at last intel released the xeon phi accelerator. the interesting part of it is not the speedup it's the programming. in the ideal case you get a speedup of 2x or 3x without reprogramming( needs a recompile) because it can use the same x86 bytecode.

it takes me wonder if sesi invest time in to it because if it's true it will save allot of time in the opencl optimization.

some infromation:
http://semiaccurate.com/2012/11/12/what-does-it-take-to-code-for-a-xeon-phi/ [semiaccurate.com]

the card prices are in the 2k$ area same as the nvidia tesla cards.

Yes, or to use for mantra batch rendering…

I m also interesting to have a comment from sesi about this card and houdini

have fun
Thomas

It looks like it will be hard to ignore Phi, as computing facilities seem to be enthusiastic about it. I'm sure SESI already lurks around.

+1!

I believe the Phi has a lot of potential being (on paper) easier to develop for than OpenCL/Cuda. Mantra support would be great to speed up heavy renders.