diff --git a/tensorflow/python/kernel_tests/pooling_ops_test.py b/tensorflow/python/kernel_tests/pooling_ops_test.py
index 224405ace67..e87cccac1e6 100644
--- a/tensorflow/python/kernel_tests/pooling_ops_test.py
+++ b/tensorflow/python/kernel_tests/pooling_ops_test.py
@@ -39,6 +39,15 @@ from tensorflow.python.platform import test
 from tensorflow.python.platform import tf_logging
 
 
+def GetDeviceScope(self, use_gpu=False):
+  if context.executing_eagerly():
+    if use_gpu and test.is_gpu_available():
+      return ops.device("GPU:0")
+    return ops.device("CPU:0")
+  else:
+    return self.session(use_gpu=use_gpu)
+
+
 def GetTestConfigs(include_nchw_vect_c=False):
   """Get all the valid tests configs to run.
 
@@ -802,12 +811,7 @@ class PoolingTest(test.TestCase):
       # Generate numbers in a narrow range, so that there are many duplicates
       # in the input.
       tensor_input = np.random.random_integers(0, 3, input_shape).astype(dtype)
-      def get_device_scope():
-        if context.executing_eagerly():
-          return ops.device("GPU:0")
-        else:
-          return self.cached_session(use_gpu=True)
-      with get_device_scope():
+      with self.cached_session(use_gpu=False):
         t = constant_op.constant(tensor_input, shape=input_shape)
         _, argmax_op = nn_ops.max_pool_with_argmax(t, ksize, strides, padding)
         argmax = self.evaluate(argmax_op)
@@ -840,7 +844,7 @@ class PoolingTest(test.TestCase):
                [True, False, [0, 1, 3, 5, 0, 2, 6, 8]]]
 
     for use_gpu, include_batch_in_index, argmax_exp in configs:
-      with self.session(use_gpu=use_gpu):
+      with GetDeviceScope(self, use_gpu=use_gpu):
         t = constant_op.constant(tensor_input, shape=[2, 3, 3, 1])
         out_op, argmax_op = nn_ops.max_pool_with_argmax(
             t,
@@ -868,7 +872,7 @@ class PoolingTest(test.TestCase):
                [True, False, [0, 1, 3, 5, 0, 2, 6, 8]]]
 
     for use_gpu, include_batch_in_index, argmax in configs:
-      with self.session(use_gpu=use_gpu):
+      with GetDeviceScope(self, use_gpu):
         orig_in = constant_op.constant(orig_input, shape=[2, 3, 3, 1])
         t = constant_op.constant(tensor_input, shape=[2, 2, 2, 1])
         argmax_t = constant_op.constant(