Intel backend

[Zaneham]

Ill be working on making this run on Intel GPU's so if anyone has the hardware to test any of my implementations on please feel free to reach out :-)

[Zaneham]

If anyone wants to drop what Intel GPU they have I'm more than happy to have a go at that first. If not, I think my grandparents can let me use their old desktop just to run the tests (this will benefit the two people running CUDA on GMAx4500, my Grandad hasn't updated his computer since the GFC).

[octaveoclx]

We have Intel Arc A770, can help your code running on it.

[octaveoclx]

In the meantime, is it possible for you to make an OpenCL on everything without LLVM? I normally stay away from CUDA regardless of a swathe of software running on CUDA.

[Zaneham]

Hello @octaveoclx, re your message on issue #16 as well as here I do plan on supporting Intel once I get some more ISA docs as well as shoring up my AMD support. Yes, Apple as well as any other GPU/TPU I can add over time. Thank you so much for your offer!

In terms of OpenCL thats an interesting problem and I have looked at it. I'm going to shore up this Compiler first before I move onto a project like that but that is something I can do.

[Zaneham]

In the meantime, is it possible for you to make an OpenCL on everything without LLVM? I normally stay away from CUDA regardless of a swathe of software running on CUDA.

Actually, I have thought about it and the BarraCUDA IR doesn't really care about where the code came from, doing OpenCL would just be swapping out a new frontend and making sure it hits the parser correctly. I'll keep this in my backpocket for now because I might even have to change my project name (lol!)