代码唯一的区别是place = fluid.CPUPlace()改成了place = fluid.CUDAPlace(0)
在执行exe.run(test_program,xxx)的时候，报错：
temporary_buffer::allocate: get_temporary_buffer failed
terminate called after throwing an instance of 'thrust::system::system_error'
what(): device free failed: unspecified launch failure
Aborted