[llvm] AMDGPU: Add range attribute to mbcnt intrinsic callsites (PR #189191)

[via llvm-commits]

================
@@ -1422,11 +1422,31 @@ GCNTTIImpl::instCombineIntrinsic(InstCombiner &IC, IntrinsicInst &II) const {
 
     break;
   }
-  case Intrinsic::amdgcn_mbcnt_hi: {
+  case Intrinsic::amdgcn_mbcnt_hi:
     // exec_hi is all 0, so this is just a copy.
     if (ST->isWave32())
       return IC.replaceInstUsesWith(II, II.getArgOperand(1));
-    break;
+    [[fallthrough]];
+  case Intrinsic::amdgcn_mbcnt_lo: {
+    ConstantRange AccRange = computeConstantRange(II.getArgOperand(1),
+                                                  /*ForSigned=*/false);
+    if (AccRange.isFullSet())
+      return nullptr;
+
+    // TODO: Can raise upper bound by inspecting first argument.
----------------
ruiling wrote:

I mean we may not need https://github.com/llvm/llvm-project/commit/686987a540bc176bceaad43ffe530cb3e88796d5 after this change. I will have a try after you merge this one.
By inspecting the first argument, we have chance to raise the lower bound or reduce the upper bound. But cannot `raise upper bound`. Seems you made a typo here. Could you fix before submit?

https://github.com/llvm/llvm-project/pull/189191