[PATCH] D24973: [AMDGPU] Choose VMCNT, EXPCNT, LGKMCNT masks and shifts based on the isa version

[Konstantin Zhuravlyov via llvm-commits]

kzhuravl added a comment.

I am also planning to introduce 2 new functions for encoding/decoding *CNTs instead of doing pretty much the same math in a few different places, but in a separate patch


https://reviews.llvm.org/D24973