What is a quantum energetic advantage?
The development of quantum computers was primarily motivated by their computational advantage, that is, the fact that quantum computers can solve tasks which are too complex for classical computer. The notion of quantum energetic advantage is instead quite recent [A. Auffèves, Quantum technologies need a quantum energy initiative, PRX Quantum 3, 10.1103/prxquantum.3.020101 (2022)]. It denotes the situations where quantum computers solve tasks using less energy than the most efficient classical computer, even at the cost of losing the computational advantage. This advantage is particularly relevant for algorithms where quantum computers offer exponential speed-ups. A proof of concept of the existence of an energetic advantage was provided in [M. Fellous-Asiani et al., Optimizing resource efficiencies for scalable full-stack quantum computers, PRX Quantum 4, 040319 (2023)] for the case of a full-scale superconducting quantum computer. The search for these quantum energetic advantages has since become a new motivation to examine the resource consumption of quantum computing and quantum computers of various technologies.
Why can quantum computers have a quantum energetic advantage?
The most basic answer to that question also answers the question of why quantum computers can have a computational advantage: because quantum computers and classical computers work very differently. Let us consider the example of boson sampling—a task which involves generating samples from the output distribution of indistinguishable photons passing through a linear optical circuit. The time and energy taken by a classical computer for this task quickly increases exponentially with the size of the sample. Photonic quantum computers, instead, can perform this task natively, simply by sending indistinguishable photons through a large enough circuit. The energy consumed by the quantum computer consists mainly of cryogenic power and increases with the sampling size, but sub-exponentially, which results into an energetic quantum advantage.
How to find a quantum energetic advantage?
A methodology called Metric-Noise-Resource has been developed in order to determine the existence of a quantum energetic advantage for a given quantum computer and algorithm [M. Fellous-Asiani et al., Optimizing resource efficiencies for scalable full-stack quantum computers, PRX Quantum 4, 040319 (2023)]. The principle is to identify the sources of noise in the hardware, the resources which one can spend to mitigate those errors and to operate the computer, and a performance metric for a given task or algorithm. Then, the goal is to maximize the metric while minimizing the resources consumed, both for quantum computer and for the best available classical computer. If the quantum computer uses less energy for the same performance metric, then there is an energetic advantage.
Frequently asked questions about quantum energetic advantage
- Isn’t the connection between energy consumption and computation time a simple rule of thumb? True, if the computation time is known, then deducing the energy consumed is straightforward. The point of studying the energy efficiency alongside the computational efficiency is that there typically is a tradeoff between the accuracy of an algorithm and the energy efficiency. One then has to find the optimal working point of the device. For instance, it might not always be necessary to use the quantum computer at its maximum accuracy, and lowering the accuracy can lead to significant energy savings.
- Could a rebound effect lead to a higher energy consumption in the future, even if quantum computers are more energy efficient? As for any technology, the rebound effect is a risk. There is however a higher awareness of this effect today with the development of quantum technologies compared to the time when classical computers were developed, with initiatives such as the Quantum Energy Initiative (https://quantum-energy-initiative.org/).