[Mlir-commits] [mlir] [mlir] Add `arith-int-range-narrowing` pass (PR #112404)

[Jakub Kuderski]

================
@@ -216,6 +473,18 @@ void mlir::arith::populateIntRangeOptimizationsPatterns(
                DeleteTrivialRem<RemUIOp>>(patterns.getContext(), solver);
 }
 
+void mlir::arith::populateIntRangeNarrowingPatterns(
+    RewritePatternSet &patterns, DataFlowSolver &solver,
+    ArrayRef<unsigned> bitwidthsSupported) {
+  // Cmpi uses args ranges instead of results, run it with higher benefit,
----------------
kuhar wrote:

```suggestion
  // CmpI uses args ranges instead of results, run it with higher benefit,
```

https://github.com/llvm/llvm-project/pull/112404