diff --git a/paddle/operators/get_places_op.cc b/paddle/operators/get_places_op.cc
index 96a019ac799d3477f7319562ab0f7bc26f8a6560..6373df2158d75a6c1f1707c14e5965e5575c14d2 100644
--- a/paddle/operators/get_places_op.cc
+++ b/paddle/operators/get_places_op.cc
@@ -14,6 +14,9 @@ limitations under the License. */
 
 #include "paddle/framework/op_registry.h"
 #include "paddle/platform/place.h"
+#ifdef PADDLE_WITH_CUDA
+#include "paddle/platform/gpu_info.h"
+#endif
 
 namespace paddle {
 namespace operators {
@@ -26,7 +29,7 @@ class GetPlacesOp : public framework::OperatorBase {
       : OperatorBase(type, inputs, outputs, attrs) {}
   void Run(const framework::Scope &scope,
            const platform::DeviceContext &dev_ctx) const override {
-    auto use_gpu = Attr<bool>("use_gpu");
+    std::string device_type = Attr<std::string>("device_type");
     auto trainer_count = Attr<int>("trainer_count");
 
     auto out_var_name = Output("Out");
@@ -36,11 +39,16 @@ class GetPlacesOp : public framework::OperatorBase {
 
     auto &places = *(out_var->GetMutable<std::vector<platform::Place>>());
     places.resize(trainer_count);
-    if (use_gpu) {
+    if (device_type == "CUDA") {
+#ifdef PADDLE_WITH_CUDA
+      PADDLE_ENFORCE_LT(trainer_count, GetCUDADeviceCount());
       for (int i = 0; i < trainer_count; i++) {
         places.emplace_back(platform::GPUPlace(i));
       }
-    } else {
+#else
+      PADDLE_THROW("'GPUPlace' is not supported in CPU only device.");
+#endif
+    } else if (device_type == "CPU") {
       for (int i = 0; i < trainer_count; i++) {
         places.emplace_back(platform::CPUPlace());
       }
@@ -55,9 +63,11 @@ class GetPlacesOpProtoMaker : public framework::OpProtoAndCheckerMaker {
       : OpProtoAndCheckerMaker(proto, op_checker) {
     AddOutput("Out", "vector of Place");
     AddAttr<int>("trainer_count", "(int)trainer count").SetDefault(1);
-    AddAttr<bool>("use_gpu", "(bool)use gpu").SetDefault(false);
+    AddAttr<std::string>("device_type",
+                         "(string), deivce type can be \"CPU\" and \"CUDA\"")
+        .InEnum({"CPU", "CUDA"});
     AddComment(R"DOC(
-GetPlaces Operator.
+Returns a list of places based on flags. The list will be used for parallel execution.
 
 )DOC");
   }
diff --git a/python/paddle/v2/fluid/layers/utils.py b/python/paddle/v2/fluid/layers/utils.py
index aa6857e088b721bfef78e4ae684622e862e6838f..b71d8d93573244f85818e975235833cee9a8d5a7 100644
--- a/python/paddle/v2/fluid/layers/utils.py
+++ b/python/paddle/v2/fluid/layers/utils.py
@@ -8,14 +8,14 @@ from ..framework import Variable
 __all__ = ['get_places']
 
 
-def get_places(use_gpu, trainer_count):
+def get_places(trainer_count, device_type="CPU"):
     helper = LayerHelper('get_places', **locals())
     out_places = helper.create_tmp_variable(dtype=helper.input_dtype())
     helper.append_op(
         type='get_places',
         outputs={"Out": [out_places]},
         attrs={
-            "use_gpu": use_gpu,
+            "device_type": device_type,
             'trainer_count': trainer_count,
         })
 
diff --git a/python/paddle/v2/fluid/tests/test_layers.py b/python/paddle/v2/fluid/tests/test_layers.py
index 4a03ca68fffca556ba3747f64bb858d0e735d33e..82d7b03a8dbd2f2cd71b0d5b83e795faabaa6246 100644
--- a/python/paddle/v2/fluid/tests/test_layers.py
+++ b/python/paddle/v2/fluid/tests/test_layers.py
@@ -173,7 +173,7 @@ class TestBook(unittest.TestCase):
     def test_get_places(self):
         program = Program()
         with program_guard(program):
-            x = layers.get_places(use_gpu=True, trainer_count=4)
+            x = layers.get_places(trainer_count=4)
         print(str(program))