Atos introduces “Q-score”, the first universal quantum metrics, applicable to all programmable quantum processors. Atos’ Q-score measures a quantum system’s effectiveness at handling real-life problems, those which cannot be solved by traditional computers, rather than simply measuring its theoretical performance. Q-score reaffirms Atos’ commitment to deliver early and concrete benefits of quantum computing. Over the past five years, Atos has become a pioneer in quantum applications through its participation in industrial and academic partnerships and funded projects, working hand-in-hand with industrials to develop use-cases which will be able to be accelerated by quantum computing.
“Faced with the emergence of a myriad of processor technologies and programming approaches, organizations looking to invest in quantum computing need a reliable metrics to help them choose the most efficient path for them. Being hardware-agnostic, Q-score is an objective, simple and fair metrics which they can rely on,” said Elie Girard, Atos CEO. “Since the launch of ‘Atos Quantum’ in 2016, the first quantum computing industry program in Europe, our aim has remained the same: advance the development of industry and research applications, and pave the way to quantum superiority.”
What does Q-score measure?
Today the number of qubits (quantum units) is the most common figure of merit for assessing the performance of a quantum system. However, qubits are volatile and vastly vary in quality (speed, stability, connectivity, etc.) from one quantum technology to another (such as supraconducting, trapped ions, silicon and photonics), making it an imperfect benchmark tool. By focusing on the ability to solve well-known combinatorial optimization problems, Atos Q-score will provide research centers, universities, businesses and technological leaders with explicit, reliable, objective and comparable results when solving real-world optimization problems.
Q-score measures the actual performance of quantum processors when solving an optimization problem, representative of the near-term quantum computing era (NISQ – Noisy Intermediate Scale Quantum). To provide a frame of reference for comparing performance scores and maintain uniformity, Q-score relies on a standard combinatorial optimization problem, the same for all assessments (the Max-Cut Problem, similar to the well-known TSP – Travelling Salesman Problem, see below). The score is calculated based on the maximum number of variables within such a problem that a quantum technology can optimize (ex: 23 variables = 23 Q-score or Qs).
Atos will organize the publication of a yearly list of the most powerful quantum processors in the world based on Q-score. Due in 2021, the first report will include actual self-assessments provided by manufacturers.
Based on an open access software package, Q-score is built on 3 pillars:
- Application driven: Q-score is the only metrics system based on near-term available quantum algorithms and measures a quantum system’s capacity to solve practical operational problems;
- Openness and ease of use: Universal and free, Q-score benefits from Atos’ technology-neutral approach. Its software package, including tools and methodology, does not require heavy computation power to calculate the metrics;
- Objectiveness and reliability: Atos combines a hardware-agnostic, technology-agnostic approach with a strong expertise in algorithm design and optimization acquired working with major industry clients and technology leaders in the quantum field. The methodology used to build Q-score will be made public and open to assessment.
A free software kit, which enables Q-score to be run on any processor will be available in Q1 2021. Atos invites all manufacturers to run Q-score on their technology and publish their results.
Thanks to the advanced qubit simulation capabilities of the Atos Quantum Learning Machine (Atos QLM), its powerful quantum simulator, Atos is able to calculate Q-score estimates for various platforms. These estimates take into account the characteristics publicly provided by the manufacturers. Results range around a Q-score of 15 Qs, but progress is rapid, with an estimated average Q-score dating from one year ago in the area of 10 Qs, and an estimated projected average Q-score dating one year from now to be above 20 Qs.
Q-score has been reviewed by the Atos Quantum Advisory Board, a group of international experts, mathematicians and physicists authorities in their fields, which met on
December 4, 2020.
Understanding Q-score using the Travelling Salesman Problem (TSP)
Today’s most promising application of quantum computing is solving large combinatorial optimization problems. Examples of such problems are the famous TSP problem and the less notorious but as important Max-Cut problem.
Problem statement: a traveler needs to visit N number of cities in a round-tour, where distances between all the cities are known and each city should be visited just once. What is the absolute shortest possible route so that he visits each city exactly once and returns to the origin city?
Simple in appearance, this problem becomes quite complex when it comes to giving a definitive, perfect answer taking into account an increasing number of N variables (cities). Max-Cut is a more generic problem, with a broad range of applications, for instance in the optimization of electronic boards or in the positioning of 5G antennas.
Q-score evaluates the capacity of a quantum processor to solve these combinatorial problems.
Q-score, Quantum Performance, and Quantum Superiority
While the most powerful High Performance Computers (HPC) worldwide to come in the near term (so called “exascale”) would reach an equivalent Q-score close to 60, today we estimate, according to public data, that the best Quantum Processing Unit (QPU) yields a Q-score around 15 Qs. With recent progress, we expect quantum performance to reach Q-scores above 20 Qs in the coming year.
Q-score can be measured for QPUs with more than 200 qubits. Therefore, it will remain the perfect metrics reference to identify and measure quantum superiority, defined as the ability of quantum technologies to solve an optimization problem that classical technologies cannot solve at the same point in time.
As per the above, Atos estimates quantum superiority in the context of optimization problems to be reached above 60 Qs.