diff --git a/orttraining/orttraining/models/pipeline_poc/main.cc b/orttraining/orttraining/models/pipeline_poc/main.cc
index a80aa08e9f..7f423b9ee0 100644
--- a/orttraining/orttraining/models/pipeline_poc/main.cc
+++ b/orttraining/orttraining/models/pipeline_poc/main.cc
@@ -105,11 +105,12 @@ int main(int argc, char* argv[]) {
 
   InferenceSession session_object{so, *env};
 
+  Status st;
   CUDAExecutionProviderInfo xp_info{static_cast<OrtDevice::DeviceId>(world_rank)};
-  session_object.RegisterExecutionProvider(std::make_unique<CUDAExecutionProvider>(xp_info));
+  st = session_object.RegisterExecutionProvider(std::make_unique<CUDAExecutionProvider>(xp_info));
+  ORT_ENFORCE(st == Status::OK(), "MPI rank ", world_rank, ": ", st.ErrorMessage());
 
   std::string model_at_rank;
-  Status st;
   if (world_rank == 0) {
     st = session_object.Load(params.model_stage0_name);
     ORT_ENFORCE(st == Status::OK(), "MPI rank ", world_rank, ": ", st.ErrorMessage());
@@ -214,4 +215,4 @@ int main(int, char* []) {
   ORT_NOT_IMPLEMENTED("P2P demo currently requires CUDA to run.");
 }
 
-#endif
\ No newline at end of file
+#endif