[Mlir-commits] [mlir] [mlir][x86] Add extra patterns for AMX nano-kernel lowering (acc with zero constant) (PR #222513)

[llvmlistbot at llvm.org]

================
@@ -27,6 +27,24 @@ using namespace mlir::x86;
 
 namespace {
 
+// Return true if the operation is a constant dense element
+// attribute.
+static bool isZeroVectorConstant(Operation *op) {
+  auto constantOp = dyn_cast<arith::ConstantOp>(op);
+  if (!constantOp)
+    return false;
+
+  auto vectorType = dyn_cast<VectorType>(constantOp.getType());
+  if (!vectorType)
+    return false;
----------------
z6jh89mtvm-alt wrote:

1159018553

https://github.com/llvm/llvm-project/pull/222513