[llvm] [AMDGPU] Add llvm.amdgcn.dead intrinsic (PR #123190)

[Diana Picus via llvm-commits]

rovka wrote:

> > Unfortunately, this kind of phi node will get optimized away and the backend won't be able to figure out that it's ok to use the active lanes of `%inactive.vgpr*` inside `shader`.
> 
> Does InstSimplify/InstCombine optimize away a phi if all but one of its inputs is undefined?

FWIW phi node elimination makes similar decisions (i.e. even if the middle-end leaves the phi nodes in place, we still get the long live ranges after phi-node-elim).

https://github.com/llvm/llvm-project/pull/123190