# QUANTUMLANDAUER — Result In Brief

Project ID:
628912

Funded under:
FP7-PEOPLE

Country:
Portugal

Domain:
Fundamental Research

## The physics of forgetting

Landauer's principle states that the erasure of information always takes a little energy. EU-funded physicists have found out how to minimise this unavoidable cost of quantum computation.

© Aigars Reinholds, Shutterstock

In 1961, the physicist Rolf Landauer argued that resetting one bit of information in a computer memory – for example, to set a binary digit to zero – is inevitably accompanied by the generation of heat. This principle provides a bridge between information theory and physics.

Landauer's principle sheds light on several issues arising in information processing. It implies a limit on how low the energy dissipation and thus, energy consumption of a computer can be. This lower limit was the focus of the QUANTUMLANDAUER (Beating Landauer's limit in the quantum regime) study.

During the project's lifetime, physicists explored possible departures from the Landauer framework that would permit more energy-efficient information erasure. They approached the problem of unavoidable heat dissipation from a probabilistic perspective.

Within the Landauer framework, information erasure relies on the interaction between an object and a thermal reservoir. This always results in a minimum quantity of heat being dissipated that is proportional to the entropy reduction incurred by the object. However, this lower bound is reached only in certain situations. Furthermore, it is not achievable for any given thermal reservoir. To determine the achievable minimum heat dissipation, physicists explicitly optimised all possible operators acting on the composite system of object and reservoir.

In addition, the QUANTUMLANDAUER team found a way to minimise the amount of energy transferred to a thermal reservoir as heat. This is applicable to the case where the probability of preparing an object in a pure state is smaller than the maximum probability of information erasure.

Using tools from the theory of majorisation, physicists characterised the equivalence class of the operators that bring the probability of information erasure to the desired value. Furthermore, they showed the trade-off between maximum probability of information erasure and minimum heat dissipation.

Heat dissipation in computer chips is hindering their miniaturisation. In quantum computers that exploit the laws of quantum physics to achieve greater processing power, the Landauer limit is already being confronted. QUANTUMLANDAUER findings, published in the New Journal of Physics were, therefore, timely.

Landauer's principle sheds light on several issues arising in information processing. It implies a limit on how low the energy dissipation and thus, energy consumption of a computer can be. This lower limit was the focus of the QUANTUMLANDAUER (Beating Landauer's limit in the quantum regime) study.

During the project's lifetime, physicists explored possible departures from the Landauer framework that would permit more energy-efficient information erasure. They approached the problem of unavoidable heat dissipation from a probabilistic perspective.

Within the Landauer framework, information erasure relies on the interaction between an object and a thermal reservoir. This always results in a minimum quantity of heat being dissipated that is proportional to the entropy reduction incurred by the object. However, this lower bound is reached only in certain situations. Furthermore, it is not achievable for any given thermal reservoir. To determine the achievable minimum heat dissipation, physicists explicitly optimised all possible operators acting on the composite system of object and reservoir.

In addition, the QUANTUMLANDAUER team found a way to minimise the amount of energy transferred to a thermal reservoir as heat. This is applicable to the case where the probability of preparing an object in a pure state is smaller than the maximum probability of information erasure.

Using tools from the theory of majorisation, physicists characterised the equivalence class of the operators that bring the probability of information erasure to the desired value. Furthermore, they showed the trade-off between maximum probability of information erasure and minimum heat dissipation.

Heat dissipation in computer chips is hindering their miniaturisation. In quantum computers that exploit the laws of quantum physics to achieve greater processing power, the Landauer limit is already being confronted. QUANTUMLANDAUER findings, published in the New Journal of Physics were, therefore, timely.