NVIDIA presents SLI accelerated physics rendering

NVIDIA will presented its new SLI Physics technology later this week at the Game Developer's Conference in San Jose.

In October of last year, ATI discussed the opportunity to use the massive floating point computation power of a graphics processor to enable accelerated physics - but Nvidia is first to introduce the feature to the market: Called "SLI Physics," the feature will offloads physics calculations from the CPU to the graphics processor and promises to bring movie-type effects from crashing cars and speeding bullets to the PC screen - all with smooth frame rates. A new software driver for Nvidia's graphics cards will use the second graphics processor to enable the feature in future games.

AGEIA was one of the first manufacturers that talked about launching a PPU (Physics Processing Unit), called the PhysX. The company is still waiting for the first games that support this to launch their cards and it's getting kinda clear what ATI and NVIDIA will do to counter AGEIA's PhysX cards.

Instead of launching separate PPUs they will rely on multiple graphics cards to do the task. This will possibly be slower but NVIDIA has the advantage that it has already shipped millions of SLI cards.

Multiple graphics processors provide plenty of excess floating point capability to run more than just graphics and overcome the limitations of the main processor of a computer system. Physics appears to be the first and major new application graphics chips will be aiming for.

NVIDIA will soon release a new driver to enable this feature. The company says the new drivers will simulate object and particle effects like fluid, smoke and dust, all within the GPU. More details at TG Daily.