Nvidia releases magic physx drivers?

D

defiant007

2[H]4U
Joined
Feb 27, 2006
Messages
3,497
http://www.xtremesystems.org/forums/showthread.php?t=191765

db4tws7h.png


okaks9su.png


itw29hlm.png


Anyone able to test these?
 
Their answer to the 48-- success? Holding back drivers until ATI outperforms them?
Probably not, I am just a paranoid asshole :p
 
Those tests are bullshit. They are just *optimizations* for the vantage test by new nVidia drivers. The drivers enables the card to use the physics engine which helps in passing those numbers(something like that) which were previously disabled in the earlier drivers.

In games the 4850 and 9800GTX+ perform head to head and that's what matters. I have not bothered in the recent past to even install those benchmarks. They are just meant for misguiding the customers.
 
Switch said:
Those tests are bullshit. They are just *optimizations* for the vantage test by new nVidia drivers. The drivers enables the card to use the physics engine which helps in passing those numbers(something like that) which were previously disabled in the earlier drivers.

In games the 4850 and 9800GTX+ perform head to head and that's what matters. I have not bothered in the recent past to even install those benchmarks. They are just meant for misguiding the customers.
Click to expand...
Agreed. Actual gameplay is the only thing we should be considering. Synthetic benchmarks can be overly optimistic.
 
Some people took the 177.35 and modded the .inf file in order to boost folding performance. So without a doubt, these new drivers do give a performance boost. Ive seen it with my very own eyes.
 
defiant007 said:
Might be interesting to see if these drivers make a difference to physx enabled games, are actually quite a few games with it:

http://en.wikipedia.org/wiki/Physx#Title_support

Anyone want to do some benching using GoW, R6 Vegas, Mass Effect, GRAW 2 or MoH Airborne?
Click to expand...

The only thing the Geforce PhysX drivers will improve is UT at the present time (and 3DMark Vantage, of course). The other PhysX games are not supported with GPU PhysX acceleration at this time.

The difference in UT is simply amazing though.


It should also be noted that any of the 177.xx drivers will work with PhysX, you just need to download the seperate Nvidia PhysX driver and install it.
 
Blue Falcon said:
The only thing the Geforce PhysX drivers will improve is UT at the present time (and 3DMark Vantage, of course). The other PhysX games are not supported with GPU PhysX acceleration at this time.

The difference in UT is simply amazing though.


It should also be noted that any of the 177.xx drivers will work with PhysX, you just need to download the seperate Nvidia PhysX driver and install it.
Click to expand...

Why is that? Physx is just a an api software layer, if games utilise physx in their games engine wouldn't these drivers improve performance in those games?
 
might be some exceptions, but every game with physx support can already practically be maxxed out with current cards at high rez. maybe not 2560x1600 but 1920 for sure with max settings. it'll help the slower cards far more than the higher end ones. 8600GT/S and 8800GT probably gain the most performance from the hardware physx.
 
Switch said:
Those tests are bullshit. They are just *optimizations* for the vantage test by new nVidia drivers. The drivers enables the card to use the physics engine which helps in passing those numbers(something like that) which were previously disabled in the earlier drivers.

In games the 4850 and 9800GTX+ perform head to head and that's what matters. I have not bothered in the recent past to even install those benchmarks. They are just meant for misguiding the customers.
Click to expand...

How is that BS? Nvidia have spent a long time building in decent physics support into their GPUs and no doubt countless millions in development, and now we can benchmark it, all of a sudden it's BS?

Wow...just wow.
 
Frostex said:
How is that BS? Nvidia have spent a long time building in decent physics support into their GPUs and no doubt countless millions in development, and now we can benchmark it, all of a sudden it's BS?

Wow...just wow.
Click to expand...

Do you always jump through hoops when Nvidia tell you to?
 
APOLLO said:
Agreed. Actual gameplay is the only thing we should be considering. Synthetic benchmarks can be overly optimistic.
Click to expand...

i bet futuremark are starting to cringe at the ever-increasing number of statements like that

hahahaha
 
gaspah said:
i bet futuremark are starting to cringe at the ever-increasing number of statements like that

hahahaha
Click to expand...
Actually, they are with the mounting crescendo of naysayers echoing similar statements throughout the fora across the net. ;)
 
Frostex said:
How is that BS? Nvidia have spent a long time building in decent physics support into their GPUs and no doubt countless millions in development, and now we can benchmark it, all of a sudden it's BS?

Wow...just wow.
Click to expand...
http://www.legitreviews.com/article/731/1/

Have a look at the numbers in this review. They are done using the same cards with the drivers in question. The gaming performance is nowhere near in sync with the results posted above. And he has tested quite a few old and new games. Hence i call those nos/benchmarks BS.
 
Somnambulator said:
might be some exceptions, but every game with physx support can already practically be maxxed out with current cards at high rez. maybe not 2560x1600 but 1920 for sure with max settings. it'll help the slower cards far more than the higher end ones. 8600GT/S and 8800GT probably gain the most performance from the hardware physx.
Click to expand...

Really? And what of all the games posted above?
 
I could care less about the Gar..Vantage scores but the UT3 scores are crazy. In any case free stuff is always welcome in my world, regardless if I use it or not.
 
There are no magic drivers and no magic performance increases in any card from any company. And as many said, 3D Mark and Vantage scores are worthless.

As far as PhysX, I'm sure that there will be much to gain over time, as more developers use it in their games, which is something I like. Being able to fully utilize my GPU for other things than just graphics.
 
Darn! I only just installed 177.35 drivers.

How come 177.39 isn't on the nvidia website yet though?
 
GenBanks said:
Darn! I only just installed 177.35 drivers.

How come 177.39 isn't on the nvidia website yet though?
Click to expand...

Because it's an unreleased or 'leaked' beta driver. These drivers can be leaked from a variety of sources ranging from one of the board partners (EVGA, BFG, XFX, etc) to an unscrupulous e-press website. Generally around these parts lots of people are running one of the latest unreleased betas rather than anything offered at the official nvidia website - myself included. I love new drivers. :D
 
Question: As more games make use of physx, and knowing that physx and other game calculations are done at the same time on the GPU, is the bus size if the card going to be needing to carry any extra data now? (The new physx data)

I am wondering if this type of driver update may start to show a bigger performance gap between the 280 and the GX2.

If it's true that physx enabled GPU's, could make games that support the physx engine run faster, but will also require your GPU's bus to move more data, we may indeed see the 280's scores move up more than a GX2's...

Probably splitting hairs here, but just wondering. :D


UPDATE: There is currently no SLI support, so the 280 will for sure shine here over the GX2.
 
I want to try this later today on 8800GTS and 8800GT cards. :D
 
You must log in or register to reply here.
Back
Top