Nvidia Responds to AMD's Claim of PhysX Failure

Not to take sides here, as I own an Nvidia card fwiw. But I came to the same conclusion as Richard Huddy before ever knowing he made this statement. It struck me when toying around with PhysX on Batman Arkham Asylum.

I disabled card PhysX and let the CPU handle them just to see how it performed. Strangely, my CPU usage barely increased at all and framerates suffered immensely as a result - same thing reportedly occurs with ATI cards.

The physics being calculated on this application are not particularly intensive from a visual standpoint, especially not when compared to say what GTA IV does (which relies solely on the CPU). They are just terribly optimized and by my estimation intentionally gimped when handled by the CPU.

Anyone can connect the dots and understand why this is so. It's just stupid because I bet a quad core CPU, or even a triple core paired with say a measly 9800 GT can max out PhysX and the in-game settings if the CPU handled the PhysX without being gimped. But since it is gimped, owners of such a card pretty much cannot run PhysX.

If game developers added true multicore support in the first place i bet this would have never even come up even the newest games like borderlands have bad multicore support I know almost nobody with single core cpus these days the devs need to step up

I am of the same opinion as AMD here.

The response sounded quite well founded. Don't think Nvidia is to blame on this one.

My first thought is PhysX has been a market failure since it's Ageia days. Nvidia is just using this proprietary gimmick to hock more GPUs. I was stunned when Nvidia bought Ageia, but I guess the price was right, and their in-house development was lagging. The list of games using PhysX is just sad, and the performance hit with PhysX enabled is rough. Makes you wonder how big of a carrot Nvidia has to dangle out there to get the developers to bite.

I think PhsyX is a bunch of bull anyway. Any game I have seen that utilizes it isn't anything special. Like, I mean.. the physics aspects implemented in the game. I.E... they wouldn't be missed. (Arkham Asylum is a good game) They wouldn't need an GPU. It's just a money maker in my opinion. Garbage.

It is true, it has been proven many times like FoSjizzleDizzle explained.

I own nVidia as well, but their anti-competitive acts are really starting to piss me off.

Luckily DX11 will make PhysX completely useless anyway.

As a amateur game developer I was intrigued by Physx since it's a significantly cheaper route then Havok. However, I found that Physx just had more problems then it's worth. For instance, 1) Physx makes use heavy of the PCI bus when in pure hardware mode. How fast can it really be if your utilizing the PCI bus which has a maximum bandwidth of 133megabytes/second(or 4 megs per frame) 2) Nvidia has been caught already locking the competition out of Ageia physics 3) the pci express bus is quoted by microsoft as SLOW and needs to be shared by the graphics card 4) No direct hlsl interface with the physics directly(you have to use a C++ call to get around it 5)bullet physics is free and offers cross platform gpu based physics 6) To write custom physics with ageia you will need to write a event handler that will have to be invoked by the C++ api on a PER ACTOR/ENTITY basis. This can be a problem if you wish to have LOTS of entities/actors.

So yeah not too crazy about Ageia and havok is costly as well. Anyways that's my 2 cents.

'We continue to invest substantial resources into improving PhysX support on ALL platforms--not just for those supporting GPU acceleration.'

this made me lol.

if they are so intent to support multi-platforms, why is their primary platform (PC GPU acceleration) locked out in the presence of competitor hardware?

Each corporation or business wants to drive it's competitors out of the marketplace, but doesn't want to pay the price for this in ant-competitive practice law suites. Nvidia is no better than ATI/AMD (or A^2 = A Squared as I like to think of them), or Intel.

"Our PhysX SDK API is designed such that thread control is done explicitly by the application developer, not by the SDK functions themselves."

To me, this means that it is up to the game developers to optimize thread control for multi-core CPU. It is not nVidia's fault that game developers choose to only spend time making phsyx work with GPU and not optimize it for multi-core CPU use.

Can AMD point to changes within the code that can show that performance of Physx has deterioted on multi-core CPU if you compare pre-Nvida Ageia API versus present day Nvidia Physx API?

Then we'll know who is telling the truth? If there is no deterioration, then nVidia is not in the wrong. Why would they spend resources making Physx work better on multi-core cpus. That is just a dumb business decision unless they see the value of doing so. It may be that they should do that or risk phsyx being ditched as a widely used physics engine.

Any modern CPU can pretty well handle physics without any slowdowns or affecting fps. All it takes is some multi-core optimizing, with Octacores CPUs coming, games are barely taking advantage of 2 CPUS, that's really really Sad but good for GPU makers. GTA graphics are badly optimized because it was ported from consoles. Best games never used PhysX on my book. Crysis, Stalker, MW1/2, NFS , Unreal 3 had physX in only 1 map and the physx rain was really really annoying during gameplay.

Developers wake Up please, It is Time to make use of Quad cores that have been in the market for 3 Years! Thats ridiculous seriously Q6600 release date January 7, 2007! God developers some is paying them to not do so,selling more gpus?

i wonder why all the forum people are down here?

could it be the badges? hmmm

EDIT: Quad cores were in the market for 4 years actually not 3. Core 2 Extreme QX6700 Release date : 2 Nov 2006 http://www.techspot.com/review/27- [...] 6700-quad/

Were mad gamers because PhysX is actually Slowing the Multicore game optimization process. Look how bad games are multicore optimized while they can run perfectly on a Quad Core if done efficiently. They would rather let the GPU make the physics calculations and forget about the CPU.

wow some times i wish they would stop take hits at each other

Wait, does this mean that PhysX isn't a waste of time and money?

Sure the PhysX SDK may allow developers to choose to make their PhysX apps multi-threaded, and many will (3dmark).

But the real question is, does NVIDIA pressure The Way It Is Meant To Be Played game developers to not multi-thread their PhysX apps.

To me physx is like the US economy both are rigged to be in favor of a select few while every one else either pays the price or miss out. Personally I hope that Nvidia suffers for a few years in order to learn from their on going mistakes and become a better company to its competitors as well most of all us the consumer plus if it wasn't for gamers this company would have gone belly up during the .com bubble. If they had ATI or 3DFX would be the main players with Matrox offering a few consumer cards that didn't cost a fortune but offered much of the same quality as they do today. Perhaps if some other player came on the scene with out Intel sucking every thing up if Nvidia failed then maybe a little Renaissance for PC gaming would pop up like what happened with 3dfx but with out better long lasting results. I wish that Amiga was still around and popular like they were.

Being an avid PC and game player I 100% fully support Nvidias PhysX. Since replacing my former Nvidia GPU with a 5850 I am not as pleased with the Ati cards' physics code. In several games, the physics preformance is poor when comparing to a lessor Nvidia card. Even though there is an increase in FPS from my 9800gtx+ in games like Crysis, Batman AA, GTA4, and Far Cry2, the physics somehow feels "off" or the calculations allow the realism in the Ati card feels faster. Example, the trees sway way too fast and unrealistically in FC2 while in Crysis the barrels and throwing of an object and the way it "bounces" and "rolls" of a rocky edge is not as fluid and realistic as the PhysX run engine.

Anyone else notice this?

I found the manipulation of words by Nadeem Mohammad.
The theme of the whole controversy was Nvidia skimping performance with a competitor's card.
Nadeem's main theme or point was "...there have been no changes to the SDK code which purposely reduces the software performance of PhysX or its use of CPU multi-cores."
What does this mean? That Physx will use multicores. Whether or not it will use multi cores at the same efficiency as Nvidia cards vs ATI card is a different issue. Notice how they never tackle the main issue which was Nvidia's anti competitive acts. They also say it's up to the developer to program the threading. We already saw with Batman Asylum that Nvidia locked AA out of ATI cards because of "copyrights", we can already guess they added incentives or threats to forget to optimize physx for ATI cards.

Shame one you Nvidia, you really need to step up your game or at least use PR as well as Apple that you can steal hundreds or thousands of dollars and still have happy clients. First ludicrous Fermi claims, then fake Fermi GPU's, now anti-competitive actions. I understand you been bullied already by Intel's anti-Nvidia campaign against the Ion, but seriously can't you learn by their anti-trust mistakes?