Drive lines come at a cost.

Scaling superconducting quantum processors is not only a matter of adding qubits, but also of managing the wiring into the cryostat.

One promising direction is cryogenic multiplexing: sharing control lines across many qubits using fast, low-dissipation switches. This naturally raises an important question: if qubits have to wait for control, how much performance do we lose?

In our latest paper, Marvin Richter, Ingrid Strandberg, Simone Gasparinetti and Anton Frisk Kockum address this trade-off quantitatively, together with collaborators at Chalmers University of Technology.

The results are promising:
– Two-qubit gate couplers can share control lines with no penalty up to a connectivity-defined limit.
– For single-qubit gates, serialization overhead grows only logarithmically with the number of qubits per line.
– Relative to total circuit duration, the overhead remains nearly constant as the processor grows.

Read the paper here: https://journals.aps.org/prxquantum/abstract/10.1103/82cj-lfzy 👈