diff --git a/tests/src/CMakeLists.txt b/tests/src/CMakeLists.txt
index 0037846e03..44becb0ed3 100644
--- a/tests/src/CMakeLists.txt
+++ b/tests/src/CMakeLists.txt
@@ -124,7 +124,9 @@ make_hip_executable (hipMathFunctionsHost hipMathFunctions.cpp hipSinglePrecisio
 make_hip_executable (hipMathFunctionsDevice hipMathFunctions.cpp hipSinglePrecisionMathDevice.cpp hipDoublePrecisionMathDevice.cpp)
 make_hip_executable (hipIntrinsics hipMathFunctions.cpp hipSinglePrecisionIntrinsics.cpp hipDoublePrecisionIntrinsics.cpp hipIntegerIntrinsics.cpp)
 make_hip_executable (hipPointerAttrib hipPointerAttrib.cpp) 
-make_hip_executable (hipMultiThreadStreams hipMultiThreadStreams.cpp) 
+make_hip_executable (hipMultiThreadStreams1 hipMultiThreadStreams1.cpp)
+make_hip_executable (hipMultiThreadStreams2 hipMultiThreadStreams2.cpp) 
+
 make_hip_executable (hipStreamL5 hipStreamL5.cpp)
 target_link_libraries(hipMathFunctionsHost m)
 
@@ -142,7 +144,8 @@ make_test(hipMemset --N 256M  --memsetval 0xa6 )  # big copy
 make_test(hipGridLaunch " " )
 make_test(hipEnvVarDriver " " )
 make_test(hipPointerAttrib " " )
-make_test(hipMultiThreadStreams " " )
+make_test(hipMultiThreadStreams1 " " )
+make_test(hipMultiThreadStreams2 " " )
 
 make_test(hipMemcpy  " " )
 make_test(hipMemcpyAsync  " " )
diff --git a/tests/src/hipMultiThreadStreams.cpp b/tests/src/hipMultiThreadStreams1.cpp
similarity index 100%
rename from tests/src/hipMultiThreadStreams.cpp
rename to tests/src/hipMultiThreadStreams1.cpp
diff --git a/tests/src/hipMultiThreadStreams2.cpp b/tests/src/hipMultiThreadStreams2.cpp
new file mode 100644
index 0000000000..c72d087db0
--- /dev/null
+++ b/tests/src/hipMultiThreadStreams2.cpp
@@ -0,0 +1,86 @@
+#include<iostream>
+#include"test_common.h"
+#include<thread>
+#define N 1000
+
+template<typename T>
+__global__ void Inc(hipLaunchParm lp, T *Array){
+int tx = hipThreadIdx_x + hipBlockIdx_x * hipBlockDim_x;
+Array[tx] = Array[tx] + T(1);
+}
+
+void run1(size_t size, hipStream_t stream){
+	float *Ah, *Bh, *Cd, *Dd, *Eh;
+
+	hipMallocHost(&Ah, size);
+	hipMallocHost(&Bh, size);
+	hipMalloc(&Cd, size);
+	hipMalloc(&Dd, size);
+	hipMallocHost(&Eh, size);
+
+	for(int i=0;i<N;i++){
+		Ah[i] = 1.0f;
+	}
+
+	hipMemcpyAsync(Bh, Ah, size, hipMemcpyHostToHost, stream);
+	hipMemcpyAsync(Cd, Bh, size, hipMemcpyHostToDevice, stream);
+	hipLaunchKernel(HIP_KERNEL_NAME(Inc), dim3(N/500), dim3(500), 0, stream, Cd);
+	hipMemcpyAsync(Dd, Cd, size, hipMemcpyDeviceToDevice, stream);
+	hipMemcpyAsync(Eh, Dd, size, hipMemcpyDeviceToHost, stream);
+	HIPCHECK(hipDeviceSynchronize());
+	HIPASSERT(Eh[10] == Ah[10] + 1.0f);
+}
+
+
+void run(size_t size, hipStream_t stream1, hipStream_t stream2){
+	float *Ah, *Bh, *Cd, *Dd, *Eh;
+	float *Ahh, *Bhh, *Cdd, *Ddd, *Ehh;
+
+	hipMallocHost(&Ah, size);
+	hipMallocHost(&Bh, size);
+	hipMalloc(&Cd, size);
+	hipMalloc(&Dd, size);
+	hipMallocHost(&Eh, size);
+	hipMallocHost(&Ahh, size);
+	hipMallocHost(&Bhh, size);
+	hipMalloc(&Cdd, size);
+	hipMalloc(&Ddd, size);
+	hipMallocHost(&Ehh, size);
+
+	hipMemcpyAsync(Bh, Ah, size, hipMemcpyHostToHost, stream1);
+	hipMemcpyAsync(Bhh, Ahh, size, hipMemcpyHostToHost, stream2);
+	hipMemcpyAsync(Cd, Bh, size, hipMemcpyHostToDevice, stream1);
+	hipMemcpyAsync(Cdd, Bhh, size, hipMemcpyHostToDevice, stream2);
+	hipLaunchKernel(HIP_KERNEL_NAME(Inc), dim3(N/500), dim3(500), 0, stream1, Cd);
+	hipLaunchKernel(HIP_KERNEL_NAME(Inc), dim3(N/500), dim3(500), 0, stream2, Cdd);
+	hipMemcpyAsync(Dd, Cd, size, hipMemcpyDeviceToDevice, stream1);
+	hipMemcpyAsync(Ddd, Cdd, size, hipMemcpyDeviceToDevice, stream2);
+	hipMemcpyAsync(Eh, Dd, size, hipMemcpyDeviceToHost, stream1);
+	hipMemcpyAsync(Ehh, Ddd, size, hipMemcpyDeviceToHost, stream2);
+	HIPCHECK(hipDeviceSynchronize());
+	HIPASSERT(Eh[10] = Ah[10] + 1.0f);
+	HIPASSERT(Ehh[10] = Ahh[10] + 1.0f);
+}
+
+int main(int argc, char **argv){
+	HipTest::parseStandardArguments(argc, argv, true);
+
+
+	hipStream_t stream[3];
+	for(int i=0;i<3;i++){
+		HIPCHECK(hipStreamCreate(&stream[i]));
+	}
+
+	const size_t size = N * sizeof(float);
+
+	std::thread t1(run1, size, stream[0]);
+	std::thread t2(run1, size, stream[0]);
+	std::thread t3(run, size, stream[1], stream[2]);
+	t1.join();
+//	std::cout<<"T1"<<std::endl;
+	t2.join();
+//	std::cout<<"T2"<<std::endl;
+	t3.join(); 
+	passed();
+}
+