Challenges
Datasets
Workspaces
Papers
Discussions
Leaderboard
Log inSign up
Challenges
Datasets
Workspaces
Papers
Discussions
Leaderboard
Blog
Job Board
Q3AS

© 2026 Aqora Quantum S.A.S.

TermsPrivacyLegal Notice

IEEE Transactions on Nanotechnology

A Review of Ising Machines Implemented in Conventional and Emerging Technologies

Tingting Zhang, Qichao Tao, Bailiang Liu, Andrea Grimaldi, Eleonora Raimondo, Manuel Jiménez, M.J. Avedillo, Juan Núñez, B. Linares-Barranco, Teresa Serrano‐Gotarredona, Giovanni Finocchio, Jie Han

Published Jan 1, 2024• 35 citations
Publisher

Abstract

Ising machines have received growing interest as efficient and hardware-friendly solvers for combinatorial optimization problems (COPs). They search for the absolute or approximate ground states of the Ising model with a proper annealing process. In contrast to Ising machines built with superconductive or optical circuits, complementary metal-oxide-semiconductor (CMOS) Ising machines offer inexpensive fabrication, high scalability, and easy integration with mainstream semiconductor chips. As low-energy and CMOS-compatible emerging technologies, spintronics and phase-transition devices offer functionalities that can enhance the scalability and sampling performance of Ising machines. In this article, we survey various approaches in the process flow for solving COPs using CMOS, hybrid CMOS-spintronic, and phase-transition devices. First, the methods for formulating COPs as Ising problems and embedding Ising formulations to the topology of the Ising machine are reviewed. Then, Ising machines are classified by their underlying operational principles and reviewed from a perspective of hardware implementation. CMOS solutions are advantageous with denser connectivity, whereas hybrid CMOS-spintronic and phase-transition device-based solutions show great potential in energy efficiency and high performance. Finally, the challenges and prospects are discussed for the Ising formulation, embedding process, and implementation of Ising machines.

DOI
10.1109/tnano.2024.3457533

Related papers

A Review of Simulation Algorithms of Classical Ising Machines for Combinatorial optimization

Tingting Zhang, Qichao Tao, Bailiang Liu +1 more

Combinatorial optimization problems are difficult to solve due to the space explosion in an exhaustive search. Using Ising model-based solvers can efficiently find near-optimal solutions by minimizing the energy of a nonlinear Hamiltonian system. In contrast to Ising machines based on quantum mechanics, classical Ising machines using conventional technologies, such as the complementary metal-oxide-semiconductor, offer efficient implementations with competitive performance. In this paper, we briefly review recently developed simulation algorithms of classical Ising machines. These algorithms are classified by considering various inherent mechanisms in the simulation of physical phenomena. Then, strategies to improve the simulation efficiency are discussed by generalizing their characteristics and behaviours. These simulation algorithms are key for improving the efficiency of classical Ising machines in solving combinatorial optimization problems.

IEEE International Symposium on Circuits and Systems proceedings•May 28, 2022•17 citations

VLSI Implementation of an Annealing Accelerator for Solving Combinatorial Optimization Problems

Yuan‐Ho Chen, Hsin-Tung Hua, Chin-Fu Nien +1 more

Although quantum computing is expected to supplant traditional computing in several application fields, its adoption is hampered by temperature and economic constraints. To overcome these hurdles, researchers have proposed complementary metal-oxide-semiconductor (CMOS) annealing circuits. These circuits draw inspiration from quantum computing algorithms such as quantum annealing and aim to achieve near-quantum benefits by leveraging traditional CMOS technologies. This paper introduces an Ising-model-based hardware architecture that can be applied to combinatorial optimization problems (COPs). With its ability to express quadratic unconstrained binary optimization (QUBO) formulations as polynomials, the Ising model facilitates the encapsulation of multiple solutions and mapping onto fully connected architecture. The proposed annealing accelerator utilizes traditional circuit technologies, including pseudo-random number generators (PRNGs), to realize the required algorithms. The chip proposed herein, implemented using Taiwan Semiconductor Manufacturing Company (TSMC) 90-nm CMOS technology, operates at 50 MHz and covers an area of $3.24{\mathrm{m}}{{\mathrm{m}}^2}$3.24mm2. Experimental results demonstrate the excellent performance of this annealing accelerator in terms of area and power consumption, indicating its promise for use in solving COPs rapidly.

IEEE Nanotechnology Magazine•Apr 18, 2024•0 citations

Device‐to‐System Perspectives on Scalable Ising Machines for Physics‐Driven Computing

Sangwook Youn, Hwiho Hwang, Jaeung Lee +2 more

ABSTRACT Combinatorial optimization problems (COPs) are fundamental to numerous applications but remain challenging due to their exponentially large solution spaces and the limitations of conventional computing architectures. Ising machines have emerged as a promising alternative by mapping optimization problems onto the energy minimization of the Ising Hamiltonian. Various hardware implementations, including quantum annealers, digital annealers, and coherent Ising machines, have been explored based on distinct physical mechanisms. More recently, emerging device‐based approaches have gained significant attention for their potential to enable scalable and energy‐efficient optimization. In this review, we present a comprehensive overview of Ising machines from both theoretical and hardware perspectives, with particular emphasis on three emerging‐device‐based platforms: crossbar array‐based, stochastic p‐bit, and oscillator‐based Ising machines. We establish a comparative device‐to‐system perspective that traces how distinct physical and device properties translate into computational primitives, hardware architectures, and ultimately system‐level scalability. Platform‐specific bottlenecks, including array nonidealities and peripheral overhead in crossbar‐based systems, device variability and probability‐generation overhead in p‐bit systems, and synchronization and coupling‐network complexity in oscillator‐based systems, are quantitatively and qualitatively analyzed. Finally, we discuss future directions toward scalable and energy‐efficient Ising hardware through device‐to‐system co‐optimization across devices, circuits, architectures, and algorithms.

Advanced Quantum Technologies•Oct 1, 2026•0 citations

Solving Boltzmann optimization problems with deep learning

Fiona Knoll, John T. Daly, Jess Meyer

Abstract Decades of exponential scaling in high-performance computing (HPC) efficiency is coming to an end. Transistor-based logic in complementary metal-oxide semiconductor (CMOS) technology is approaching physical limits beyond which further miniaturization will be impossible. Future HPC efficiency gains will necessarily rely on new technologies and paradigms of computing. The Ising model shows particular promise as a future framework for highly energy-efficient computation. Ising systems are able to operate at energies approaching thermodynamic limits for energy consumption of computation. Ising systems can function as both logic and memory. Thus, they have the potential to significantly reduce energy costs inherent to CMOS computing by eliminating costly data movement. The challenge in creating Ising-based hardware is in optimizing useful circuits that produce correct results on fundamentally nondeterministic hardware. The contribution of this paper is a novel machine learning approach, a combination of deep neural networks and random forests, for efficiently solving optimization problems that minimize sources of error in the Ising model. In addition, we provide a process to express a Boltzmann probability optimization problem as a supervised machine learning problem.

arXiv:2401.17408•npj Unconventional Computing•Aug 5, 2024•3 citations

A Fully-Connected Ising Model Embedding Method and Its Evaluation for CMOS Annealing Machines

Daisuke Oku, Kotaro Terada, Masato HAYASHI +3 more

Combinatorial optimization problems with a large solution space are difficult to solve just using von Neumann computers. Ising machines or annealing machines have been developed to tackle these problems as a promising Non-von Neumann computer. In order to use these annealing machines, every combinatorial optimization problem is mapped onto the physical Ising model, which consists of spins, interactions between them, and their external magnetic fields. Then the annealing machines operate so as to search the ground state of the physical Ising model, which corresponds to the optimal solution of the original combinatorial optimization problem. A combinatorial optimization problem can be firstly described by an ideal fully-connected Ising model but it is very hard to embed it onto the physical Ising model topology of a particular annealing machine, which causes one of the largest issues in annealing machines. In this paper, we propose a fully-connected Ising model embedding method targeting for CMOS annealing machine. The key idea is that the proposed method replicates every logical spin in a fully-connected Ising model and embeds each logical spin onto the physical spins with the same chain length. Experimental results through an actual combinatorial problem show that the proposed method obtains spin embeddings superior to the conventional de facto standard method, in terms of the embedding time and the probability of obtaining a feasible solution.

IEICE Transactions on Information and Systems•Aug 31, 2019•33 citations

A Continuous-Time Ising Machine using Coupled Inverter Chains Featuring Fully-Parallel One-Shot Spin Updates

Chengshuo Yu, Junjie Mu, Kevin Tshun Chuan Chai +2 more

Unconventional computers based on superconducting qubits [1] and analog/digital hardware accelerators [2]–[5] have been developed and solved hard combinatorial optimization problems (COPs) with higher efficiency than classical computers by integrating the Ising models in hardware and observing their natural convergence behavior towards the lower Ising Hamiltonian (or energy). Ouantum annealer based on superconducting qubits [1] has demonstrated its capability of finding solutions to COPs using the quantum tunneling effect. However, the quantum computer requires operating at ultra-low temperatures (15-20mK), consumes excessive power (>10kW), and is not scalable yet to solve real-world problems. Recently, ASIC hardware accelerators [2]–[5] using low-cost CMOS technologies have implemented the Ising model by integrating many physical spins (i.e., computing units of the Ising model) to accelerate the finding of solutions to hard COPs. The hardware accelerators, so-called Ising machines, solved hard COPs much faster than classical computers and consumed several orders of magnitude lower energy. Nevertheless, they still have to overcome the fundamental scalability issues for solving large-scale COPs in the real world. The performance of existing discrete-time Ising machines has been limited by the exponentially increasing number of operation cycles when the problem size increases for the large-scale COPs. in addition, prior works have not been able to integrate massive random number generators (RNG), the essential building blocks of the Ising machines for finding better solutions, that could occupy a significant hardware footprint. Prior works have implemented them off-chip, or a few have been integrated on-chip and shared by many spins. This paper proposes a novel continuous-time analog Ising machine based on coupled inveiter chain circuits to address the scalability issues of the existing discrete-time Ising machines. Besides the benefits of the continuous-time operation, which significantly lowers the computing energy and latency, we introduce an equalization method for finding better solutions without RNGs.

Custom Integrated Circuits Conference•Apr 1, 2023•10 citations
Ising machines have received growing interest as efficient and hardware-friendly solvers for combinatorial optimization problems (COPs). They search for the absolute or approximate ground states of the Ising model with a proper annealing process. In contrast to Ising machines built with superconductive or optical circuits, complementary metal-oxide-semiconductor (CMOS) Ising machines offer inexpensive fabrication, high scalability, and easy integration with mainstream semiconductor chips. As low-energy and CMOS-compatible emerging technologies, spintronics and phase-transition devices offer functionalities that can enhance the scalability and sampling performance of Ising machines. In this article, we survey various approaches in the process flow for solving COPs using CMOS, hybrid CMOS-spintronic, and phase-transition devices. First, the methods for formulating COPs as Ising problems and embedding Ising formulations to the topology of the Ising machine are reviewed. Then, Ising machines are classified by their underlying operational principles and reviewed from a perspective of hardware implementation. CMOS solutions are advantageous with denser connectivity, whereas hybrid CMOS-spintronic and phase-transition device-based solutions show great potential in energy efficiency and high performance. Finally, the challenges and prospects are discussed for the Ising formulation, embedding process, and implementation of Ising machines.
Combinatorial optimization problems are difficult to solve due to the space explosion in an exhaustive search. Using Ising model-based solvers can efficiently find near-optimal solutions by minimizing the energy of a nonlinear Hamiltonian system. In contrast to Ising machines based on quantum mechanics, classical Ising machines using conventional technologies, such as the complementary metal-oxide-semiconductor, offer efficient implementations with competitive performance. In this paper, we briefly review recently developed simulation algorithms of classical Ising machines. These algorithms are classified by considering various inherent mechanisms in the simulation of physical phenomena. Then, strategies to improve the simulation efficiency are discussed by generalizing their characteristics and behaviours. These simulation algorithms are key for improving the efficiency of classical Ising machines in solving combinatorial optimization problems.
Although quantum computing is expected to supplant traditional computing in several application fields, its adoption is hampered by temperature and economic constraints. To overcome these hurdles, researchers have proposed complementary metal-oxide-semiconductor (CMOS) annealing circuits. These circuits draw inspiration from quantum computing algorithms such as quantum annealing and aim to achieve near-quantum benefits by leveraging traditional CMOS technologies. This paper introduces an Ising-model-based hardware architecture that can be applied to combinatorial optimization problems (COPs). With its ability to express quadratic unconstrained binary optimization (QUBO) formulations as polynomials, the Ising model facilitates the encapsulation of multiple solutions and mapping onto fully connected architecture. The proposed annealing accelerator utilizes traditional circuit technologies, including pseudo-random number generators (PRNGs), to realize the required algorithms. The chip proposed herein, implemented using Taiwan Semiconductor Manufacturing Company (TSMC) 90-nm CMOS technology, operates at 50 MHz and covers an area of $3.24{\mathrm{m}}{{\mathrm{m}}^2}$3.24mm2. Experimental results demonstrate the excellent performance of this annealing accelerator in terms of area and power consumption, indicating its promise for use in solving COPs rapidly.
ABSTRACT Combinatorial optimization problems (COPs) are fundamental to numerous applications but remain challenging due to their exponentially large solution spaces and the limitations of conventional computing architectures. Ising machines have emerged as a promising alternative by mapping optimization problems onto the energy minimization of the Ising Hamiltonian. Various hardware implementations, including quantum annealers, digital annealers, and coherent Ising machines, have been explored based on distinct physical mechanisms. More recently, emerging device‐based approaches have gained significant attention for their potential to enable scalable and energy‐efficient optimization. In this review, we present a comprehensive overview of Ising machines from both theoretical and hardware perspectives, with particular emphasis on three emerging‐device‐based platforms: crossbar array‐based, stochastic p‐bit, and oscillator‐based Ising machines. We establish a comparative device‐to‐system perspective that traces how distinct physical and device properties translate into computational primitives, hardware architectures, and ultimately system‐level scalability. Platform‐specific bottlenecks, including array nonidealities and peripheral overhead in crossbar‐based systems, device variability and probability‐generation overhead in p‐bit systems, and synchronization and coupling‐network complexity in oscillator‐based systems, are quantitatively and qualitatively analyzed. Finally, we discuss future directions toward scalable and energy‐efficient Ising hardware through device‐to‐system co‐optimization across devices, circuits, architectures, and algorithms.
Abstract Decades of exponential scaling in high-performance computing (HPC) efficiency is coming to an end. Transistor-based logic in complementary metal-oxide semiconductor (CMOS) technology is approaching physical limits beyond which further miniaturization will be impossible. Future HPC efficiency gains will necessarily rely on new technologies and paradigms of computing. The Ising model shows particular promise as a future framework for highly energy-efficient computation. Ising systems are able to operate at energies approaching thermodynamic limits for energy consumption of computation. Ising systems can function as both logic and memory. Thus, they have the potential to significantly reduce energy costs inherent to CMOS computing by eliminating costly data movement. The challenge in creating Ising-based hardware is in optimizing useful circuits that produce correct results on fundamentally nondeterministic hardware. The contribution of this paper is a novel machine learning approach, a combination of deep neural networks and random forests, for efficiently solving optimization problems that minimize sources of error in the Ising model. In addition, we provide a process to express a Boltzmann probability optimization problem as a supervised machine learning problem.
Combinatorial optimization problems with a large solution space are difficult to solve just using von Neumann computers. Ising machines or annealing machines have been developed to tackle these problems as a promising Non-von Neumann computer. In order to use these annealing machines, every combinatorial optimization problem is mapped onto the physical Ising model, which consists of spins, interactions between them, and their external magnetic fields. Then the annealing machines operate so as to search the ground state of the physical Ising model, which corresponds to the optimal solution of the original combinatorial optimization problem. A combinatorial optimization problem can be firstly described by an ideal fully-connected Ising model but it is very hard to embed it onto the physical Ising model topology of a particular annealing machine, which causes one of the largest issues in annealing machines. In this paper, we propose a fully-connected Ising model embedding method targeting for CMOS annealing machine. The key idea is that the proposed method replicates every logical spin in a fully-connected Ising model and embeds each logical spin onto the physical spins with the same chain length. Experimental results through an actual combinatorial problem show that the proposed method obtains spin embeddings superior to the conventional de facto standard method, in terms of the embedding time and the probability of obtaining a feasible solution.
Unconventional computers based on superconducting qubits [1] and analog/digital hardware accelerators [2]–[5] have been developed and solved hard combinatorial optimization problems (COPs) with higher efficiency than classical computers by integrating the Ising models in hardware and observing their natural convergence behavior towards the lower Ising Hamiltonian (or energy). Ouantum annealer based on superconducting qubits [1] has demonstrated its capability of finding solutions to COPs using the quantum tunneling effect. However, the quantum computer requires operating at ultra-low temperatures (15-20mK), consumes excessive power (>10kW), and is not scalable yet to solve real-world problems. Recently, ASIC hardware accelerators [2]–[5] using low-cost CMOS technologies have implemented the Ising model by integrating many physical spins (i.e., computing units of the Ising model) to accelerate the finding of solutions to hard COPs. The hardware accelerators, so-called Ising machines, solved hard COPs much faster than classical computers and consumed several orders of magnitude lower energy. Nevertheless, they still have to overcome the fundamental scalability issues for solving large-scale COPs in the real world. The performance of existing discrete-time Ising machines has been limited by the exponentially increasing number of operation cycles when the problem size increases for the large-scale COPs. in addition, prior works have not been able to integrate massive random number generators (RNG), the essential building blocks of the Ising machines for finding better solutions, that could occupy a significant hardware footprint. Prior works have implemented them off-chip, or a few have been integrated on-chip and shared by many spins. This paper proposes a novel continuous-time analog Ising machine based on coupled inveiter chain circuits to address the scalability issues of the existing discrete-time Ising machines. Besides the benefits of the continuous-time operation, which significantly lowers the computing energy and latency, we introduce an equalization method for finding better solutions without RNGs.