Cunqa: Distributed Quantum Computing Emulator for HPC Enables Evaluation of Three Quantum Communication Models

Distributed quantum computing represents a crucial step toward realising the full potential of quantum computation, and researchers are increasingly exploring how to integrate these future systems into existing high-performance computing environments. Jorge Vázquez-Pérez, Daniel Expósito-Patiño, and Marta Losada, along with colleagues at the Galicia Supercomputing Center and the Universidad de Santiago de Compostela, now present CUNQA, a novel open-source emulator designed to simulate distributed quantum computers within standard HPC infrastructure. This achievement allows scientists to test and evaluate different distributed quantum computing models, including those with no communication, classical communication, and quantum communication, before the hardware becomes available, effectively bridging the gap between theory and practical implementation. By implementing these models and utilising the well-known Quantum Phase Estimation algorithm, CUNQA represents the first tool capable of emulating all three distributed quantum computing schemes within a high-performance computing context, paving the way for future advancements in quantum acceleration and scalable quantum computation.

There is a growing momentum toward treating quantum computers as accelerators, integrating them into the heterogeneous architectures of high-performance computing (HPC) environments. This work introduces CUNQA, an open-source distributed quantum computing (DQC) emulator designed for HPC environments, allowing researchers to test, evaluate, and study DQC before physical quantum computers become widely available. CUNQA implements three DQC models, no-communication, classical-communication, and quantum-communication, enabling comprehensive investigation of different architectural approaches. This work details the programming considerations, emulation techniques, and implementation specifics of this innovative tool.

Hybrid Quantum-Classical Computing Architectures

Current research demonstrates a convergence between high-performance computing (HPC) and quantum computing (QC), with researchers moving beyond theoretical studies to integrate QC with existing HPC infrastructure. The central theme is building hybrid architectures that augment HPC with quantum capabilities for specific tasks, rather than replacing it entirely. A significant challenge lies in developing the necessary software infrastructure to support this integration, particularly in programming models, compilers, and simulation tools. Simulation and emulation are crucial due to the limited availability of actual quantum hardware, allowing researchers to test algorithms and explore system designs.

Building the necessary system-level infrastructure, including resource management, scheduling, and communication protocols, is also a major focus. Researchers are investigating how to optimally partition computations between classical and quantum processors, and how to design hardware and software that facilitate this co-execution. Optimizing performance and scalability is key, addressing communication overhead and data transfer bottlenecks. While still in its early stages, research is beginning to explore potential applications of hybrid HPC-QC systems in areas like materials science, drug discovery, finance, and machine learning. The research landscape is rapidly evolving, with a strong emphasis on building the software and system-level infrastructure needed to realize the potential of hybrid HPC-QC systems, shifting the focus from theoretical exploration to practical implementation and integration.

CUNQA Emulates Distributed Quantum Computing Architectures

Scientists have developed CUNQA, a groundbreaking emulator designed to explore the future of distributed quantum computing within high-performance computing environments. This tool emulates three distinct distributed quantum computing models, no-communication, classical-communication, and quantum-communication, allowing researchers to test and study these architectures before physical quantum computers are built. CUNQA is the first tool capable of emulating all three schemes within a standard HPC infrastructure, utilizing existing resources to simulate quantum processing units. The core of CUNQA lies in its “virtual QPUs”, classical processes running on HPC resources that faithfully simulate the behaviour of real quantum processors, functioning as classical accelerators within the system.

The team supports three simulators: AerSimulator version 0. 16. 0, a specific commit of MQT-DDSIM, and a custom CunqaSimulator developed for testing. Experiments demonstrate CUNQA’s ability to emulate co-located and on-node architectures. The system utilizes a layered software stack, with resource management handled by the user, and a middleware layer connecting the interface with the virtual QPUs. Implemented in both Python and C++, CUNQA delivers a powerful new tool for quantum computing research, enabling exploration of complex distributed architectures and paving the way for future advancements in quantum computation.

Distributed Quantum Computing Emulator for HPC Environments

CUNQA represents a significant advance in the field of distributed quantum computing, delivering the first open-source emulator designed to operate within high-performance computing environments. This tool faithfully replicates the behaviour of distributed quantum systems, allowing researchers to test and evaluate different architectures before physical implementation becomes feasible. CUNQA distinguishes itself by emulating three distinct distributed quantum computing models, those employing no communication, classical communication, and quantum communication, offering a comprehensive platform for exploration. The system achieves this emulation through simulated quantum processing units, or virtual QPUs, which run as classical processes on existing HPC resources.

This approach allows CUNQA to model the interaction of multiple quantum processors without requiring access to actual quantum hardware. Importantly, the tool focuses on emulating the models of distributed quantum computation, rather than the individual QPUs themselves, and is specifically designed for co-located and on-node architectures within an HPC context. The authors acknowledge that CUNQA currently discards the standalone model of distributed quantum computing, considering it incompatible with the accelerator paradigm. Nevertheless, CUNQA provides a valuable platform for the quantum computing community, enabling detailed investigation of distributed quantum algorithms and architectures within the constraints of current technology.