[Mlir-commits] [mlir] [mlir][CSE] Fix double-counting of numCSE statistic (PR #189802)

[llvmlistbot at llvm.org]

llvmbot wrote:


<!--LLVM PR SUMMARY COMMENT-->

@llvm/pr-subscribers-mlir

Author: lonely eagle (linuxlonelyeagle)

<details>
<summary>Changes</summary>

This PR fixes a regression where the numCSE statistic was being incremented twice for a single operation elimination. The numCSE counter is already internally incremented within the replaceUsesAndDelete function. Manually incrementing it again after the function call leads to an inaccurate total count.

---
Full diff: https://github.com/llvm/llvm-project/pull/189802.diff


1 Files Affected:

- (modified) mlir/lib/Transforms/CSE.cpp (-1) 


``````````diff
diff --git a/mlir/lib/Transforms/CSE.cpp b/mlir/lib/Transforms/CSE.cpp
index 289cfb8408c37..b8f3999d45f59 100644
--- a/mlir/lib/Transforms/CSE.cpp
+++ b/mlir/lib/Transforms/CSE.cpp
@@ -298,7 +298,6 @@ LogicalResult CSEDriver::simplifyOperation(ScopedMapTy &knownValues,
   // Look for an existing definition for the operation.
   if (auto *existing = knownValues.lookup(op)) {
     replaceUsesAndDelete(knownValues, op, existing, hasSSADominance);
-    ++numCSE;
     return success();
   }
 

``````````

</details>


https://github.com/llvm/llvm-project/pull/189802