I'm not looking for improving frame-rates, I'm looking for physics that is believeable, like what Crysis is offering.[/qoute]
The only 2 physics items you would have that would be different between the two implementations would be the interaction physics with the foliage & terrain and then the possible use of objects affecting gameplay (pipes being blow off a parked truck killing the people standing beside it). The question is what is the load difference between those and the 'shiny physics'. So far neither solution has shown serious game dependant physics.
Honestly, you can't expect uber awesome physics without frame-rate loss, and a PPU cannot contribute to increasing frame-rates, since it doesn't do rendering, only the physics calculations.
Sure it can contribute to frame-rate increase, it doesn't have to do rendering it need to free up resources. Part of the selling feature was to offload the task from the CPU, which should free it up from dips. If we say a game is CPU bound, then if that CPU load overhead is impacted largley enough by physics, then the PPU should improve that and thus increase FPS by removing the system limitation. That was one of the selling point (free up the CPU), it could also increase the level of physics, but they seemed to have gone overboard with it and ended up drooping frames which should be an inssue if it's a 'global' improvement, improving one area at the cost of others is not what everyone wants. If the load of 5.1 audio turned you 80fps sterea FPS into a 20fps 5.1 game, would you say 'great' it's so much more immersive? There needs to be balance.
And with all the extra data going to the GPU telling it to render all these extra effects, something's bound to give.
Except that the VPUs weren't at 100% load, so the giving is probably once again at the CPU level since the PPU stops there first, then tell the CPU what to send to the VPU.
Ageia is a pioneer in the field, but now as it has broken the ice, others will follow, and we shall start seeing better solutions in order to provide the physics we all desire.
I doubt it. If they weren't going against 2 hardware giants and one of the bigger names in game physics developement I'd say they had a great chance because no one else was even considering it, but they have a tough battle ahead of them and I doubt we'll see many more entrants into the market. Ther was rumoured to be one other (can't remember who right now), but even during the recent discussions their name hasn't come up again.
The best thing Ageia did WAS push the envelope and get people thinking about this. I'm not sure if HavokFX or SLi-Physics or ATis implementation would've seen as much development effort if Ageia weren't there to push them along, but they do deserve Kudos for getting the ball rolling, now they just have to make sure it keeps rolling without being crushed by it.