diff --git a/ttnn/cpp/ttnn/operations/core/to_layout/to_layout_op.cpp b/ttnn/cpp/ttnn/operations/core/to_layout/to_layout_op.cpp
index f948be785c3..d0757378ea6 100644
--- a/ttnn/cpp/ttnn/operations/core/to_layout/to_layout_op.cpp
+++ b/ttnn/cpp/ttnn/operations/core/to_layout/to_layout_op.cpp
@@ -62,7 +62,9 @@ Tensor to_layout_impl_on_device(
 
     if (!requires_padding_change(tensor_arg, layout)) {
         if (layout == ttnn::ROW_MAJOR_LAYOUT) {
-            TT_FATAL(!dtype.has_value(), "dtype cannot be specified when converting to ROW_MAJOR_LAYOUT!");
+            TT_FATAL(
+                !dtype.has_value() || dtype.value() == tensor_arg.dtype(),
+                "dtype cannot be specified when converting to ROW_MAJOR_LAYOUT!");
             return ttnn::untilize(tensor_arg, output_memory_config, use_multicore_untilize);
         }
         return ttnn::tilize(tensor_arg, output_memory_config, dtype, use_multicore_tilize);
@@ -71,7 +73,9 @@ Tensor to_layout_impl_on_device(
     auto tensor_shape = tensor_arg.get_logical_shape();
 
     if (layout == ttnn::ROW_MAJOR_LAYOUT) {
-        TT_FATAL(!dtype.has_value(), "dtype cannot be specified when converting to ROW_MAJOR_LAYOUT!");
+        TT_FATAL(
+            !dtype.has_value() || dtype.value() == tensor_arg.dtype(),
+            "dtype cannot be specified when converting to ROW_MAJOR_LAYOUT!");
 
         if (tensor_arg.is_sharded()) {
             const auto memory_config = tensor_arg.memory_config();