diff --git a/Paddle b/Paddle
index 0759e99d8a4..d5b4570dd98 160000
--- a/Paddle
+++ b/Paddle
@@ -1 +1 @@
-Subproject commit 0759e99d8a4ba233850dbffe87954a2b6a628776
+Subproject commit d5b4570dd98bf59a1ea81fb04fc5ef62f8826e59
diff --git a/backends/npu/CMakeLists.txt b/backends/npu/CMakeLists.txt
index 19cdd447e33..c0cc808f2ce 100644
--- a/backends/npu/CMakeLists.txt
+++ b/backends/npu/CMakeLists.txt
@@ -54,6 +54,15 @@ include(third_party)
 add_dependencies(${CUSTOM_NPU_NAME} third_party)
 target_link_libraries(${CUSTOM_NPU_NAME} PRIVATE ${PADDLE_CORE_LIB})
 
+# testing
+if (WITH_TESTING)
+  set(PYTHON_SOURCE_DIR "${CMAKE_CURRENT_SOURCE_DIR}/../../python")
+  add_subdirectory(tests)
+  add_custom_command(OUTPUT ${CMAKE_CURRENT_BINARY_DIR}/tests/.timestamp
+      COMMAND cp -r ${CMAKE_SOURCE_DIR}/tests ${CMAKE_CURRENT_BINARY_DIR})
+  add_custom_target(python_tests ALL DEPENDS ${CMAKE_CURRENT_BINARY_DIR}/tests/.timestamp)
+endif()
+
 configure_file(${CMAKE_CURRENT_SOURCE_DIR}/setup.py.in
     ${CMAKE_CURRENT_BINARY_DIR}/setup.py)
 
diff --git a/backends/npu/kernels/sgd_kernel.cc b/backends/npu/kernels/sgd_kernel.cc
index c40cbd61b46..56797e12d63 100644
--- a/backends/npu/kernels/sgd_kernel.cc
+++ b/backends/npu/kernels/sgd_kernel.cc
@@ -22,7 +22,10 @@ void SGDKernel(const Context& dev_ctx,
                const phi::DenseTensor& param_var,
                const phi::DenseTensor& learning_rate,
                const phi::DenseTensor& grad_var,
-               phi::DenseTensor* param_out) {
+               paddle::optional<const phi::DenseTensor&> master_param,
+               bool multi_precision,
+               phi::DenseTensor* param_out,
+               phi::DenseTensor* master_param_out) {
   aclrtStream stream = static_cast<aclrtStream>(dev_ctx.stream());
   dev_ctx.template Alloc<T>(param_out);
 
diff --git a/backends/npu/tests/CMakeLists.txt b/backends/npu/tests/CMakeLists.txt
new file mode 100644
index 00000000000..29908b97b59
--- /dev/null
+++ b/backends/npu/tests/CMakeLists.txt
@@ -0,0 +1,33 @@
+# Copyright (c) 2022 PaddlePaddle Authors. All Rights Reserved.
+#
+# Licensed under the Apache License, Version 2.0 (the "License");
+# you may not use this file except in compliance with the License.
+# You may obtain a copy of the License at
+#
+# http://www.apache.org/licenses/LICENSE-2.0
+#
+# Unless required by applicable law or agreed to in writing, software
+# distributed under the License is distributed on an "AS IS" BASIS,
+# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+# See the License for the specific language governing permissions and
+# limitations under the License
+
+function(py_test_modules TARGET_NAME)
+    set(options SERIAL)
+    set(oneValueArgs "")
+    set(multiValueArgs MODULES DEPS ENVS)
+    cmake_parse_arguments(py_test_modules "${options}" "${oneValueArgs}" "${multiValueArgs}" ${ARGN})
+
+    add_test(NAME ${TARGET_NAME}
+        COMMAND ${CMAKE_COMMAND} -E env PYTHONPATH=${PYTHON_SOURCE_DIR}:$ENV{PYTHONPATH} ${py_test_modules_ENVS}
+        python ${PYTHON_SOURCE_DIR}/tools/test_runner.py ${py_test_modules_MODULES}
+            WORKING_DIRECTORY ${CMAKE_CURRENT_BINARY_DIR})
+
+    if (py_test_modules_SERIAL)
+        set_property(TEST ${TARGET_NAME} PROPERTY RUN_SERIAL 1)
+    endif()
+endfunction()
+
+py_test_modules(test_MNIST_model MODULES test_MNIST_model)
+
+add_subdirectory(unittests)
diff --git a/backends/npu/tests/unittests/CMakeLists.txt b/backends/npu/tests/unittests/CMakeLists.txt
new file mode 100644
index 00000000000..49715de4818
--- /dev/null
+++ b/backends/npu/tests/unittests/CMakeLists.txt
@@ -0,0 +1,20 @@
+# Copyright (c) 2022 PaddlePaddle Authors. All Rights Reserved.
+#
+# Licensed under the Apache License, Version 2.0 (the "License");
+# you may not use this file except in compliance with the License.
+# You may obtain a copy of the License at
+#
+# http://www.apache.org/licenses/LICENSE-2.0
+#
+# Unless required by applicable law or agreed to in writing, software
+# distributed under the License is distributed on an "AS IS" BASIS,
+# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+# See the License for the specific language governing permissions and
+# limitations under the License
+
+file(GLOB TEST_OPS RELATIVE "${CMAKE_CURRENT_SOURCE_DIR}" "test_*.py")
+string(REPLACE ".py" "" TEST_OPS "${TEST_OPS}")
+
+foreach(TEST_OP ${TEST_OPS})
+    py_test_modules(${TEST_OP} MODULES ${TEST_OP})
+endforeach(TEST_OP)
diff --git a/backends/npu/tests/unittests/test_softmax_op_npu.py b/backends/npu/tests/unittests/test_softmax_op_npu.py
new file mode 100644
index 00000000000..7f4083b82e6
--- /dev/null
+++ b/backends/npu/tests/unittests/test_softmax_op_npu.py
@@ -0,0 +1,120 @@
+# Copyright (c) 2022 PaddlePaddle Authors. All Rights Reserved.
+#
+# Licensed under the Apache License, Version 2.0 (the "License");
+# you may not use this file except in compliance with the License.
+# You may obtain a copy of the License at
+#
+#    http://www.apache.org/licenses/LICENSE-2.0
+#
+# Unless required by applicable law or agreed to in writing, software
+# distributed under the License is distributed on an "AS IS" BASIS,
+# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+# See the License for the specific language governing permissions and
+# limitations under the License.
+
+from __future__ import print_function
+
+import numpy as np
+import unittest
+import sys
+from tests.op_test import OpTest
+import paddle
+import paddle.fluid as fluid
+from paddle.fluid import core
+
+paddle.enable_static()
+SEED = 2021
+
+
+class TestSoftmax(OpTest):
+    def setUp(self):
+        self.set_npu()
+        self.place = paddle.CustomPlace('ascend', 0)
+        self.op_type = "softmax"
+        self.init_dtype()
+
+        x = np.random.random([3, 3]).astype(self.dtype)
+        np_out = np.exp(x) / np.sum(np.exp(x), axis=1, keepdims=True)
+        self.inputs = {'X': x}
+
+        self.attrs = {}
+        self.outputs = {'Out': np_out}
+
+    def set_npu(self):
+        self.__class__.use_custom_device = True
+        self.__class__.no_need_check_grad = True
+
+    def init_dtype(self):
+        self.dtype = np.float32
+
+    def test_check_output(self):
+        self.check_output_with_place(self.place)
+
+
+class TestSoftmaxNet(unittest.TestCase):
+    def _test(self, run_npu=True):
+        main_prog = paddle.static.Program()
+        startup_prog = paddle.static.Program()
+        main_prog.random_seed = SEED
+        startup_prog.random_seed = SEED
+        np.random.seed(SEED)
+
+        a_np = np.random.random(size=(4, 32)).astype('float32')
+        b_np = np.random.random(size=(4, 32)).astype('float32')
+        label_np = np.random.randint(2, size=(4, 1)).astype('int64')
+
+        with paddle.static.program_guard(main_prog, startup_prog):
+            a = paddle.static.data(name="a", shape=[4, 32], dtype='float32')
+            b = paddle.static.data(name="b", shape=[4, 32], dtype='float32')
+            label = paddle.static.data(
+                name="label", shape=[4, 1], dtype='int64')
+
+            c = paddle.multiply(a, b)
+            d = paddle.sqrt(c)
+
+            # 4 x 128
+            fc_1 = fluid.layers.fc(input=d, size=128)
+            # 4 x 2
+            prediction = fluid.layers.fc(input=fc_1, size=2)
+
+            # 4 x 2
+            prob = fluid.layers.softmax(prediction, axis=1)
+
+            cost = fluid.layers.cross_entropy(input=prob, label=label)
+            loss = fluid.layers.mean(cost)
+            sgd = fluid.optimizer.SGD(learning_rate=0.01)
+            sgd.minimize(loss)
+
+        if run_npu:
+            place = paddle.CustomPlace('ascend', 0)
+        else:
+            place = paddle.CPUPlace()
+
+        exe = paddle.static.Executor(place)
+        exe.run(startup_prog)
+
+        print("Start run on {}".format(place))
+        for epoch in range(100):
+
+            pred_res, loss_res = exe.run(
+                main_prog,
+                feed={"a": a_np,
+                      "b": b_np,
+                      "label": label_np},
+                fetch_list=[prediction, loss])
+            if epoch % 10 == 0:
+                print("Epoch {} | Prediction[0]: {}, Loss: {}".format(
+                    epoch, pred_res[0], loss_res))
+
+        return pred_res, loss_res
+
+    def test_npu(self):
+        cpu_pred, cpu_loss = self._test(False)
+        npu_pred, npu_loss = self._test(True)
+
+        self.assertTrue(np.allclose(npu_pred, cpu_pred, rtol=1e-2))
+        self.assertTrue(np.allclose(npu_loss, cpu_loss, rtol=1e-2))
+
+
+if __name__ == '__main__':
+    unittest.main()
diff --git a/python/__init__.py b/python/__init__.py
new file mode 100644
index 00000000000..513558501a0
--- /dev/null
+++ b/python/__init__.py
@@ -0,0 +1,13 @@
+#   Copyright (c) 2022 PaddlePaddle Authors. All Rights Reserved.
+#
+# Licensed under the Apache License, Version 2.0 (the "License");
+# you may not use this file except in compliance with the License.
+# You may obtain a copy of the License at
+#
+#     http://www.apache.org/licenses/LICENSE-2.0
+#
+# Unless required by applicable law or agreed to in writing, software
+# distributed under the License is distributed on an "AS IS" BASIS,
+# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+# See the License for the specific language governing permissions and
+# limitations under the License.
diff --git a/python/tests/__init__.py b/python/tests/__init__.py
new file mode 120000
index 00000000000..f4c1ea4faab
--- /dev/null
+++ b/python/tests/__init__.py
@@ -0,0 +1 @@
+../../Paddle/python/paddle/fluid/tests/unittests/__init__.py
\ No newline at end of file
diff --git a/python/tests/op_test.py b/python/tests/op_test.py
new file mode 120000
index 00000000000..3f65a6e3b4f
--- /dev/null
+++ b/python/tests/op_test.py
@@ -0,0 +1 @@
+../../Paddle/python/paddle/fluid/tests/unittests/op_test.py
\ No newline at end of file
diff --git a/python/tests/testsuite.py b/python/tests/testsuite.py
new file mode 120000
index 00000000000..c9eaa456157
--- /dev/null
+++ b/python/tests/testsuite.py
@@ -0,0 +1 @@
+../../Paddle/python/paddle/fluid/tests/unittests/testsuite.py
\ No newline at end of file
diff --git a/python/tests/white_list b/python/tests/white_list
new file mode 120000
index 00000000000..a5b9f97d72a
--- /dev/null
+++ b/python/tests/white_list
@@ -0,0 +1 @@
+../../Paddle/python/paddle/fluid/tests/unittests/white_list
\ No newline at end of file
diff --git a/python/tools/__init__.py b/python/tools/__init__.py
new file mode 120000
index 00000000000..393955a4b42
--- /dev/null
+++ b/python/tools/__init__.py
@@ -0,0 +1 @@
+../../Paddle/tools/__init__.py
\ No newline at end of file
diff --git a/python/tools/static_mode_white_list.py b/python/tools/static_mode_white_list.py
new file mode 120000
index 00000000000..908dbd07063
--- /dev/null
+++ b/python/tools/static_mode_white_list.py
@@ -0,0 +1 @@
+../../Paddle/tools/static_mode_white_list.py
\ No newline at end of file
diff --git a/python/tools/test_runner.py b/python/tools/test_runner.py
new file mode 120000
index 00000000000..278cedb0304
--- /dev/null
+++ b/python/tools/test_runner.py
@@ -0,0 +1 @@
+../../Paddle/tools/test_runner.py
\ No newline at end of file
diff --git a/scripts/paddle_ci.sh b/scripts/paddle_ci.sh
index d0c09af5b39..717df5b576b 100755
--- a/scripts/paddle_ci.sh
+++ b/scripts/paddle_ci.sh
@@ -42,10 +42,10 @@ function custom_npu_test() {
     pip install hypothesis
     pip install ${WORKSPACE_ROOT}/Paddle/build/python/dist/*whl
 
-    # custom_npu install
+    # custom_npu build and install
     cd ${WORKSPACE_ROOT}/PaddleCustomDevice/backends/npu
     mkdir build && cd build
-    cmake .. -DWITH_TESTING=ON -DWITH_KERNELS=ON
+    cmake .. -DWITH_TESTING=ON
     if [[ "$?" != "0" ]];then
         exit 7;
     fi
@@ -55,10 +55,9 @@ function custom_npu_test() {
     fi
     pip install dist/*.whl
 
-    # simple test now
+    # run ut
     ut_total_startTime_s=`date +%s`
-    cd ${WORKSPACE_ROOT}/PaddleCustomDevice/backends/npu/tests
-    python test_MNIST_model.py 
+    ctest --output-on-failure
     EXIT_CODE=$?
     ut_total_endTime_s=`date +%s`
     echo "TestCases Total Time: $[ $ut_total_endTime_s - $ut_total_startTime_s ]s"