[Mlir-commits] [mlir] [mlir] Add direct vectorization lowering for `tensor.pack` ops (PR #78660)

[Han-Chung Wang]

================
@@ -1585,6 +1706,30 @@ vectorizeLinalgOpPrecondition(LinalgOp linalgOp,
   return success();
 }
 
+static LogicalResult
+vectorizePackOpPrecondition(tensor::PackOp packOp,
+                            ArrayRef<int64_t> inputVectorSizes) {
+  auto padValue = packOp.getPaddingValue();
+  if (padValue && getConstantIntValue(padValue) != std::nullopt) {
+    LDBG("pad value is not constant: " << packOp << "\n");
+    return failure();
+  }
+
+  ArrayRef<int64_t> resultTensorShape = packOp.getSourceType().getShape();
+  if (failed(isValidMaskedInputVector(resultTensorShape, inputVectorSizes)))
+    return failure();
+
+  if (llvm::any_of(packOp.getInnerTiles(), [](OpFoldResult v) {
+        std::optional<int64_t> res = getConstantIntValue(v);
+        return !res.has_value();
----------------
hanhanW wrote:

nit: we can merge these two lines to `return getConstantIntValue(v).has_value()`.

https://github.com/llvm/llvm-project/pull/78660