Fully-integrated control stacks for quantum computing, part 1: stack overview

Reaching NISQ applications hinges on improvements in the gate fidelity and qubit number. Qblox supports this with time-efficient, ultralow-noise, and cost-effective control stacks. We introduce the Cluster system which incorporates processors capable of sequencing pulses, their parameters, and measurement operations in real time. This architecture speeds up experiments by orders of magnitude as it avoids the overhead caused by software-controlled loops. This speed-up is realized by multi-parameter real-time pulse modification by on-board data processing (integrating, averaging, binning) of readout signals and storing up to 131072 measurement results per experimental run. The state-of-the-art signal noise level (14 nV/√Hz @ 1 MHz) supports improved gate fidelities and the low gain and offset drift (a few ppm/K) reduces the need for recalibrations. The Cluster supports many qubit platforms with its wide frequency range from DC to 18.5 GHz while occupying less volume than 1 liter per controlled qubit. Quantify -an open-source python framework- manages the hardware stack, which allows hybrid scheduling of gate-level and pulse-level descriptions. This full-stack approach opens a fast track for gate optimizations and scaling efforts towards running NISQ applications.