|
|
Nvidia helping to bring PhysX to ATi Cards
Author: Harry ButlerPublished: 9th July 2008
Nvidia are helping to bring PhysX to your Radeon cards.
For those not in the know, Nvidia aquired PhysX developer Ageia back in February, and has successfully integrated PhysX support onto it's latest GT200 core graphics cards, with PhysX support for any Geforce 8-series or above card being released in the next few months.
In response AMD licensed Intel's Havok physics API (yeah, it's complicated) to include Havok acceleration on it's own Radeon graphics cards.
However, an intrepid team of software developers over at NGOHQ.com have been busy porting Nvidia's CUDA based PhysX API to work on AMD Radeon graphics cards, and have now received official support from Nvidia - who is no doubt delighted to see it's API working on a competitor's hardware (as well as seriously threatening Intel's Havok physics system.)
As cheesed off as this might make AMD, which is unsurprisingly not supporting NGOHQ's work, it could certainly be for the betterment of PC gaming as a whole. If both AMD and Nvidia cards support PhysX, it'll remove the difficult choice for developers of which physics API to use in games. We've been growing more and more concerned here at bit-tech at the increasingly fragmented state of the physics and graphics markets, and anything that has the chance to simplify the situation for consumers and developers can only be a good thing.
Still hoping for a unified physics API? Or is it still early days for physics in games? Let us know in the forums.
Recent articles on bit-tech.net
Comments (15)Stats: 0.120 seconds
It'll making everything far better for us, and make the corporations far more respectable.
http://www.khronos.org/news/press/releases/khronos_launches_heterogeneous_computing_initiative/
In the mean time we're stuck. :p
AMD claim they won't support CUDA as it isn't open source.
Because PhysX is on dedicated hardware?
http://www.cgl.uwaterloo.ca/~csk/its.html
Don't make me call the Punctuation Police...
I bet they're feeling pretty smug though, having physics solutions owned by their competitors running on their cards while their competitors themselves are stuck on a single solution only :-P
So you've written in the article "it looks like Nvidia is indulging in some cross brand tactics to promote it is PhysX API". Which just isn't accurate.
On topic though: a unified physics API would be rather nice, although to be honest has anyone seen PhysX actually do anything properly useful in any game, ever?
I saw the tech demos....
.. They were pretty cool...
A unified method of performing general computing tasks on the graphics card would be more useful than having everyone use the same physics API; you wouldn't want another DirectX vs OpenGL scenario arising (where one completely dominates the other, for PC games at least). If nVidia makes CUDA an open-source specification and therefore allow all graphics card manufacturers to implement it we'd have a far better situation than if everyone adopted PhysX instead.
So I guess, in my opinion, the questioner is asking the wrong thing. Being forced to use a single proprietary platform for a wide array of things is generally a bad idea, but having an open specification for how to communicate with a graphics card to have it perform general computing tasks for you is a fantastic one.
Unifying the physics market would slow the advances it makes considerably; we've already seen that nVidia slows down with its technological progression when it's dominating the market =)