I understand the reasoning behind disallowing PhysX support as it pertains to the CPU test especially if a single card is used. But this is a completely different situation. FutureMark should code Vantage so that the CPU test will not run if PhysX support is enabled or it runs as an unapproved score. One wouldn't think that would not be too difficult to implement.

Well, I think that FM launched Vantage in April 2008, though first Nvidia drivers that allowed NV GFX cards to implement PhysX processing was in August 2008.
I am pretty sure that a patch could make things work as you are suggesting..

[TheKarmakazi]

yeah thats piss poor policy to punish people for nvidias physics protocols. (sorry for alliteration there :P)

[T_Flight]

Why not just leave the Physics alone? It's there to boost performance. If they don;t wanna use it, then that's not representative of the performance of the card. Are they just looking for arbitrary numbers, or some kind of actual performance measure?

[Praz]

Well, I think that FM launched Vantage in April 2008, though first Nvidia drivers that allowed NV GFX cards to implement PhysX processing was in August 2008.
I am pretty sure that a patch could make things work as you are suggesting..

Sure they could. I understood and fully supported the original decision to disallow PhysX. But this latest decision if true is the result of nothing more then laziness or a spat between them and nVidia.

[SNiiPE_DoGG]

Why not just leave the Physics alone? It's there to boost performance. If they don;t wanna use it, then that's not representative of the performance of the card. Are they just looking for arbitrary numbers, or some kind of actual performance measure?

the point is that physics runs while there is no graphical load on the GPU and is classified as a CPU test

[Gautam]

Score one for the guy who's got a pair of 4870X2's but only two GTX 280's.

j/k

Considering Futuremark's actions over the past year or so, it's hard not to think that there is more than a little favoritism going on there...

[dread77]

@ Praz... Man, most probably it's BOTH...

[T_Flight]

the point is that physics runs while there is no graphical load on the GPU and is classified as a CPU test

Man, I have been away too long. What has happened to this benchmark? If this is true they have completely destroyed it, and it has nothing to do with Physics, but an issue with their software.

nVidia obviously went with driver level coding to improve efficiency. It's not nVidia's or the user's fault if 3DM can't get their software to function correctly. Ain't no way I'm getting mixed up with these games they are playing. This is BAD news.

Man, this is worse than I thought. I'm gonna have to find a Graphical benchmark that works and can be validated. I could post others I have, but don't wanna be questioned "Is that real?". Dang man 3DM used to be THE benchmark to have and was #1 on my list. It's depressing now.

[bench2die]

still cant even publish i7 3dmark03 runs, and at first people with i7 and nvidia cards couldn't publish scores either.

IMO someone should make a hall of fame here and just with a published orb link that is not using physics be put on the list lol forget futuremark

[Ket]

.48 drivers suck anyway, get the .60s off guru3d, uber fast drivers

[G H Z]

Considering Futuremark's actions over the past year or so, it's hard not to think that there is more than a little favoritism going on there...

If FM chose to approve PhysX, other people would reach the same conclusion. They need to keep the numbers somewhat comparable, or risk an even bigger mess.

[K404]

The worst thing about all of this is that 3DVantage took 2 years to develop and it looks friggin awful...then they made us pay for it

ok...back on topic....aye, a mess. Either side could release a fix.

[Harshal]

Regressive decision.
IF ATI in today's date cant offer graphics + physics via driver (&/or hardware) support its their loss. Why should NV be made to restrict their out in open cards + driver which can handle both??

FM should be pushing the right things and not limiting it. If nothing then they should have a filter for 180.XX+ drivers.

[Charles Wirth]

Its pretty bad that Nvidia is being held back on superior technology because of features that work. We all know they should have come up with a patch months ago so this would not still be an issue.

[ThugsRook]

they should allow physX as a second database of results.
this is futuremarks problem, not nvidias.

[K404]

If it comes down to it....how much difference does FM make to nV (or ATI etc) perception and resulting sales? Who bases their hardware decisions on the HoF?

[cadaveca]

The problem is that as you add cpu cores, the test changes, but because of how it works, it doesn't when you run a VGA. They would have to re-work the test so that it was comparable with all vgas, as the workload would have to vary for every card that has more/less shaders...not only the graphical part of the test changes, but also the workload(as the numebr of planes and such varies already).

Of course, this could be patched, but it would then invalidate all previous scores, as the test would be significantly different.


Now, unfortuantely, nV is playing the same card here...Phys-X cards WILL NOT WORK in Vantage with ATI cards either, due to the nV gpu-phys-X drivers.

So, nV have made it so that only nV benefits from Phys-X. Why should futuremark support this? Why should they have to work to approve blatantly monopolistic policies?

Of course, noone mentions teh ATI side of Phys-X, because almost noone has a real Phys-X card. But I do, so I see the whole story everyone else does not.

[BenchZowner]

The problem is that as you add cpu cores, the test changes, but because of how it works, it doesn't when you run a VGA. They would have to re-work the test so that it was comparable with all vgas, as the workload would have to vary for every card that has more/less shaders...not only the graphical part of the test changes, but also the workload(as the numebr of planes and such varies already).

Of course, this could be patched, but it would then invalidate all previous scores, as the test would be significantly different.


Now, unfortuantely, nV is playing the same card here...Phys-X cards WILL NOT WORK in Vantage with ATI cards either, due to the nV gpu-phys-X drivers.

So, nV have made it so that only nV benefits from Phys-X. Why should futuremark support this? Why should they have to work to approve blatantly monopolistic policies?

Of course, noone mentions teh ATI side of Phys-X, because almost noone has a real Phys-X card. But I do, so I see the whole story everyone else does not.

How 'bout leaving the benchmark test as is, and allow ATi ( AMD ) to introduce and use their GPU PhysX ?
Oh wait... there's something missing...

[zanzabar]

Its pretty bad that Nvidia is being held back on superior technology because of features that work. We all know they should have come up with a patch months ago so this would not still be an issue.

what do u mean superior technology, i didnt know that buying some1 and then not licensing the hardware support was superior, they bought the guy that ported it and now there is nothing for the openCL platform. and its not like physX is enven usefull out side of vantage

also havoc is a much better engine and its more portable, physX is a joke in comparison and should have died a long time ago, even the fluid dynamics physics is better

[cadaveca]

How 'bout leaving the benchmark test as is, and allow ATi ( AMD ) to introduce and use their GPU PhysX ?
Oh wait... there's something missing...

Why can't nVidia get teh original Phys-Xcards to work with ATi cards? Why have they removed this functionality? Futuremark might be a bit more willing to work something out had nVidia not limited Phys-X to thier gpus only...

I mean, I think gpu phys-X is great, I would LOVE to see it legal, but at the same time I cannot agree with supporting something that might potentially be illegal...nV using thier monoply on Phys-X to influence the gpu market...

[JacobF]

There is already alot of work being done on this, 180.48 will be approved really soon.