Distributing Quantum Computation Across Multiple NISQ Computers

The current quantum landscape is characterised by a wide variety of quantum computers, each with unique capabilities and limitations. Despite its rapid advancement, this field is still in its early stages and lacks essential software engineering principles such as resilience and customisability by design. To provide developers with a quantum technology stack that embeds these principles by design, we propose leveraging the heterogeneity of existing quantum computers. Specifically, we introduce a novel distribution methodology, termed shot-wise, to distribute the quantum computation workload of a single circuit execution (i.e., its shots) across multiple quantum computers. This shot-wise methodology is highly customisable, allowing for the em-bedding of user-specific and scenario-specific requirements into the distribution process. Requirements can be expressed through constraints, such as the geographical location of the QPUs, and their architecture, or ensuring or avoiding specific combinations of QPUs and quantum providers. Additionally, requirements can be defined by one or more objective functions to be minimised or maximised, such as reducing waiting time, computation cost, or energy footprint. By also combining constraints and objective functions, our shot-wise methodology enhances flexibility and improves the overall resilience of quantum computations against potential quantum computer failures. Preliminary results indicate that this approach offers error rates comparable to, or lower than, those of single-device execution even in all the tested scenarios with different numbers of qubits, families of circuits and number of QPUs to distribute the shots.