[Mlir-commits] [mlir] [mlir] [memref] Compile-time memref.alloc Scheduling/Merging optimization (PR #95882)

[Krzysztof Drewniak]

================
@@ -0,0 +1,11 @@
+// RUN: mlir-opt -allow-unregistered-dialect -split-input-file -verify-diagnostics --merge-alloc %s
+
+func.func @block() {
+  %mref = memref.alloc() : memref<8 x f32>
+  %mref2 = memref.alloc() : memref<8 x f32>
+  // expected-error at +1 {{expecting RegionBranchOpInterface or LoopLikeOpInterface for merge-alloc}}
+  "some.block"() ({
+   ^bb0:
+    "some.use"(%mref) : (memref<8 x f32>) -> ()
+   }) : () -> ()
+}
----------------
krzysz00 wrote:

Nit: newline

https://github.com/llvm/llvm-project/pull/95882