Parallel hybrid quantum-classical compute model

Hybrid computing in which classical hardware is augmented with quantum processing units (QPUs) is widely adopted in the noisy intermediate-scale quantum (NISQ) era, contending with many independent, often repetitive tasks in order to measure an observable comprised of many Pauli strings or to achieve reliable statistics in the presence of noise. Such computations are expected to greatly benefit from the simultaneous operation of an array of independent QPUs, akin to the parallel computing paradigm widespread in high performance computing (HPC). Here we report our findings in simulations with the variational quantum eigensolver algorithm, where different terms of the problem Hamiltonian are measured concurrently. The resulting groups of circuit instances are executed by virtual QPUs ran by standard message passing interface (MPI) processes, opening a door to parallel quantum computing based on the concerted operation of classical HPC nodes and virtual QPUs within a single hybrid virtual HPC platform, with potentially massive gains in efficiency and computational scale.