Software & Tools
Pulse-Level Control
Programming a quantum computer at the level of the analog microwave or laser pulses that implement each gate.
Definition
Behind every quantum gate in a circuit is a carefully shaped pulse of microwaves or laser light applied to the qubits. Pulse-level control means working with those pulses directly, choosing their shape, strength, and timing, instead of using the gates that the manufacturer has already calibrated. It gives researchers and advanced users a way to squeeze more performance out of hardware, to build new kinds of gates, and to study the physics of the device.
In depth
Superconducting qubits are driven with microwave pulses of tens of nanoseconds. A simple Gaussian pulse can excite unwanted transitions, such as leakage to the second excited state of a transmon, so engineers use shaped pulses such as DRAG, which adds a quadrature component to cancel them. Calibration finds the amplitude that gives a precise pi rotation, the frequency that matches the qubit, and the timing for two-qubit gates, and it must be repeated frequently because parameters drift over hours.
Direct pulse access has several uses. Researchers design faster or more robust gates using optimal control, implement gates that are not in the standard set, such as arbitrary two-qubit rotations, and study noise by probing the device. It enables error mitigation methods, for example stretching pulses to amplify noise for zero-noise extrapolation, and it allows hardware-efficient variational circuits tuned to the device's native interactions.
Access varies by vendor. IBM offered a pulse interface through Qiskit Pulse and OpenPulse for several years, but retired it from the main Qiskit library in 2025 as it shifted to higher-level primitives and error-mitigation services. Other companies such as Quantum Machines with its QUA language and Zurich Instruments provide control systems that give laboratories full pulse-level access, and some cloud devices allow pulse programming for approved users. Control-system latency and the integration of real-time feedback are active areas of development.
Technical Definition
A pulse schedule specifies waveforms, such as Gaussian, DRAG, or flat-top envelopes with carrier frequency and phase, played on hardware channels at defined times. A gate is a calibrated pulse sequence implementing a target unitary, with parameters tuned by experiments such as Rabi oscillations, Ramsey fringes, and randomized benchmarking. Optimal control methods such as GRAPE and CRAB design pulses that maximize fidelity under constraints.
Visual Explanation: An Analogy
A pianist playing printed music relies on the notes, while a composer can shape each key press, with exactly how hard and how long. Gates are the notes, and pulse-level control is how the key is pressed.
Key facts
| Item | Detail |
|---|---|
| Typical pulse lengths | Tens of nanoseconds for superconducting single-qubit gates; microseconds for ions |
| Common shapes | Gaussian and DRAG envelopes |
| Calibration | Rabi, Ramsey, and benchmarking experiments, repeated as parameters drift |
| Optimal control | GRAPE and CRAB algorithms design high-fidelity pulses |
| Tools | Quantum Machines QUA, Zurich Instruments, Qblox, and lab-built systems |
Where the idea comes from
Pulse control comes from nuclear magnetic resonance and atomic physics, where pulse sequences have been used since the 1950s. In quantum computing, the DRAG technique of Motzoi and coauthors in 2009 and the GRAPE algorithm from NMR (Khaneja et al., 2005) became standard tools.
Real-World Use Cases
- Calibrating and optimizing gates on new devices.
- Designing faster gates, robust gates, and custom operations.
- Implementing pulse-based error mitigation such as noise amplification.
- Studying qubit physics, noise, and crosstalk.
How it differs from related ideas
- Gate-level programming uses calibrated gates, while pulse-level programming controls the waveforms that realize them.
- Optimal control designs pulses numerically, while hand-tuned pulses rely on standard shapes.
- Pulse access gives more freedom but demands more expertise and re-calibration.
Common Misconceptions
- Pulse-level control does not automatically give better results. It needs careful calibration and can introduce errors.
- Most users do not need pulse access. Gate-level tools are adequate for typical algorithms.
- Pulses are analog, and small errors in amplitude and timing matter.
Frequently asked questions
What is pulse-level control?
Programming the actual microwave or laser pulses that implement quantum gates, instead of using pre-calibrated gates.
Why would someone need it?
To design better gates, calibrate devices, study noise, and implement specialized operations.
Is pulse access available on cloud devices?
It depends on the vendor. Some offer it to approved users, and laboratory control systems provide full access.
Related Terms
Keep exploring
- Algorithm: Quantum Error Mitigation Techniques
- Algorithm: Quantum Signal Processing (QSP)
- Hardware: IBM Heron (r1, r2, and r3)
- Hardware: IQM Garnet