[Mlir-commits] [mlir] [SCFToGPU] Convert scf.parallel+scf.reduce to gpu.all_reduce (PR #122782)

[Adam Siemieniuk]

================
@@ -648,6 +653,30 @@ ParallelToGpuLaunchLowering::matchAndRewrite(ParallelOp parallelOp,
       rewriter.setInsertionPointAfter(parent);
       leftNestingScope = true;
       seenSideeffects = false;
+    } else if (auto reduceOp = dyn_cast<scf::ReduceOp>(op)) {
+      // Convert scf.reduction op
+      auto parentLoop = op->getParentOfType<ParallelOp>();
+      if (!parentLoop || op->getOperands().size() != 1) {
+        return failure();
+      }
+      auto operand = op->getOperands().front();
+      auto newValue = cloningMap.lookupOrNull(operand);
+      if (!newValue) {
----------------
adam-smnk wrote:

`gpu.all_reduce` is more restrictive in terms of operand types. Extra checks are needed here.

https://github.com/llvm/llvm-project/pull/122782