The Classical Advances Needed to Make Quantum Computers Tick
Quantum computers promise to solve problems beyond the most powerful supercomputers, but operating them takes a large amount of classical computing. As qubit counts rise, the classical hardware and software that support calibration and error correction must scale alongside them. Companies including Nvidia, Q-CTRL, IBM Quantum, Riverlane, and Google Quantum AI are building tools to accelerate these classical tasks. The article focuses on how calibration is being automated.
Key Takeaways
- Operating a quantum computer requires substantial classical computing for tasks like calibration and error correction.
- In April, Nvidia announced new AI-based software to accelerate the classical tasks that enable quantum computers.
- Q-CTRL built intelligent calibration software that analyzes each measurement, diagnoses failures, and adjusts before retrying.
- Done by hand, calibration can require a Ph.D. and take days or even weeks, which is not scalable.
- Calibration parameters drift over time, so Q-CTRL's software performs runtime recalibration to correct them.
Stats & Key Facts
- #Nvidia announced the new AI-based software in April
The Role of Classical Computing in Quantum
Qubits are temperamental and unreliable compared with classical chips.
- ›Classical chips operate flawlessly and can run trillions of operations without error.
- ›Qubits require regular calibration and complex error-correcting schemes to stay on track.
- ›Calibration and error correction are fundamentally classical problems requiring dedicated classical hardware.
As quantum computers grow, the scale of classical resources will need to rise in lockstep. For the foreseeable future, quantum computers are expected to be hybrid devices that pair qubits with a healthy dose of classical computing.
Industry Efforts to Support Quantum
Several companies are building classical tools for quantum systems.
- ›Sydney-based Q-CTRL developed an automatic calibration algorithm and now uses Nvidia's agent-based system.
- ›IBM Quantum is developing similar tools.
- ›Cambridge, England-based Riverlane develops quantum-error correction.
- ›Google Quantum AI is also working on these tools.
Adam Zalcman, a quantum software engineer at Google Quantum AI, said the cheapest and fastest way to execute most computer programs is to run them on a classical computer, even when a quantum computer is available. He expects every practical and efficient quantum-computer architecture to incorporate fast classical devices.
Why Tuning Quantum Hardware Is Hard
Qubits come in many physical forms, each needing careful calibration.
- ›Qubit types include superconducting circuits, trapped ions, neutral atoms, and individual photons.
- ›Calibration turns the bare metal of the hardware into a controllable qubit, says Jay Guilmart of Q-CTRL.
- ›The first stage, called bring up, finds each qubit's resonant frequency, coherence time, sensitivity to control pulses, and interaction strength with neighbors.
Done by hand, the process still requires someone with a Ph.D. and can take days or weeks, says Guilmart. Because it is not scalable, there is a growing drive to automate it.
Automating Calibration
Each calibration step depends on results from the previous step.
- ›Rather than running a predefined script, Q-CTRL built software that examines each measurement result.
- ›The software diagnoses failures and adjusts its approach before retrying.
- ›After each step it decides whether to proceed, go back, or recreate the step.
Guilmart described the loop: after each step the system analyzes the data and decides whether it is okay to proceed, must go back to the previous step, or must recreate the step.
Calibration Drift and Recalibration
Calibration is not a one-and-done process.
- ›Key parameters drift over time and gradually degrade performance.
- ›Q-CTRL's software performs runtime recalibration to nudge parameters back into place.
- ›There is a limit to how much on-the-fly adjustment is practical.
Frequently Asked Questions
Why do quantum computers need classical computing?
Calibration and error correction are fundamentally classical problems that require dedicated classical hardware, so quantum computers are expected to remain hybrid devices.
What did Nvidia announce in April?
Nvidia announced new AI-based software to accelerate the classical tasks that enable quantum computers.
Why is manual qubit calibration a problem?
Done by hand, calibration can require a Ph.D. and take days or even weeks, which does not scale as qubit counts rise.
How does Q-CTRL automate calibration?
Q-CTRL built software that examines each measurement, diagnoses failures, and adjusts its approach before retrying, deciding after each step whether to proceed, go back, or recreate the step.
What is runtime recalibration?
Because calibration parameters drift over time and degrade performance, Q-CTRL's software performs runtime recalibration to nudge parameters back into place.
As qubit counts grow, automating the classical work of calibration and error correction is becoming essential to making quantum computers practical.
Continue Learning
Comments
Sign in to join the conversation