PhysX - What went wrong and what's next

Conclusion:
"If PhysX dies, it won't be because the hardware has failed; it hasn't. It'll be because it was a small fish in a small pond already filled with sharks. It's Beagle Two up against NASA, and while you have to admire its pluckiness, its homemade, tinfoil and sellotape approach was never going to be able to compete with infinitely rich giants that could turn its ideas into just one tiny mass-produced component amongst thousands."

[RioGrande]

first baby!

[bootsielon]

1. Being bought by an important PC hardware manufacturer like nVidia, AMD or Intel, that way their chipset can be included with every high-end package or even with some low-end.

2. Being bought by an important software company that designs with PhysX in mind for improved performance. EA wouldn't be bad, but their games suck and are made for the lowest common denominator, so they wouldn't. Epic would be ideal. They could make the standards for physics cards to be compliant with, and spurr some competition from other manufacturers.

[Plasmana]

Your right, there are only two ways to succeed. They are:

1) Mirosoft DirectPhysics

2) OpenGL Physics

Although I'm willing to bet the first would be enough.

The problem with any hardware only approach is that it takes us back to the "old days" when developers had to write code to support multiple hardware vendors. Game developers can't afford to bet the kind of money it takes to develop a game today on a hardware vendor standard. Microsoft needs to get into the game to create a neutral standard that software developers can trust is future proofed.

I also find it doubtful that even a heavy hitter like EA could provide the necessary thrust to push something like this through. They could give the some traction, but EA would know that Microsoft could pull the carpet out from under them.

[JasonPC360PS3Wii]

If they want to survive, PC is big but its not big enough.

[Thugbot187]

If they can get into the next console, and this is the kind of edge consoles would need over PC's.

[sak500]

Unlike 3dfx which bought something revolutionary to the gaming world with their add-on cards, when the PCs were only using CPUs to do everything except display video on the screens. Even with that technology, initially only handful of companies supported it but once the public saw what these cards could do then the market exploded.

But now here with dual core and quad core cps becoming mainstream and with 8 and 16 on the horizon, functions like physics could easily be done on one of the cores. The need to add another expensive card in already brimmed to teeth gamer's pc does not make much sense. However if their code was built into DX10+ and automatically all games would benefit from having this card then maybe it might be picked up by the purest of hardcore gamer. However i would not, now only if another add-on card comes out which can provide even better gfx which rival CGI then i'm all game.

[THWIP]

...was a sh1tty game, with very unimpressive physics (nothing really "new"); even making it free, couldn't change that fact. THAT is why PhysX hasn't, and won't take off. It offers NOTHING that can't be done with the Havoc 2, or the Euphoria engine from Lucas Arts. When a $300-$400 console (360) can make use of such advanced physics, WITHOUT additional hardware, then there's obviously no need for PhysX in a PC gaming rig that will cost 2-3 times more, on average.

[WilliamRLBaker]

It didn't die because the hardware failed? lets see it was expensive and stupid no one used it, HArdly any games supported it, you could get cheaper and betta with havok physics and a good cpu...so yes the PHYSX hardware did fail.

[bootsielon]

Multi-core CPU
Multi-core GPU
Multi-core PPU
Multi-core AIPU

Each should have at least 32 cores. That might sound like a lot and too expensive. But supposedly massive market CPUs will hit the 80-core target, so thinking that a PS4 would have 128 cores including GPU, PPU and AIPU would be pretty much standard in 2012, unlike the Cell (today) which is pretty cutting-edge with 8-cores (okay, 1 is disabled, but it's still an 8-core processor).

That way PPUs (like PhysX) would have a chance to succeed.