Total: 1
Thermodynamic cost of communication is a major factor in the thermodynamic cost of real world computers, both biological and digital. However, little is known about the fundamental features of this cost. Addressing this hole in the literature, our first major contribution is a strictly positive lower bound on on the unavoidable entropy production (EP) in any physical system that implements a communication channel. Specifically, we derive a relationship between the rate of information transmission across a channel and the minimal thermodynamic cost incurred by any physical systems that implements that transmission. We use this relationship to show that under certain conditions, transmitting information through multiple, high-noise channels (reverse multiplexing) can be thermodynamically more efficient than using a single, low noise channel. We then address the thermodynamic costs inherent in the computational front-ends and back ends to communication channels, which are used to implement error correcting codes over those channels an essential component of modern communication systems. Our second major contribution is a strictly positive lower bound on the minimal EP of any physical system that implements those encoding and decoding algorithms. We use this bound to make a detailed comparison of the minimal thermodynamic costs across various linear codes and error rates. Because our results are like the second law, applying independently of the physical details of the system, they provide valuable insights into the trade offs between decoding error and thermodynamic efficiency. This gives a deeper understanding of the interplay between algorithmic design of error-correcting codes and physical constraints in communication systems.