Quantum computers do not run in isolation; rather, they are embedded in quantum-classical hybrid architectures. In these setups, a quantum processing unit communicates with a classical device in near-real time. To enable efficient hybrid computations, it is mandatory to optimize quantum-classical hybrid code. To the best of our knowledge, no previous work on the optimization of hybrid code nor on metrics for which to optimize such code exists. In this work, we take a step towards optimization of hybrid programs by introducing seven optimization routines and three metrics to evaluate the effectiveness of the optimization. We implement these routines for the hybrid quantum language Quil and show that our optimizations improve programs according to our metrics. This lays the foundation for new kinds of hybrid optimizers that enable real-time collaboration between quantum and classical devices.
Quantum computers do not run in isolation; rather, they are embedded in quantum-classical hybrid architectures. In these setups, a quantum processing unit communicates with a classical device in near-real time. To enable efficient hybrid computations, it is mandatory to optimize quantum-classical hybrid code. To the best of our knowledge, no previous work on the optimization of hybrid code nor on metrics for which to optimize such code exists. In this work, we take a step towards optimization of hybrid programs by introducing seven optimization routines and three metrics to evaluate the effectiveness of the optimization. We implement these routines for the hybrid quantum language Quil and show that our optimizations improve programs according to our metrics. This lays the foundation for new kinds of hybrid optimizers that enable real-time collaboration between quantum and classical devices.
Quantum computing has become an increasingly relevant topic in computer science due to the development of functional quantum computers in recent years. Quantum processing units (QPUs) use quantum particles in the form of qubits as information carriers. By using properties of quantum particles, quantum algorithms have been developed that can solve problems in polynomial time, for which only exponentially scaling classical algorithms are known. QPUs will not run on their own in the foreseeable future, due to high error rates and low stability of qubits. Instead, quantum devices will be one component in a quantum-classical (hybrid) computing system (distributed system). Quantum computers embedded in distributed hybrid quantum-classical systems get instructions from a main CPU, execute them and report their results. To fully utilize quantum calculations, a QPU should be able to communicate with a CPU in real-time. This means the qubits information content should stay stable during communication time. In this thesis, we will look at quantum computing systems with real-time feedback between a QPU and a CPU. As quantum computers are going to be used in real-time hybrid systems, we need optimization routines for code running on these systems. Existing works on the optimization of quantum code focus on the optimization of quantum circuits only, and neglect potential classical calculations that are necessary on a hybrid distributed system. We want to examine which kind of optimization routines are possible and sensible for hybrid quantum-classical computations. We will evaluate some of todays quantum programming languages (QPLs), especially with respect to their ability to support and optimize real-time hybrid calculations. We will find that most languages support real-time calculations to some extent, but are more adapted to the programming approach of static circuit creation. Additionally, none of the QPLs offer optimization routines targeted at quantum-classical calculations. Therefore, we will examine options for real-time quantum-classical optimization. For this, we use the low-level QPL Quil. We will introduce optimization operations and apply them to real-time quantum-classical algorithms. We will present metrics to evaluate the performance of hybrid calculations. The results of applying our optimization operations to Quil code will be evaluated against our performance metrics. We will find that we are in principle able to optimize hybrid quantum-classical programs. We encourage more research in this field. QPLs will become more abstract in the future and need more compilation steps until they can be executed on hardware. This will make the optimization of quantum-classical hybrid code more relevant.
One of the primary focuses in the rapidly changing scenery of quantum computing is one of the most promising methodologies for transgressing the current feat of quantum technologies: a better mix of quantum and classical paradigms of computations.This paper suggests a new hybrid quantum-classical processor architecture that augments the computational speed and enhances efficiency by utilizing powers emanating from the quantum and classical processing units.The architecture proposed in the study incorporates quantum processing units with classical processing units, with CPUs interfaced tightly to support dynamic task allocation with a view toward computational requirements.Hybrid processor design caters to the performance of quantum algorithms that usually involve classical preprocessing and postprocessing, such as the Variational Quantum Eigensolver and Quantum Approximate Optimization Algorithm.These algorithms are essential for any complicated optimization, the simulation of quantum systems, and applications in machine learning.We provide an insightful processor architecture design, understanding the communication protocol details between QPU and CPU, the scheduling mechanism for hybrid tasks, and the optimization techniques employed to minimize data transfer overhead while maximizing computational throughput.A lot of these implementation challenges, like quantum error correction, managing the coherence time, and classical-quantum data conversion, are something that innovations in hardware design and software algorithms can achieve.In this paper, the performance of the hybrid processor is, for the first time, fully benchmarked based on simulation and experimental setups.Resultsshow that this hybrid approach outperforms standalone quantum and classical processors on specific computational tasks by ten times in optimization problems and five times for machine learning.These include case studies showing that for some problems, the processor can solve real-world problems much more efficiently than any existing quantum or classical system.Such case studies include quantum chemistry simulations, cryptographic tasks, and training machine learning models.We turn to the scalability and flexibility of the hybrid processor to extend the horizon of these case studies and to bridge the gap between the current quantum computing capabilities and the race toward fully functional quantum computers.This new hybrid quantum-classical processor architecture makes significant progress in quantum computing by providing a pragmatic, workable solution for harnessing quantum mechanics but using the maturity of classical computing.The results show how to proceed with the research involving the hybrid system and the many potential applications of such systems across optimization, cryptography, and artificial intelligence.This study is significant as it offers a practical approach to overcoming the limitations of current quantum and classical processors, providing a pathway to more efficient and scalable computational solutions.
The rapid evolution of Large Language Models (LLMs) has transformed natural language processing, enabling sophisticated applications across various sectors. However, the substantial computational demands associated with training and deploying LLMs result in significant energy consumption and carbon emissions. This study introduces an optimized hybrid quantum-classical framework that integrates variational quantum algorithms (VQAs) with accelerated classical learning techniques. By harnessing quantum computing for complex non-linear optimization and employing prompt learning to minimize full model retraining, the proposed approach enhances both computational efficiency and sustainability. Simulation outcomes indicate that the hybrid method can reduce energy usage by up to 30% and shorten computation time by 25% relative to conventional classical approaches, without diminishing model accuracy. These improvements are substantiated through quantitative analysis and visualized energy metrics. The adaptability of the framework supports its application in diverse areas, including sustainable energy management, supply chain optimization, and environmentally conscious transportation systems. Nevertheless, the broader implementation of such hybrid solutions remains constrained by current quantum hardware capabilities and integration challenges with classical infrastructure. The findings underscore the potential of hybrid quantum-classical optimization as a pathway toward sustainable AI development. Future research should prioritize advancements in quantum hardware reliability and interdisciplinary collaboration to accelerate practical adoption, thereby supporting global efforts in energy efficiency and environmental responsibility.
In recent years, the number of hybrid algorithms that combine quantum and classical computations has been continuously increasing. These two approaches to computing can mutually enhance each others’ performances thus bringing the promise of more advanced algorithms that can outmatch their pure counterparts. In order to accommodate this new class of codes, a proper environment has to be created, which enables the interplay between the quantum and classical hardware.For many of these hybrid processes the coherence time of the quantum computer arises as a natural time constraint, making it crucial to minimize the classical overhead. For ion-trap quantum computers however, this is a much less limiting factor than with superconducting technologies, since the relevant timescale is on the order of seconds instead of microseconds. In fact, we show that the operating time-scales of trapped-ion quantum computers are compatible with the execution speed of the Python programming language, enabling us to develop an interpreted scheme for real-time control of quantum computations. In particular, compilation of all instructions in advance is not necessary, unlike with superconducting qubits. This keeps the implementation of hybrid algorithms simple and also lets users benefit from the rich environment of existing Python libraries.In order to show that this approach of interpreted quantum-classical computations (IQCC) is feasible, we bring real-world examples and evaluate them in realistic benchmarks.
This paper presents a qualitative analysis of hybrid quantum-classical computing approaches aimed at solving complex optimization problems using near-term quantum processors. Hybrid algorithms leverage the strengths of quantum and classical computing to tackle computationally intensive tasks often constrained by current quantum hardware limitations. Through an extensive literature review and synthesis of recent empirical studies, we benchmark various hybrid algorithms, such as the Quantum Approximate Optimization Algorithm (QAOA) and Variational Quantum Eigensolver (VQE), focusing on their performance, scalability, and practical applicability on noisy intermediate-scale quantum (NISQ) devices. The study highlights the advantages, challenges, and future prospects of integrating hybrid quantum-classical computation in optimization domains, providing a comprehensive framework and qualitative insights into benchmarking methodologies and critical performance metrics.
Continuous improvements in quantum computing hardware are exposing the need for simultaneous advances in software. Large-scale implementation of quantum algorithms requires rapid and automated compilation routines such as circuit synthesis and optimization. As systems move towards fault-tolerance, programming frameworks and compilers must also be capable of compiling and optimizing programs comprising both classical and quantum code. This work takes a step in that direction by providing an implementation of Shor's factoring algorithm, compiled to elementary quantum gates using PennyLane and Catalyst, a library for quantum just-in-time (QJIT) compilation of hybrid workflows. We demonstrate that with QJIT compilation, the algorithm is compiled once per bit width of N, the integer being factored, even when N-specific optimizations are applied to circuit generation based on values determined at runtime. The implementation is benchmarked up to 32-bit N, and both the size of the compiled program and the pure compilation time are found to be constant (under 3 seconds on a laptop computer), meaning code generation becomes tractable even for realistic problem sizes.