[Mlir-commits] [mlir] [memref] Simplify loads from reinterpret_cast of 1D contiguous memrefs (PR #188459)

[Andrzej Warzyński]

================
@@ -220,3 +221,301 @@ func.func private @negative_plain_copy(%src : memref<1x1xf32>,
   : memref<1x1xf32> to memref<1x1xf32>
   return
 }
+
+
+// -----
+
+//===----------------------------------------------------------------------===//
+// Positive tests
+//===----------------------------------------------------------------------===//
+
+/// For rank-1 MemRefs, expansion/collapsing may be considered on either side.
+
+// CHECK-LABEL: func.func private @expand_scalar(
+// CHECK-SAME:    %[[SRC:.*]]: memref<1xi64>) {
+func.func private @expand_scalar(%src : memref<1xi64>) {
+  // CHECK:       %[[C0:.*]] = arith.constant 0 : index
+  %c0 = arith.constant 0 : index
+  // CHECK-NOT:   memref.reinterpret_cast
+  %reinterpret_cast = memref.reinterpret_cast %src
+    to offset: [0], sizes: [1, 1, 1], strides: [1, 1, 1] : memref<1xi64>
+      to memref<1x1x1xi64>
----------------
banach-space wrote:

[nit] With your current formatting, it is a bit hard to see what the input and the output shapes are. Could use the following format consistently throughout the patch? It will make it easier to compare examples.
```suggestion
  %reinterpret_cast = memref.reinterpret_cast %src
    to offset: [0], sizes: [1, 1, 1], strides: [1, 1, 1] 
    : memref<1xi64> to memref<1x1x1xi64>
```

https://github.com/llvm/llvm-project/pull/188459