diff --git a/paddle/operators/rowwise_add_op.h b/paddle/operators/rowwise_add_op.h
index 232135c38de68d4002e044972b282b43a1374c72..771c5d7c0acf341027348355d4f548e2d0012b04 100644
--- a/paddle/operators/rowwise_add_op.h
+++ b/paddle/operators/rowwise_add_op.h
@@ -63,7 +63,7 @@ class RowwiseAddGradKernel : public framework::OpKernel {
 
     // https://eigen.tuxfamily.org/dox/unsupported/TensorBase_8h_source.html
     // colwise add
-    Eigen::array<int, 1> dims{{1}}; /* dimension to reduce */
+    Eigen::array<int, 1> dims{{0}}; /* dimension to reduce */
     EigenVector<T>::Flatten(*db).device(place) = OutGrad.sum(dims);
   }
 };
diff --git a/python/paddle/v2/framework/tests/test_rowwise_add_op.py b/python/paddle/v2/framework/tests/test_rowwise_add_op.py
index 29d72e850099734a9828ccceed47cc0a57fc3d6b..45d569da29d13cf8e2a3cb9d67c2d01e8b365453 100644
--- a/python/paddle/v2/framework/tests/test_rowwise_add_op.py
+++ b/python/paddle/v2/framework/tests/test_rowwise_add_op.py
@@ -20,7 +20,7 @@ class RowwiseAddGradOpTest(GradientChecker):
     def test_rowwise_add(self):
         op = create_op("rowwise_add")
         inputs = {
-            "X": np.random.uniform(0.1, 1, [10, 10]).astype("float32"),
+            "X": np.random.uniform(0.1, 1, [5, 10]).astype("float32"),
             "b": np.random.uniform(0.1, 1, [10]).astype("float32")
         }
         self.check_grad(op, inputs, set(["X", "b"]), "Out")