please correct my understanding.

Even though python allows multi-threading it is similar to the normal mechanism(single thread) as at an instance only one thread has access to the interpreter. And it's like dividing the job into small pieces and keep them in queue till they get the workspace (interpreter). Also, additional lock acquire and release happens. And this leads to slowness compared the other two mechanisms( regular and multiprocessing). So multi-threading is not really parallelly doing the job.