Breakthrough in quantum computing stabilises qubits in milliseconds
A team of researchers has developed a faster way to stabilise superconducting qubits—a key challenge in quantum computing. Traditional calibration methods struggle to keep up with rapid changes in qubit behaviour. The new approach uses on-FPGA processing to adjust parameters in milliseconds, preventing performance drops over time.
The project, led by Malthe A. Marciniak, Rune T. Birke, and Johann B. Severin, targets parameter drift—a problem where qubit properties shift faster than standard CPU-based systems can correct. Their solution moves the entire workflow—pulse generation, data collection, analysis, and adjustments—onto an FPGA chip. This eliminates delays caused by sending data back and forth to a central processor.
Tests showed qubit coherence times (T1 values) fluctuating within just two seconds. The team's system sustained over 74,000 consecutive recalibrations, keeping gate operations precise through continuous, closed-loop optimisation. Tasks like readout calibration, spectroscopy, and pulse tuning now run far quicker than before.
One benchmark, Clifford randomised gate testing, completed in just 107 milliseconds. The research also mapped out the best balance between sampling speed and measurement accuracy, helping operators choose optimal settings. By cutting latency, the method ensures qubits stay stable even as their environment changes.
The on-FPGA workflow allows quantum systems to recalibrate almost instantly, maintaining high performance without interruptions. This could reduce errors in quantum computations and extend the usable time of fragile qubit states. The findings provide a practical path for handling drift in real-world quantum devices.
