Do you really not see the problem with features that lock out one brand or the other? When AMD does advanced lighting, or TressFX, it runs on everyone's hardware. That's the difference that people are pointing out.
If AMD had created their own API, do you honestly think they would allow NVidia users to use TressFX? TressFX runs ons on NVidia hardware because it uses DirectCompute, which was created by Microsoft.
As long as it's an advantage for NVidia to lock PhysX to NVidia hardware only, they will continue to do so. There's no incentive for them to create an open standard with AMD on this..
And the only reason why it's been advantageous for them so far is because PhysX has been the only game in GPU accelerated physics over the years.
Havok at one time was going to be ported to OpenCL, but after it was purchased by Intel, it was canceled because it's in Intel's best interests to keep physics on the CPU.. And of course AMD's Bullet uses OpenCL, but it is nowhere near as refined or as comprehensive as PhysX.
I can't even think of a single game off the top of my head that uses Bullet.. :\ While NVidia have consistently and relentlessly been polishing and refining PhysX over the years till it's now indisputably the best game physics middleware option out there.
I personally wish NVidia would port PhysX to OpenCL, because only then will hardware accelerated physics
REALLY take off. But like I said, it's not going to happen unless pressure is applied to NVidia, preferably in the form of some hard competition.
