Mobile Support

[ddutchie]

Hello Keijiro

Great implementation and optimization. Learning a lot about GPU compute by studying this code.

I am trying to get this to run on an iPhone X. I believe it could, just need to limit the kernel usage.

I might implement this myself and offer a drop down enum setting for mobile and desktop.

Have you any suggestions where to start in the code. I'll then make a pull request so you can see if it's worth it.

Regards

[keijiro]

I tried it with my iPhone X but immediately abandoned the idea as the performance was quite low.

I think the half-precision arithmetic pipeline should be utilized on mobiles to gain performance. Although I abandoned the idea based on the results from the FP16ComputeTest experimentation, it may become an effective approach on mobile GPUs.