Sign in with
Sign up | Sign in
Your question

Nvidia cripples PhysX performance: True or False ?

Last response: in Graphics & Displays
Share
a b Î Nvidia
April 19, 2010 3:07:29 PM

This is a attempt of w*w.ngohq.com to generate hits with bogus inflammatory info.
I will tell you the problem, the game should not try implement physX when there is no physX certified gpu present. The game played fine without PhysX , so whats the problem ?
Here is note for the game.

PhysX (hardware supported by nVIDIA GPUs only, acceleration and additional effects are exclusive to the Windows version)[3]


Also physxinfo.com came to this conclusion about the web site that made the claim

Quote:
Thus, we see two explanations of original statement:

1) Performance impairment really does exist, but does occur only when non-Nvidia GPU is spotted in the system.

2) Mr. Eran “Regeneration” Badit has provided not very accurate information, either without malicious intent or with it.
April 19, 2010 6:22:46 PM

notty22 said:
This is a attempt of w*w.ngohq.com to generate hits with bogus inflammatory info.
I will tell you the problem, the game should not try implement physX when there is no physX certified gpu present. The game played fine without PhysX , so whats the problem ?
Here is note for the game.

PhysX (hardware supported by nVIDIA GPUs only, acceleration and additional effects are exclusive to the Windows version)[3]


Also physxinfo.com came to this conclusion about the web site that made the claim

Quote:
Thus, we see two explanations of original statement:

1) Performance impairment really does exist, but does occur only when non-Nvidia GPU is spotted in the system.

2)Mr. Eran “Regeneration” Badit has provided not very accurate information, either without malicious intent or with it.


Impairment does exist? Not performance loss, "impairment." Now, where I come from, impairment and crippling are synonyms.

However, they didn't test explanation 1, so.... we don't know. If the scores are significantly lower if you run CPU physx with an ATI card, then yes, performance was either without malicious intent or with it, crippled.

Just one more time we need a separate bench, this one is seriously incomplete.


April 19, 2010 6:51:38 PM

You've read into it wrong. Nvidia crippled the FP64 compute function on its Geforce line of cards but not the Tesla line (to make the Tesla line sell more).

Physx works off the FP64 compute function but because it doesn't need all of the Gigaflops 100% FP64 functionality would give and can work off just a 25% capacity.
April 19, 2010 10:32:01 PM

Griffolion said:
You've read into it wrong. Nvidia crippled the FP64 compute function on its Geforce line of cards but not the Tesla line (to make the Tesla line sell more).

Physx works off the FP64 compute function but because it doesn't need all of the Gigaflops 100% FP64 functionality would give and can work off just a 25% capacity.


I don't think you read it at all. We're not talking about how they crippled GeForce Fermis, we're talking about how Nvidia's newest Physx software 'supposedly' crushed CPU PhysX performance with ATI cards.

After re-reading everything, I can completely surmise that the testing physx-info did is largely irrelevant to whether or not Nvidia crippled performance. While they showed that CPU physx with an NV card didn't change, the accusation is that CPU physx changed when an ATI card is present. In effect, accusing NV of hurting performance of physx for ATI based systems.

Need reliable testing of performance of CPU physx with an ATI card, old software to new software. Until then, we have on accusation (not very detailed, either) and can draw no conclusion. Sounds to me like the update didn't go properly, in all honesty.
a b Î Nvidia
April 20, 2010 5:17:32 AM

The Article mentioned in this thread says......"Framerate measured by FRAPS 3.0.1." and thats a straight quote from it, I didn't know FRAPs was used to measure frame rates as a standard..... actually I haven't even come across it as any sort of standard except for people capturing gameplay to be posted online .....
I don't think I or actually anyone should be taking that review seriously.
Benchmarks are usually based on standard benchmarking software, as far as FRAPs is concerned, a lot of stuff comes into play when you look at the framrate you get on the screen.....
I think it's a bad review, it may not be wrong, but it still isn't the right review....
!