Thoughts
The premise of physics acceleration by dedicated hardware (and we mean something other than the CPU in this case) is a solid one. While there's always a downside to adding a discrete piece of hardware to a PC to do something better -- cost, noise, heat and power consumption in the main -- the upsides usually make it worth it.However, much like the early days of 3D graphics (take a moment to remember your first encounter with GLQuake, or the Tomb Raider patch if you were in to 3D graphics back then, for those are sweet moments), the PhysX PPU needs traction with a killer title or two to make the world sit up and notice. GR:AW that ain't. Bet on Soldier that ain't. Rise of Nations that ain't.
I remember the feeling the first time I fired up GLQuake on my Voodoo Graphics accelerator. I get almost the same feeling when I fire it up now. I don't get that feeling when I play PhysX-enabled titles, since the effects presented in the first wave of supporting titles isn't as overwhelming and persistent. And it remains to be seen where the first killer title will come from.
Developer support is pretty strong, though, with Unreal Engine 3.0 the biggest current proponent in terms of game-building middleware that uses it. We wait and see.
Technically, the API and devrel needs to mature to encourage developers to make the investment, and if you are making it you have to think hard about its integration, lest you run into problems with use latency of return data when using it (the usual data path is CPU-PPU-CPU-GPU, adding in an extra data roundtrip to the PPU IC before rendering takes place).
GPU effects physics (just Havok FX™ right now) is the easier bet for a developer looking to tack on effects-based stuff to an existing engine, too, especially if they already license Havok Physics™. That means we pretty much just wait and see what happens.
Currently there's no reason to spend the £200+ to acquire one. The current effects in the supported games aren't worth the price and potential performance drop. Cell Factor and awesome UE3.0 games, where art thou? Without those the PhysX hardware is a curiosity for the gamer.
That the physics API space for games is fragmented -- with Ageia and Havok plowing the field on their own, and with the Khronos group and Microsoft not announcing any Open or DirectPhysics making it seem like it'll be that way for a while -- isn't good for adoption either.
Outside of gaming there's some scope for the PhysX hardware to be used for general purpose parallel floating point computation, given its parallel nature, but the API simply doesn't cater for that in any meaningful way just yet, the interconnect it's sat on somewhat limits the usefulness in that respect, and it's a 32-bit Windows device only at the time of writing. The goodly chunk of on-board storage counts in its favour for that, though. Another distant possibility.
The limited number of titles and their disappointing use of the PhysX PPU means that this first look is almost not worth the bother, but it serves us well as an intro to what it's capable of, hinting at future promise. We'll revisit the PhysX PPU in due course.
