Geforce : Force Within

The idea of using the GPU to assist in non-gaming stuff like video encoding seems pretty cool.

seems to be the answer to microsoft chosing Havok for all theyr game studios and future releases (as does Intel and AMD/ATI with theyr hardware support). Word on the streets is havok will be part of the next DirectX

Quote from Thrik

The idea of using the GPU to assist in non-gaming stuff like video encoding seems pretty cool.

yeah i am actually pondering about writing my bachelor assignment about GCGPU (general computing on gpu). there's much more power to tap into from the gpu than is used mostly nowadays.

Yhm. I installed all these freebies except that full game thing and honestly i cant underdstand why this shit weights so much Honestly, i cant find anything new that my card or system couldnt do before

If you run that fluid demo, and switch between software physics and hardware physics, you will see a difference...

It will aslo support the "what it should look like" physics and post process filter button in Diablo3.

Nice! what perfect timing for me. I just bought myself a 8800GTX (found it on sale for a nice price) so now I get even more value!

Do you need multiple cards or can it work on just one? For games like UT3 for example that could use a physics card for stuff will my one gfx card do graphics and physics or would I need a second card to dedicate to physics for in games and stuff?

One will do both gfx and physics

In some ways this is cool, but I don't get the point of pulling additional workload to the GPU when most computers are GPU bound anyway when playing games. Outside games in Photoshop and whatnot it's about damn time though that they start using the GPGPU.

Presumably it'll be optimal when running a multi-GPU configuration. Then you've effectively got the choice between GPU power and PPU power, rather than having to buy a dedicated PPU card and being stuck with a paperweight for games that don't use it.

The GPU is specialised hardware, so it makes sense to load some stuff onto it, even if it means there's less available for graphics processing (i.e. by dedicating only 10% of a specialised GPU, you can perform certain types of calculations that would require a significant amount of CPU time, even with multiple cores available ).