Add hipStreamSynchronize to hipModuleLoadDataMultThreadOnMultGPU test

hipStreamDestroy() by definition doesn't guarantee waiting for all stream operations.

Change-Id: Ibfd21c48d71741ebcbcb9898c8b5ac235bdc570b


[ROCm/hip commit: 2299459ce0]

projects/hip/tests/src/runtimeApi/module/hipModuleLoadDataMultThreadOnMultGPU.cpp

@@ -94,6 +94,8 @@ void run(const std::vector<char>& buffer, int deviceNo) {
             HIP_LAUNCH_PARAM_END};
   HIPCHECK(hipModuleLaunchKernel(Function, 1, 1, 1, LEN, 1, 1, 0, stream, NULL, (void**)&config));
 
+  HIPCHECK(hipStreamSynchronize(stream));
+
   HIPCHECK(hipStreamDestroy(stream));
 
   HIPCHECK(hipModuleUnload(Module));