is it possible to implement CUDA python within grasshopper iron python?

hello vet hoppers 

i have been looking into GPU accelerated parallel computing a lot lately. 

There are  a number of slow components when numbers increase. 

to name a few, surface split, join curves (pure evil!!), kangaroo reset (my personal record: 11hours for reset of kangaroo)

but mostly i am very envious of game developers and animators. I have seen a set up with 7 GTX TITANs hooked up as a rendering machine. I have never seen a performance like that from one computer before.

i see developments with PyCUDAor other CUDA support for python. before i turn into a computer science major and apply to standford or something, i want to ask , "can Iron Python work with CUDA 5.0 or can it work with Anaconda ?

what does it takes before one can write CUDAmalloc inside grasshopper ironpython component and it actaully works?