[flang-commits] [flang] [Flang] Add opt-in affine loop optimization pipeline (PR #191854)

[Tom Eccles via flang-commits]

================
@@ -191,6 +194,54 @@ void createDefaultFIROptimizerPassPipeline(mlir::PassManager &pm,
   config.setRegionSimplificationLevel(
       mlir::GreedySimplifyRegionLevel::Disabled);
   pm.addPass(mlir::createCSEPass());
+
+  // Affine loop optimization pipeline (opt-in via --enable-affine-loop-opt).
+  if (enableAffineLoopOpt) {
+    pm.addPass(mlir::createCanonicalizerPass(config));
+    pm.addPass(mlir::createCSEPass());
+
+    addNestedPassToAllTopLevelOperations<PassConstructor>(
+        pm, fir::createSimplifyDoLoopPass);
+
+    pm.addPass(mlir::createCanonicalizerPass(config));
+    pm.addPass(mlir::createCSEPass());
+
+    pm.addPass(mlir::createLoopInvariantCodeMotionPass());
+    pm.addPass(fir::createLoopInvariantCodeMotion());
----------------
tblah wrote:

Wouldn't it be better to do LICM before SimplifyDoLoopPass so that SimplifyDoLoopPass is operating on a loop which is as simple as we can make it?

https://github.com/llvm/llvm-project/pull/191854