A Comprehensive Guide to Quantum Benchmarking: Exploring 3 Essential Characteristics, Techniques, and Challenges

As quantum computing evolves and progresses, it becomes increasingly important to establish standardized benchmarks for comparing and measuring the performance of quantum devices. In the paper “Defining Standard Strategies for Quantum Benchmarks,” researchers outline the essential characteristics of quantum benchmarks and the importance of error suppression and mitigation techniques. This blog post delves into the key aspects of quantum benchmarking, optimization techniques, and the challenges associated with establishing accurate, fair, and transparent performance measurements.

Quantum Computing Quantum Benchmarking

Characteristics of Quantum Benchmarks

According to the paper, any benchmark protocol should follow these characteristics:

  1. Randomized: Ensures that the benchmark accurately represents the device’s performance without bias or predisposition towards specific problems.
  2. Well-defined: Clear rules and guidelines should be in place to ensure that the benchmark is executed consistently across devices and over time.
  3. Holistic: The benchmark should capture all key attributes of the device being tested, including scale, quality, and speed.
  4. Device-independent: The benchmark should be applicable to any gate-based quantum computing architecture, allowing for fair comparisons across devices and generations.

Types of Benchmarks

The paper discusses several types of benchmarks, including:

  1. Application-oriented benchmarking circuits: These circuits are designed to represent different types of quantum algorithms and problems relevant to real-world applications. By testing a variety of circuits, application-oriented benchmarking can capture the overall performance of a device across different types of applications.
  2. Circuit Layer Operations per Second (CLOPS): A benchmark that measures the speed of a quantum processor by looking at the number of quantum volume (QV) layers per second that can be executed.
  3. Mirror circuits: A class of quality benchmarks that measure how well a circuit of interest can be executed on a target device. Mirror quantum volume is a fast heuristic for evaluating the potential quality of QV over a large quantum processor.

Benchmarking vs. Diagnostics

The paper also highlights the distinction between benchmarking and diagnostic protocols. While benchmarking measures average performance and provides results that can be compared across devices, diagnostic methods give insight into a device’s performance on similarly structured problems. Diagnostic methods, due to their specificity, are not suitable as standards for benchmarking.


Benchmarking protocols are designed to measure the average performance of a quantum device across a range of tasks or algorithms. The primary goal of benchmarking is to provide a means of comparing the performance of different devices or device iterations objectively.

Key attributes of benchmarking:

  1. Generalization: Benchmarking focuses on obtaining a generalized understanding of a quantum device’s performance rather than concentrating on specific tasks or error types.
  2. Comparability: Benchmarks are constructed to enable fair comparisons across different devices, technologies, and over time. The device-independent nature of benchmarks allows researchers and industry professionals to evaluate and rank quantum devices based on their overall performance.
  3. Standardization: To ensure consistent execution and results, benchmarking protocols must follow well-defined guidelines and rules.


Diagnostic methods, on the other hand, are highly sensitive to specific types of errors and are tailored to provide insights into the performance of a quantum device for a particular setting or problem. Diagnostics delve deeper into the device’s behavior, allowing researchers to identify areas of improvement or troubleshoot issues.

Key attributes of diagnostics:

  1. Specificity: Diagnostic protocols focus on identifying particular errors or assessing performance on similarly structured problems. This focus allows researchers to gain a detailed understanding of the device’s behavior in specific scenarios.
  2. Characterization: Diagnostics help researchers obtain a clear characterization of the device’s performance in a particular setting, enabling them to identify potential areas of improvement or uncover device limitations.
  3. Problem-driven: Diagnostic methods are often driven by the need to solve specific problems or achieve certain goals. As such, they are tailored to address the unique challenges and requirements of a specific problem or task.

While benchmarking and diagnostics serve different purposes, they are complementary in the broader context of quantum computing research and development. Benchmarking provides a means of comparing and evaluating quantum devices based on their general performance, while diagnostic methods offer valuable insights into device behavior for specific problems or tasks. By utilizing both benchmarking and diagnostic protocols, researchers and industry professionals can obtain a comprehensive understanding of quantum device performance, leading to more informed decision-making and fostering innovation in the field of quantum computing.

Optimization Techniques and Error Mitigation

Error suppression and mitigation techniques play a critical role in improving the quality of benchmarks. The paper presents examples of constant-time optimization techniques, such as dynamic decoupling and optimal compiling of gates, which can enhance benchmark quality with minimal overhead.

In contrast, error mitigation techniques with exponential overhead can significantly improve results, but at a much greater cost in time. It’s crucial to report the overhead incurred by using error mitigation techniques alongside the resulting improvement in quality to avoid misrepresenting device performance.

Challenges and Potential Gaming Strategies

The article warns against potential gaming strategies that could artificially improve the quality of benchmark results. For example, filtering low-frequency outcomes or using a-priori information to reduce errors could lead to misleading results. To address these issues, the paper proposes a set of rules for optimizing benchmarks, ensuring the integrity and transparency of the process.

Developing Industry Standards for Benchmarking

As the field of quantum computing matures, it is vital for industry players and researchers to adopt standardized benchmarking protocols. The development and widespread adoption of these protocols will enable the accurate comparison of quantum devices, drive innovation, and ensure that progress is measured objectively.


“Defining Standard Strategies for Quantum Benchmarks” is a vital step toward measuring and comparing the performance of quantum devices accurately. By outlining the necessary characteristics of quantum benchmarks, highlighting the importance of error suppression and mitigation techniques, and addressing potential gaming strategies, this paper sets the foundation for future benchmarking efforts in the field of quantum computing. With standardized benchmarks in place, the quantum computing community can work towards achieving a more transparent and accurate understanding of the true potential of these groundbreaking devices.


Hello! I'm a Quantum Computing Scientist based in Silicon Valley with a strong background in software engineering. My blog is dedicated to sharing the tools and trends I come across in my research and development work, as well as fun everyday anecdotes.

Articles: 22

Leave a Reply