It's again slow on the news front, but I recently discovered this interesting news item over at X-bit labs :

If you're into realtime physics (may it for games or for applications) you may have heard of Ageia, the company that created the worlds first PPU (Physics Processing Unit) which is an add-in card that sports own memory and the afore-mentioned PPU. It was meant to accelerate physics by a large amount since it's a special purpose processing unit for only physics calculations. And though Ageia seemed to be rather successfull, with raising much funding and even lining-up strong partners like ASUS and BFG it was only a matter of time until one of the big graphics companies (or even chipmakers, but both even AMD and Intel are somewhat slow in terms of innovations) would ''intervent'' by doing it's own stuff on that field.

And so short after announcing it's new line-up, Raja Kodouri (a senior architect at ATI) spread word that ATI is researching on doing physics calculations on it's GPUs. So this would (in the long-run) make the PPU add-in card obsolte and give all users (as soon as the first GPUs with physics-capabilities are wide-spread) physics acceleration, without having to spend an additional 250$ for the PPU card. And one more good thing is that, depending on how it'll be implemented, it would make physics acceleration accessible for ALL physics engines (Newton, ODE, etc.) and not just one (Novodex, in case of Ageia's PPU).

So this shure was an interesting statement from ATI and one thing you can count on too is, that NVidia is also silently working on such additions to their GPU. So this means a bright future for hardware-accelerated physics.