NVIDIA Changes GPU Requirements for PhysX

• September 3, 2009
• Written by Mark
• Leave a Comment
• Tweet

While reading the [H]ardOCP Physics Processing forum it was brought to our attention that Nvidia as changed the minimum GPU requirements capable for running PhysX. Until now Nvidia supported any Geforce 8 series GPU and above. However after reading this thread over at the [H]ardOCP Physics Processing forums it appears this has changed. The Nvidia PhysX GPU Support page no longer lists the GeForce 8500GT, 8400GS, 9400GT, 9300 mGPUor the 9300 mGPUas capable of running PhysX. This is probably due to a change in the amount of stream processors that Nvidia now requires for PhysX.

from the Nvidia PhysX FAQ:

Which NVIDIA GeForce GPUs support PhysX?
GeForce 8-series GPUs and later (with a minimum of 32 cores and a minimum of 256MB dedicated graphics memory) support PhysX.

This is a trend that will most likely continue as more and more games begin to take advantage of PhysX and more developers begin to start implementing more PhysX intensive features into games.

AI on the GPU: InstinctTech DogFighter CudaDemo

• September 3, 2009
• Written by Mark
• 2 Comments
• Tweet

In the near future we may not just be dealing with dedicated PhysX cards to run Physics calculations for games, you may also have a dedicated AI card:

From Bluesnews:

Instinct Technology has collaborated with Dark Water Studios to create a movie showing how their middleware product Instinct Studio is able to use NVIDIA’s CUDA technology to offload AI from the CPU onto the GPU of the graphics card. The trailer shows GPU-controlled flocking behavior in DogFighter that allowed them to display a ridiculous (as in over 4000) number of aircraft on the screen at once, all controlled by a single CUDA-enabled graphics card. The clip is posted on the YouTube, and here’s word:

This is a video from a case study/tech our friends from Instinct Tech (www.instinct-tech.com) did for us. We wanted to know how easily a technology like CUDA could be used from inside Instinct Studio (the game engine were using for our WIP title DogFighter [www.dark-water-studios.com]) and if the additional horse power can be utilized for some game mechanics.

The demo shows 4096 bot planes handled solely on a single GPU parallel to game rendering. The bot planes use steering behaviours for flocking, navigation and obstacle avoidance. The planes are fully lit and rendered (with shadows). The demo runs with interactive frame rate on main stream CUDA enabled graphics cards. In comparison the same simulation without utilizing CUDA achieved a similar frame rate on a decent machine with only 512 planes in our tests. The steering computation for 512 planes requires about 260.000 neighbour queries while for 4096 planes this grows to a whopping 16 Million queries. The algorithm can be easily parallelized, explaining the advantage of technologies like CUDA for this kind of problem. Even if there is a potential to optimize the algorithm for CPU the clear benefit for us is a heavily reduced development time.