Watch for Competition and Convergence In The Physics Space

As I mentioned before - this'd make a killer addition to a GPU.

People originally found 3D add-in boards to be the perfect complement to their 2D Graphics boards, but in the longer run, the integrated solution ended up winning there. Convergence comes to those who don't buy instantly.

I'll have to get a PPU board now anyway, because I'm Just That Kind Of Early Adopter (wonder if x64 drivers will be an issue? (grin)).

I can see the marketing wars between physics accelerator leaders now, as features are added while vendors scramble up the stack to provide more and better and easier features for game programmers:

"NewtonX does 30,000 bones with 16-bit accuracy".

"Well HawkinG 3000 does 60,000 bones with 32-bit accuracy, and hardware assisted broken teeth simulation".

"Ah, but NewtoNeXT will do 280,000 broken bones with 64- bit accuracy, hardware assisted broken teeth and nose, full CPU offload for many Volkswagen-crashing-into-shop scenarios, and inbuilt support for meteor storms that threaten to destroy a planet."

"One word: DirectHuman!"

"DirectSmallCrowd."

And so on, all the way up to "DirectWorldRavagedByWarOrOtherUnspecifiedDisasterButSomePeopleProbablySurvivedAndNowWantToBuildMoreShoppingMallsAndByTheWayThisIsAnFPSLikeNoOther".

Or, more hopefully by that time, they'll just have stuck it on the video card and called the overall package something fun, like "Project Kneescrape".