Here's some benchmark times I've just produced in the last hour using a rendering benchmark software called ratgpu.
***GPU+CPU comb config "Nvidia Quadro 1800m + Intel Core i7 840QM" 358.343 Seconds to render using PhysX***
***GPU Only config "Nvidia Quadro 1800m" 240.752 Seconds to render using PhysX***
***CPU Only config "Intel Core i7 840QM" 1,137.726 Seconds to render using PhysX***
You can see the GPU rendered the physics the fastest.
While running these benchmarks I found the complete opposite of the theory that we save resources by enabling the CPU to handle physics. I used the tskmgr to monitor resources while running this benchmark and found that it uses 99-100 percent "Safe to say 100 percent" value of the CPU in GPU+CPU and CPU configurations. While it used less than 15 percent using only the GPU. Not only does enabling the CPU use up all of the CPU resources. It took 1/3 times longer with GPU+CPU config and 4 times longer with just the CPU configuration. The GPU alone saves 9/10 of the CPU rersource and renders the physics 4 times faster. GPU Wins the Physics round. Surely you're better off with letting the GPU alone run PhysX applications.
The above is regurds to a previously closed thread.
The argument is not about what is more powerful. The argument is that for gaming purposes, the CPU has enough power for decent physics without needing to draw power from the GPU's main task of rendering frames. It's also an argument that is aimed at single GPU systems.
Weither it's true or not is up to opinion. But these are just speculations.[/quotemsg]