[llvm-branch-commits] [llvm] [AMDGPU] Move `getAddressableNumVGPRs` to TargetParser (PR #226821)

[Nikhil Kotikalapudi via llvm-branch-commits]

================
@@ -196,12 +196,19 @@ LLVM_ABI unsigned getVGPREncodingGranule(Triple::SubArchType SubArch,
 LLVM_ABI unsigned getTotalNumVGPRs(GPUKind AK, bool IsWave32);
 LLVM_ABI unsigned getTotalNumVGPRs(Triple::SubArchType SubArch, bool IsWave32);
 
+/// Maximum number of VGPR blocks that can be allocated in dynamic VGPR mode.
----------------
nkotikal wrote:

Is this comment necessary? Variable name seems self-explanatory

https://github.com/llvm/llvm-project/pull/226821