AMD Talks GPU Gaming Physics

[TheRealDeal]

http://www.bit-tech.net/hardware/graphics/2011/02/17/amd-manju-hegde-gaming-physics/1

A very interesting read. What is said makes a lot of sense and could be the way to get gaming physics into the mainstream.

 

[Freddie1980]

That was a worth while read. It was interesting to note that many cuda based apps inculding physx could be going the way of opencl which is a win for consumers IMO.

 

[Skellig]

Interesting stance on OpenCL.

 

[Rroff]

For someone whos apparently the co-founder of Ageia he doesn't really seem to know what hes talking about.

He also reckons that Nvidia may soon have little choice in the matter, though, as many CUDA apps, including PhysX, will end up being ported to OpenCL anyway. 'I think a lot of the CUDA applications, not just PhysX, will be ported to OpenCL,' says Hegde, 'because Nvidia's not a big enough company to take an API and make a whole industry drive towards it. I think it made sense when there was no competing API, and developers were looking for something new to improve their performance, but today there is a very viable OpenCL offering from us.'

nVidia have done it - they have a sizeable and growing industry built around it, I'm seeing 2-3 new CUDA based projects popup on my development stream every day - there is no imminent migration away from CUDA - its wishful thinking at best. Would be nice to see PhysX ported to openCL but I really don't see it happening, as far as gaming goes we are more likely to see developers move towards a directcompute based physics API if MS was to put some weight behind it.

I think one of the issues is the model that we used, where we fixed an API and didn't allow games developers to adjust it, because we didn't want them to break the hardware acceleration. It was definitely a model that's more hardware-centric for the vendor, rather than developer-centric for the content developer, and Havok by the way has a similar issue. Bullet is different – they want you to do stuff yourself, they want you to use the API and do everything underneath it – they give you a set of functionalities, but that functionality is effectively more or less the same to everybody, and any differentiation comes from really how you use it in your game story.

Forgetting that the PhysX API under nVidia it is possible to license the full source? also while the concept he talks about of tweaking underlying functionality is great for CPU based physics engines, you can cut out the stuff you don't need and really tweak it to work great with the specific effects/collisions you need for your application, it generally doesn't make sense when using a fully featured hardware accelerated physics engine where your using a broad spectrum of features.

ATI/AMD have been all talk and not enough action on this front for years and I don't see it changing now.

 

[jigger]

Hopefully this is the beginning of the end for closed standards. ATI seem set on making a big a push for an industry wide standard, and I'm sure the industry will fully support this.

Nvidia should have done the same with FizzX when they had the chance.