Graphics Cards Guide
PhysX: Then & Now
PhysX: Then & Now
PhysX began nearly seven years ago in 2002, when a company called Ageia was founded. Their business was built around PhysX, which was at that point of a time, used to refer to their physics processing unit (PPU) and their middleware PhysX engine that interfaces between their PPU hardware and the game designed to take advantage of PhysX . They also licensed out their PhysX SDK, which could handle an array of physics effects like particles, smoke, dust, fluids and even hair, to game developers. This allowed developers to create worlds where objects would react and interact realistically with one another.
The PPU was made available sometime in mid May of 2006 and was essentially a separate card that needed to be mounted onto your system's motherboard. Although initial demos with the physics engine were promising (you should check the Cellfactor demos online), PhysX never really took off back then because getting an extra card just to process physics was a luxury most couldn't afford and neither was PhysX pervasive in the gaming world, thus both reasons led to a poor adoption rate.
Fortunately, however, NVIDIA saw potential in Ageia and with PhysX being steadily implemented in a few big-ticket games (one of the first prominent games to make use of it was Unreal Tournament 3), they acquired the outfit early last year in February. The first thing NVIDIA did after acquiring the Ageia was to port over PhysX so that it could be used in tandem with CUDA to enable hardware acceleration of physics in PhysX-supported games, without the need for a separate PPU.
Months later in August 2008, NVIDIA finally integrated PhysX into its CUDA framework, allowing all GeForce 8 and above graphics cards to support PhysX processing. This made PhysX add-on cards immediately redundant for users of NVIDIA graphics cards.