PhysX to become a CUDA extension, PS4 developers rejoice

Posted by Theo Valich

Santa Clara (CA) - If you have been wondering how Ageia's PhysX SDK fits into Nvidia’s product range, the answer is actually very simple: The PhysX SDK will be translated into CUDA and, as a result, PhysX will be compliant with the current GeForce 8000 series and all future products from Nvidia, such as the upcoming GeForce 9600 GT and 9800GX2.


Truth to be told, it is hard to believe that G80-based architectures such as G80, G92, G94 would have enough horsepower to calculate PhysX and still deliver gaming experience. In a way, the future looks like this: Buy $200-$250-$300 graphics cards for SLI or Three-Way SLI, one of which one will be used for PhysX calculation. Nvidia pitched this setup already with two 7800GTX/7900GTX cards, but it didn’t exactly fly. The same thing applies for ATI's Menage-a-Trois, a combination of three X1900XTX graphics cards caught in an endless love affair.

The troubling fact is the cost of a single GeForce card versus a single PhysX card. When comparing current GeForce 8800GT vs. BFG’s PhysX card on NewEgg, we end up with a drastic difference in price ($239.99 vs. 109.99).  We hope that future graphics cards in the $100-150 range will be able to run in-game physics. Use the expensive card to do what the any gamer wants: Run games with the best possible effects.

When it comes to gaming consoles, we cannot forget the fact that Ageia plays a significant role in both Xbox 360 and Playstation 3 consoles. There is a ton of titles that run PhysX and with the conversion to CUDA, it will be interesting to see how Nvidia will address this acquisition when it comes to designing a Playstation 4 console. It is too early to talk about specs, but we bet that Nvidia is creating a complete solution for Sony, which will help to bring down the price of manufacturing. In fact, a PS4 might end up in the range of the first Xbox. We are digging our way through the tunnels below Santa Clara in order to get the details of PhysX and the PS4 development. Stay tuned.

During Nvidia's fourth-quarter financial results conference call, Nvidia shed a little more light on its acquisition of Ageia and what it plans to do with the firm's PhysX technology. CEO Jen-Hsun Huang made no announcements regarding the deal until asked in the question-and-answer session, but he was happy to divulge a decent number of details.

Huang mentioned that Nvidia's strategy is to take the PhysX engine and port it to CUDA (Compute Unified Device Architecture), which is a C-like application programming interface Nvidia developed to let programmers write general-purpose applications that can run on GPUs. All of Nvidia's existing GeForce 8 graphics processors already support CUDA, and Huang confirmed that the cards will be able to run PhysX.

"We're working toward the physics-engine-to-CUDA port as we speak. And we intend to throw a lot of resources at it. You know, I wouldn't be surprised if it helps our GPU sales even in advance of [the port's completion]. The reason is, [it's] just gonna be a software download. Every single GPU that is CUDA-enabled will be able to run the physics engine.” Hunag believes the integration will encourage people to spend more on graphics processing hardware, as well.

"Our expectation is that this it will encourage people to buy even better GPUs. It might, and probably will, encourage people to buy a second GPU for their SLI slot. And for the highest-end gamer, it will encourage them to buy three GPUs. Potentially two for graphics and one for physics, or one for graphics and two for physics," he said.

Last but not least, Huang said developers are "really excited" about the PhysX-to-CUDA port. "Finally they are able to get a physics engine accelerated into a very large population of gamers," he explained. Huang declined to comment on the time frame for the release of the first PhysX port. However, considering this will be a software implementation and Nvidia now has Ageia engineers on its payroll, the port should not take too long to complete.