Nvidia hobbling physx on CPU to make themselves look good?

[Adrie]

I was recently reading some thing on the internet and came across the article "Nvidia hobbling physx on CPU to make themselves look good?" under that it was saying something like

NVIDIA has exposed its marketing tactics. They say that PhysX is faster on a GPU than a CPU. It was proven that the only reason that PhysX is faster on a GPU is because Nvidia purposely hobbles it on the CPU. If they did not, PhysX would run faster on a modern CPU.

I really did not understand much from there, so, I decided to have little discussion on it here on this forum. Can you all here put some light on this. Thanks.

[David23]

Using the PhysX Nvidia wants to highlight the advantages of their hardware over the CPU. They want it for physics calculations and also for differentiating with AMD’s GPUs. They are with only one motive they want to make Nvidia’s hardware look good and sell more GPUs. For claiming their place in market they have made sure that Nvidia GPUs looks a lot better than the CPU.

[Archibald]

Well NVIDIA are in the market to make money, and even I would have done the same thing. They want to have a fix place so that they can be your gaming buddy. The only thing that makes me laugh about PhysX is that, their card can not run and render at the same time. This is one of the biggest failures in modern GPU history.

[Chandranath]

The rationale behind using PhysX is that Nvidia artificially blocks any other GPU from using PhysX. PhysX is a proprietary realtime physics engine middleware SDK . Going so far as to disable the functionality on their own GPUs if an ATI GPU is simply present in the system but completely unused. NVIDIA, I would love for ATI to eat you up. I would go much deeper in this discussion to hammer a few more nails in Nvidia's coffin.

[CACTUS101]

There is no problem with PhysX running on CPU, but the only thing that NVIDIA offers with the GPU is nearly 4 times better performance. That is the reason why using a graphics card allows for a greater number of more complex objects to appear in games. It has been shown that when the physics code is run on a CPU, it relies on the x87 instruction set. Modern processors also support these instruction but it is because of the chip manufacturers that are discouraging developers from using them.

[CAILYN]

Archibald, what you said is not true. A single card is capable of both rendering and calculating PhysX. Prior to the current generation, only 1 kernel could be run at a time on both ATI and NVIDIA hardware. Fermi can run multiple kernels at once now. I don't think ATI hardware has that ability yet, or at least I can't find it in the current documentation or mentioned in the architecture overviews.

[xanix]

I totally agree that NVIDIA is up to it's marketing tactics. PhysX can run on a good CPU, but NVIDIA does not want it to happen. It is also the fact that PhysX and 3d can run on a GPU but is possible only in theory, and in reality, frame-rates suffer. So, in order to do that they will sell you a GT240. PhysX is something that can PC gaming incredibly awesome.