diff --git a/cmake/CMakeLists.txt b/cmake/CMakeLists.txt
index 24269ddb00..b1fab715cc 100644
--- a/cmake/CMakeLists.txt
+++ b/cmake/CMakeLists.txt
@@ -612,7 +612,7 @@ if (onnxruntime_USE_CUDA)
   set(CMAKE_CUDA_FLAGS "${CMAKE_CUDA_FLAGS} -gencode=arch=compute_50,code=sm_50") # M series
   set(CMAKE_CUDA_FLAGS "${CMAKE_CUDA_FLAGS} -gencode=arch=compute_60,code=sm_60") # P series
   set(CMAKE_CUDA_FLAGS "${CMAKE_CUDA_FLAGS} -gencode=arch=compute_70,code=sm_70") # V series
-  set(CMAKE_CUDA_FLAGS "${CMAKE_CUDA_FLAGS} --default-stream per-thread")
+  set(CMAKE_CUDA_FLAGS "${CMAKE_CUDA_FLAGS} --default-stream legacy")
   if (NOT WIN32)
     set(CUDA_NVCC_FLAGS "${CUDA_NVCC_FLAGS} --compiler-options -fPIC")
   endif()
diff --git a/onnxruntime/core/providers/cuda/cuda_pch.h b/onnxruntime/core/providers/cuda/cuda_pch.h
index 6fd31e0c08..b32e0e59b7 100644
--- a/onnxruntime/core/providers/cuda/cuda_pch.h
+++ b/onnxruntime/core/providers/cuda/cuda_pch.h
@@ -2,9 +2,6 @@
 // Licensed under the MIT License.
 
 #pragma once
-#ifndef CUDA_API_PER_THREAD_DEFAULT_STREAM
-#define CUDA_API_PER_THREAD_DEFAULT_STREAM
-#endif
 #include <cuda.h>
 #include <cuda_runtime.h>
 #include <cublas_v2.h>