[llvm] [AMDGPU] Support i8/i16 GEP indices when promoting allocas to vectors (PR #175489)

[Harrison Hao via llvm-commits]

================
@@ -86,10 +86,11 @@ static cl::opt<unsigned>
                             "when sorting profitable allocas"),
                    cl::init(4));
 
-// We support vector indices of the form (A * stride) + B
+// We support vector indices of the form ((A >> shift) * stride) + B
----------------
harrisonGPU wrote:

Okay, I have updated.

https://github.com/llvm/llvm-project/pull/175489