Dear friends, I would like to know what kind of calculations could benefit from using the Xeon Phi mathematical co-processors with Mathematica? Is it possible to run Mathematica computational kernels on each of Phi’s physical cores? Or is it that only functions, capable of auto-parallelization, can take advantage of the additional cores by Phi? Are there any use cases or any good examples of using Phi with Mathematica? Thank you!
I’m also interested on any information related to the possibility of making use of a Phi co-processors, from within Mathematica, and just with “pure” WL.
Also, beyond the OP questions, I’m interesting to know how Mathematica license scheme is adapted to this use type of processing scheme (although, if Mathematica doesn’t support Phi, it is probably no yet adapted). I mean, I can use any amount of GPU power (I think), as long as it is controlled by the amount of CPU licenses I have (and this is probably going to be amazing on the near future :-), but being Phi co-processors of CPU “type”, how would they be seen in terms of license?
I’m also interested on any information related to the possibility of making use of a Phi co-processors, from within Mathematica, and just with “pure” WL.
Also, beyond the OP questions, I’m interested to know how Mathematica license scheme is adapted to this type of processing scheme (although, if Mathematica doesn’t support Phi, it is probably no yet adapted). I mean, I can use any amount of GPU power (I think), as long as it is controlled by the amount of CPU licenses I have (and this is probably going to be amazing on the near future :-), but being Phi co-processors of CPU “type”, how would they be seen in terms of license?