We’ve released memQ DQC, an open-source, modality-agnostic framework for compiling and scheduling quantum programs across distributed quantum networks.
Distributed quantum computing offers a promising path toward scaling beyond the limitations of individual quantum processors. By connecting multiple processors through entanglement-based links, distributed systems may be able to execute programs that would not fit on any single device. Realizing that potential, however, requires software that can translate a conventional quantum circuit into a program that accounts for both the processors and the network connecting them.
memQ DQC is designed to make that process easier to explore.
Users begin with a quantum circuit and a description of the network on which they want to run it. The network can contain processors with different numbers of qubits, different patterns of local connectivity, and different connections between processors. The framework is modality-agnostic; it is not tied to a particular qubit technology or processor architecture, and users have full control over critical hardware parameters such as gate and coherence times.
To make network configuration more accessible, memQ DQC includes a graphical Quantum Network Constructor, which allows users to define processors, local qubit connectivity, and quantum links. Researchers can therefore experiment with different hardware and network designs without modifying the input circuit.
The compiler determines how the circuit’s qubits and operations should be distributed across the available processors. When an operation involves qubits on different processors, memQ DQC inserts the remote operations required to execute it over the network. It can also account for qubits moving between processors through state teleportation. The result is a distributed quantum program that respects the connectivity and capacity of the target system.
The scheduler then assigns local and network operations to time steps while preserving the dependencies of the original circuit. Entanglement generation can be modeled as either deterministic or stochastic, allowing users to investigate how uncertainty in the network affects execution. Together, these tools generate a distributed program, estimate its EPR-pair requirements, and produce a time-resolved execution schedule.T o get started with memQ DQC, visit our documentation.
Alongside the software, we’ve written a paper examining how choices across this workflow affect the resources required to run distributed quantum programs. Our results show that compilation strategies cannot be evaluated independently of the systems they target. In one benchmark, changing the connectivity within each processor increased entanglement requirements by more than 10×. We also found that the best-performing compilation strategy varied across circuits and network configurations.
These results highlight the importance of considering algorithms, software, processor architecture, and network design together. Decisions made at one layer of a distributed quantum system can have substantial effects elsewhere.
Code: https://github.com/memQGit/DistributedCompiler
Paper: https://arxiv.org/abs/2609.15728v1
Documentation: https://memq-dqc.readthedocs.io/en/latest/