[llvm] [AMDGPU] Do not count implicit VGPRs in SIInsertWaitcnts (PR #109049)

[Matt Arsenault via llvm-commits]

================
@@ -1752,6 +1752,14 @@ bool SIInsertWaitcnts::generateWaitcntInstBefore(MachineInstr &MI,
         const bool IsVGPR = TRI->isVectorRegister(*MRI, Op.getReg());
         for (int RegNo = Interval.first; RegNo < Interval.second; ++RegNo) {
           if (IsVGPR) {
+            // Implicit VGPR defs and uses are never a part of the memory
+            // instructions description and usually present to account for
+            // super-register liveness.
+            // TODO: Most of the other instructions also have implicit uses
+            // for the liveness accounting only.
----------------
arsenm wrote:

COPY and the MOVs are the most common cases 

https://github.com/llvm/llvm-project/pull/109049