AMD to PhysX: 'Acceptable Under Conditions'
Following the NGOHQ episode with devising software that lets users accelerate GPU-based NVIDIA PhysX API on Radeon accelerators, and with NVIDIA coming in support of such an effort, a general opinion was made that NVIDIA sought an industry-wide domination of CUDA as the de-facto general purporse graphics processing (GPGPU) architecture, with putting their investment of acquiring Ageia Technologies to good use by pushing the PhysX API. Although it comes as a good news for AMD that their graphics cards that are already optimized for Havoc physics could now support PhysX acceleration, it's not in the best interests of the company that they allow the growth of CUDA and components based on it to this extent, since AMD has its own FireStream line of products and a GPGPU architecture in the making.
TG Daily spoke with Richard Huddy, Manager of Worldwide Developer Relations, and Godfrey Cheng, Director of Product Marketing, two key individuals with AMD. When it comes to the most interesting question of PhysX implementation on Radeon, Mr. Cheng says that AMD has no problems encouraging the use of feature-enhancing 'middleware', and that they have no arguments in NVIDIA going ahead with propogating their PhysX middleware as long as they don't put Radeon accelerators into a unfair disadvantage.
TG Daily spoke with Richard Huddy, Manager of Worldwide Developer Relations, and Godfrey Cheng, Director of Product Marketing, two key individuals with AMD. When it comes to the most interesting question of PhysX implementation on Radeon, Mr. Cheng says that AMD has no problems encouraging the use of feature-enhancing 'middleware', and that they have no arguments in NVIDIA going ahead with propogating their PhysX middleware as long as they don't put Radeon accelerators into a unfair disadvantage.
