Physics processing units used to cost about as much as a middle range GPU (when I last checked, two years ago), and did little to alleviate the CPU of processing responsibilities. Has this changed? Are PPUs still more or less worthless?
closed as primarily opinion-based by random♦ Apr 23 '14 at 3:12
Many good questions generate some degree of opinion based on expert experience, but answers to this question will tend to be almost entirely based on opinions, rather than facts, references, or specific expertise. If this question can be reworded to fit the rules in the help center, please edit the question.
There used to be decent updates on Physics processing at this Hardforum FAQ.
The Nehalem roadmap is planning GPUs within the processor too.
An older article before AGEIA was acquired by NVIDIA
On the whole, I am starting to think Physx is well coupled with Graphics Cards.
PPUs are far from worthless if you need them for a specific purpose. Of course you need to use a PPU suited to the type of task you wish it to do. Not all PPUs are created equal. Have a good mind towards what it realistically needs to do for you and select accordingly.
In some cases a PPU will allow you make an otherwise dead slow application doing calculations quit zippy. Certainly not worthless but you need to have an intended use.
they can be a bit pricey as this is an "Over the barrel" market. If you need it, then you need it and you will pay happily in (Or at least they would like to think so)
Nowadays with grid computing you may not need to bother with a PPU though. You can put together a computing grid that can greatly outclass the PPU cards for the same money if you are willing to take the configuration hit of building a grid to crunch the numbers for you.