Running Conventional Automatic Speech Recognition on Memristor Hardware: A Simulated Approach