Quantum Volume (QV) is a single-number metric developed by IBM to measure the performance and computational power of a quantum computer. Unlike classical computers, where performance is largely measured in terms of clock speed or the number of transistors, quantum computers require a more holistic metric. Quantum Volume aims to capture multiple aspects of quantum computing performance in one figure.
Why Quantum Volume Matters
Quantum computers are not just about how many qubits they have. You can have a machine with 1000 qubits, but if they’re too noisy or not well-connected, the computer might not be usable for any meaningful calculation. That’s why a composite metric like Quantum Volume is essential.
What Quantum Volume Measures
Quantum Volume incorporates several factors that affect how well a quantum computer can run quantum circuits:
- Number of Qubits – Total count of quantum bits available.
- Gate Fidelity – Accuracy of quantum gates (how error-prone each operation is).
- Circuit Connectivity – How easily qubits can interact with each other.
- Parallelism – Ability to execute gates on multiple qubits at the same time.
- Measurement Fidelity – Accuracy of reading out the final state of qubits.
- Compiler Efficiency – How efficiently the software compiles high-level quantum logic into hardware-executable operations.
By considering all these, Quantum Volume gives a more realistic sense of a quantum computer’s real-world usefulness.
How Quantum Volume Is Tested
To determine a system’s Quantum Volume, researchers use random quantum circuits. These are artificial workloads composed of random gate sequences applied to randomly selected qubits. Here’s how it typically works:
- Choose a number of qubits nnn.
- Generate random quantum circuits of depth nnn, applied to nnn qubits.
- Compile and run these circuits on the quantum hardware.
- Evaluate success rate: If the hardware executes them successfully (above a certain statistical threshold), then the system is said to have a Quantum Volume of 2n2^n2n.
So, if a 6-qubit random circuit of depth 6 is reliably executed, the system has a Quantum Volume of 64.
Why It’s Logarithmic
Quantum Volume increases exponentially with the number of qubits and depth of the circuit. That’s why QV = 2n2^n2n — it’s not linear because the complexity of quantum circuits doesn’t grow linearly.
Quantum Volume vs Other Metrics
Other performance indicators in quantum computing include:
- Qubit Count – Measures how many qubits are available but says nothing about quality.
- Gate Fidelity – Measures error rate of operations but not system-wide capability.
- Quantum Supremacy Benchmarks – Usually task-specific and don’t generalize across systems.
Quantum Volume is more comprehensive, as it reflects both quality and quantity across diverse workloads.
Limitations of Quantum Volume
While QV is a valuable benchmark, it’s not perfect. Its limitations include:
- Hardware-specific bias: Some systems may be optimized specifically for high QV.
- Single-number simplicity: It hides granular details about what aspect is limiting performance.
- Synthetic workload: Random circuits may not reflect real-world quantum algorithms.
Still, it remains one of the best available tools for comparing quantum systems today.
Use Cases of Quantum Volume
- Vendor Comparisons: To compare IBM, IonQ, Rigetti, and others.
- System Tuning: Engineers can improve specific components (like compiler or calibration) to raise QV.
- Research Benchmarking: As a common baseline for academic studies.
- Customer Confidence: Enterprises can judge if a quantum system is mature enough for R&D or experimentation.
Historical Progress
- In 2019, IBM achieved a QV of 16.
- By 2021, QV reached 128 with IBM’s 27-qubit systems.
- Other companies like Honeywell (now Quantinuum) and IonQ also claim increasing QV with trapped ion technologies.
This shows that quantum systems are not just increasing in size but also in reliability and practical capability.