diff --git a/paddle/fluid/inference/tensorrt/plugin/instance_norm_op_plugin.cu b/paddle/fluid/inference/tensorrt/plugin/instance_norm_op_plugin.cu
index 835dc4ac30e0b52e39dca11756dac3f391ca2846..a22714aa92f4935630c86384e90bd8e1ca3d79a4 100644
--- a/paddle/fluid/inference/tensorrt/plugin/instance_norm_op_plugin.cu
+++ b/paddle/fluid/inference/tensorrt/plugin/instance_norm_op_plugin.cu
@@ -111,6 +111,7 @@ int InstanceNormPlugin::enqueue(int batch_size, const void *const *inputs,
       handle_, CUDNN_BATCHNORM_SPATIAL_PERSISTENT, &alpha, &beta, x_desc_,
       x_ptr, y_desc_, y_ptr, b_desc_, scale_d, bias_d, 1., nullptr, nullptr,
       eps_, nullptr, nullptr);
+  return cudaGetLastError() != cudaSuccess;
 }
 
 }  // namespace plugin
diff --git a/paddle/fluid/inference/tests/api/CMakeLists.txt b/paddle/fluid/inference/tests/api/CMakeLists.txt
index 22bf27ce594963839b1cf245d273da9fd29c33ca..0125011b74a69db71121f9ead6e9c382ded0ce92 100644
--- a/paddle/fluid/inference/tests/api/CMakeLists.txt
+++ b/paddle/fluid/inference/tests/api/CMakeLists.txt
@@ -389,10 +389,9 @@ if(WITH_GPU AND TENSORRT_FOUND)
     inference_analysis_test(trt_split_converter_test SRCS trt_split_converter_test.cc
             EXTRA_DEPS ${INFERENCE_EXTRA_DEPS} 
             ARGS --infer_model=${TEST_SPLIT_CONVERTER_MODEL}/)
-    #TODO(peiyang): Fix this unitest failed on GCC8.
-    #inference_analysis_test(trt_instance_norm_test SRCS trt_instance_norm_converter_test.cc
-    #        EXTRA_DEPS ${INFERENCE_EXTRA_DEPS} 
-    #        ARGS --infer_model=${TEST_INSTANCE_NORM_MODEL}/)
+    inference_analysis_test(trt_instance_norm_test SRCS trt_instance_norm_converter_test.cc
+            EXTRA_DEPS ${INFERENCE_EXTRA_DEPS} 
+            ARGS --infer_model=${TEST_INSTANCE_NORM_MODEL}/)
     inference_analysis_test(test_analyzer_capi_gpu SRCS analyzer_capi_gpu_tester.cc
             EXTRA_DEPS ${INFERENCE_EXTRA_DEPS} paddle_fluid_c
             ARGS --infer_model=${TRT_MODEL_INSTALL_DIR}/trt_inference_test_models)