8 comments

[ 3.2 ms ] story [ 29.5 ms ] thread
Is there a working implementation of this somewhere? I remember searching for it a while back and couldn't seem to find one.
I'd love to play with OpenCL on my i7. Since no Intel implementation exists, I'll stick to CUDA on my Tesla. I am not sure learning OpenCL is a worthwhile investment yet. I am skeptical about the ability of OpenCL to run efficiently on a very heterogeneous devices.

...then again...I suppose people said the same thing about OpenGL.

Does it get compiled to all targets, or is there some sort of intermediate code that gets compiled at the end point?