> Existing no_std + no alloc crates written for other purposes can generally run on the GPU without modification.

Wow. That at first glance seems to unlock ALOT of interesting ideas.

I guess performance would be very different if things were initially assumed to run on a cpu

	▲	dvtkrlbs 6 days ago \| parent [-]
		I think it could be improved a lot by niche optimization passes on the codegen backend. Kinda like the autovectorization and similar optimizations on the CPU backends.