Natively Dynamic CUDA GPU's? repost from nzone

Your thoughts on Nvidia GPU’s utilizing haskell environments and API’s.
Should yield some faster processing results :ph34r:

Maybe even native support?
:mellow:

It’s more a matter of GHC or some other compiler writing a back end for PTX, or a bunch of people putting together a decent CUDA module together and maintaining it on hackage… not visa-versa.