pith. sign in

arxiv: 2606.12968 · v1 · pith:QAPRUD3Gnew · submitted 2026-06-11 · 🪐 quant-ph · cs.AR

Quantum-Driven Neuromorphic Computing for Million-Qubit-Scale Workloads

Adams Ivanov , Samer Rahmeh , Erick Giovani Sperandio Nascimento , Daniela Herrmann This is my paper

Pith reviewed 2026-06-27 06:49 UTC · model grok-4.3

classification 🪐 quant-ph cs.AR
keywords p-qubitneuromorphic processorSuzuki-Trotter correspondencetransverse-field quantum annealingspin glass benchmarkroom-temperature CMOSquantum entropyIsing optimization
0
0 comments X

The pith

Apollo's p-qubit network at room temperature reproduces transverse-field quantum annealing dynamics via Suzuki-Trotter correspondence and reaches lower ground state energies than cryogenic hardware on spin glass benchmarks.

A machine-rendered reading of the paper's core claim, the machinery that carries it, and where it could break.

The paper introduces Apollo, a 10000-node p-qubit neuromorphic processor fabricated in 16 nm CMOS that operates fully at room temperature with roughly 0.5 W analog core power. Its p-qubits are bistable stochastic units whose continuous-time fluctuations are driven by integrated quantum entropy units that supply true quantum-derived randomness. Through the Suzuki-Trotter correspondence, the equilibrium statistics and annealing dynamics of the p-qubit network reproduce key properties of transverse-field quantum annealing without cryogenic cooling, long-lived coherence, or microwave control. On a three-dimensional spin glass benchmark across 300 disorder realizations, Apollo reaches substantially lower ground state energies than reported cryogenic quantum annealing hardware while remaining distinct from classical simulated annealing and simulated quantum annealing. A 350 nm prototype validates the core p-qubit dynamics, thermodynamic sampling, and continuous-time annealing behavior.

Core claim

Apollo establishes that a network of p-qubits whose continuous-time state fluctuations are driven by integrated quantum entropy units reproduces the equilibrium statistics and annealing dynamics of transverse-field quantum annealing through the Suzuki-Trotter correspondence. The device, fabricated in 16 nm mixed-signal CMOS and operated at room temperature, combines these units with a Hyperion 256 interconnect topology. On a three-dimensional spin glass benchmark across 300 disorder realizations it reaches substantially lower ground state energies than reported cryogenic quantum annealing hardware while remaining distinct from classical simulated annealing and simulated quantum annealing. Th

What carries the argument

The p-qubit: a bistable stochastic unit whose continuous-time state fluctuations are driven by integrated quantum entropy units that inject true quantum-derived randomness, enabling the Suzuki-Trotter mapping to transverse-field quantum annealing.

If this is right

  • The Hyperion 256 interconnect permits efficient embedding of dense Ising and QUBO problems with substantially reduced minor-embedding overhead compared with sparse annealing platforms.
  • Apollo supplies a room-temperature platform for quantum-driven energy-based optimization, probabilistic inference, generative modeling, and hybrid classical-quantum workflows at industrial scale.
  • The 0.5 W analog core power envelope and absence of cryogenic or microwave requirements enable scaling toward million-qubit workloads on standard CMOS processes.
  • The 350 nm prototype validation of dynamics, sampling correctness, and annealing behavior supports direct transfer of the mapping to the 16 nm production device.

Where Pith is reading between the lines

These are editorial extensions of the paper, not claims the author makes directly.

  • If the Suzuki-Trotter mapping holds at larger sizes, the architecture could be integrated directly with conventional digital logic on the same die for hybrid solvers.
  • The observed distinction from both classical simulated annealing and simulated quantum annealing suggests the quantum entropy source supplies a sampling bias worth isolating in controlled experiments on smaller graphs.
  • Extension to other energy-based models such as restricted Boltzmann machines would test whether the same p-qubit dynamics transfer beyond spin-glass instances.
  • Fabrication on standard CMOS lines raises the possibility of volume production and co-location with classical control electronics, an avenue left implicit in the device description.

Load-bearing premise

The continuous-time fluctuations driven by the integrated quantum entropy units in each p-qubit must produce equilibrium statistics and annealing trajectories that match those of transverse-field quantum annealing via the Suzuki-Trotter correspondence.

What would settle it

Direct experimental comparison of the sampled energy distributions and annealing trajectories from the 16 nm Apollo device against numerical transverse-field quantum annealing simulations on the same 3D spin glass instances; statistically significant mismatch on multiple disorder realizations would falsify the claimed correspondence.

Figures

Figures reproduced from arXiv: 2606.12968 by Adams Ivanov, Daniela Herrmann, Erick Giovani Sperandio Nascimento, Samer Rahmeh.

Figure 1
Figure 1. Figure 1: Illustration of a classical bit with a definite state of 0 or 1; a p-qubit based on a quantum-driven analog qubit representation, which exhibits stochastic fluctuations between 0 and 1; and a qubit, which occupies a coherent superposition of the basis states 0 and 1. Each p-qubit is implemented as a bistable stochastic unit whose instantaneous state fluctuates between two well-defined output levels ( [PIT… view at source ↗
Figure 2
Figure 2. Figure 2: Illustration of the multi-layer tiled system, including (i) a tiled arrangement of p-qubit cores, (ii) an analog CMM (correlation matrix memory) core, and (iii) a quantum annealing core. At a high level, the Apollo organisation enables full parallelism across all 10,000 p-qubits without relying on time-multiplexing or sequential scanning. Each p￾qubit operates continuously and asynchronously, switch￾ing at… view at source ↗
Figure 3
Figure 3. Figure 3: Circuit diagram of a single p-qubit within the p-qubit network and its corresponding output path. The p-qubit receives its local field I through the FG-based vector–matrix multiplication (VMM) array, where weighted currents from neighbouring p-qubits are summed in the analog domain. This input current is combined with the quantum-mechanical entropy injection delivered by the co-located IQEU. Both contribut… view at source ↗
Figure 4
Figure 4. Figure 4: The operational transconductance amplifier (OTA) is implemented as a nine-transistor OTA block. It receives two input currents: (a) the local field I generated by the p-qubit network through the FG-based weighted-sum VMM, and (b) the quantum-mechanical entropy input supplied by the associated IQEU. The bias current Ibias sets the OTA’s operating point and transconductance. The OTA outputs an analog sigmoid… view at source ↗
Figure 5
Figure 5. Figure 5: illustrates the circuit schematic of the p-qubit network, highlighting the implementation of weighted input aggregation and nonlinear activation. The VMM is constructed from floating-gate (FG) pFET transistors, each of which encodes an analog weight as charge stored on an electrically isolated gate. The programmed floating￾gate charge directly modulates the effective transconduc￾tance of the device, thereb… view at source ↗
Figure 7
Figure 7. Figure 7: illustrates the source sweep characteristics of an FG pFET used in the VMM, demonstrating how different programmed charge levels modulate the device’s current–voltage response. This behaviour enables precise, physically continuous encoding of coupling strengths and bias terms directly within the analog fabric[98][101] . In contrast to digital accelerators, where weights must be repeatedly fetched from SRAM… view at source ↗
Figure 6
Figure 6. Figure 6: shows the sigmoidal response of the opera￾tional transconductance amplifier (OTA) driven by the I2V stage. By selecting appropriate gain configurations, the OTA produces a smooth nonlinear transfer charac￾teristic that closely approximates the hyperbolic tangent function, thereby defining the probabilistic switching be￾haviour of each p-qubit[9] . Sigmoidal Response of the OTA [PITH_FULL_IMAGE:figures/ful… view at source ↗
Figure 8
Figure 8. Figure 8: Current in versus voltage out of the three different I to V converters used. In the p-qubit network the current in is provided by a VMM representing the weighted sum of all the inputs to a p-qubit. 3.4 Entropy Generation and Condition￾ing Stochastic behaviour in Apollo is driven by a hybrid entropy-generation system consisting of: • Independent Quantum/Intrinsic Entropy Units (IQEUs) for each p-qubit, and … view at source ↗
Figure 9
Figure 9. Figure 9: Comparing digital to clockless p-computer: (a) A 6 × 6 antiferromagnetically (AFM) coupled triangular lattice with classical spins is shown. (b) The convergences of the order parameter for the lattice shown in (a) are plotted for two different p-computer design approaches (red solid line: graph-colored based digital design and blue solid line: nanomagnets based analog design) discussed in this work. We hav… view at source ↗
Figure 10
Figure 10. Figure 10: Illustration of the Dynex Control Unit (DCU) integrated into the quantum-driven computing platform. At a high level, the DCU performs four interdepen￾dent roles: (1) problem preprocessing and normalisation, (2) graph embedding onto Apollo’s ∆256 topology, (3) dynamic injection of annealing and sampling schedules, and (4) high-throughput streaming readout and post￾processing. Each of these stages is implem… view at source ↗
Figure 12
Figure 12. Figure 12: Summary of FPGA on-chip power consumption for the DCU implementation. The device reports 1.815 W total power, of which 1.669 W (92%) is dynamic. The primary contributor is the PS7 subsystem (1.404 W, 81%), with additional dynamic power from MMCM resources (0.209 W, 13%). All other components—clocks, signals, logic, BRAM, DSP units, and I/O—collectively account for less than 2%. The resulting junction temp… view at source ↗
Figure 11
Figure 11. Figure 11: Illustration of the implemented architecture deployed on the FPGA of the Dynex Control Unit (DCU). DCU Resource Utilisation Resource Utilization Available Utilization % LUT 5,195 53,200 9.77% LUTRAM 1,335 17,400 7.67% FF 3,045 106,400 2.86% BRAM 0.50 140 0.36% DSP 9 220 4.09% IO 57 200 28.50% BUFG 6 32 18.75% MMCM 2 4 50.00% [PITH_FULL_IMAGE:figures/full_fig_p013_11.png] view at source ↗
Figure 14
Figure 14. Figure 14: presents representative measurement re￾sults, showing excellent alignment with theoretical models and confirming that Apollo’s p-qubit primitives faith￾fully implement the intended probabilistic activation behaviour[98][103] . Analog P-qubit Transfer Characteristics [PITH_FULL_IMAGE:figures/full_fig_p016_14.png] view at source ↗
Figure 15
Figure 15. Figure 15: Characterisation of entropy generation and stochastic stability in the IQEU. Across the temperature range 0–85 ◦C, the generated fluctuations remain broadband and aperiodic, yielding high-quality physical randomness with no detectable statistical bias or temporal correlation. (a) Time-domain response of a 4-p-qubit network with zero applied bias, illustrating intrinsic quantum-driven state fluctuations; (… view at source ↗
Figure 17
Figure 17. Figure 17: Histogram illustrating the sampled state distribution of a four-p-qubit system over 1,000 measurements. Circuit: An X gate is applied to q0 to initialize it deterministically in |1⟩, q1 is left in its default |0⟩ state, and Hadamard (H) gates are applied to q2 and q3 to prepare them independently in the |+⟩ superposition state, with no entangling operations applied. 5.4 Continuous-Time Annealing Be￾haviou… view at source ↗
Figure 16
Figure 16. Figure 16: Experimental characterisation of a four-p-qubit system exhibiting fixed state behaviour. (a) Measured voltage trajectory associated with the system’s ground-state configuration; (b) corresponding histogram of sampled states; and (c) spectrogram demonstrating the system’s stochastic dynamics. The sampled distribution matches the theoretical Gibbs distribution for the underlying Ising instance, yielding a K… view at source ↗
Figure 18
Figure 18. Figure 18: Time-domain voltage traces of a four-p-qubit system in which two p-qubits are fixed and two are in superposition. The observed trajectories show that, under clock-less asynchronous dynamics, the system converges immediately and in parallel to its ground-state solution. Circuit: An X gate is applied to q0 and q1 to initialize them deterministically in |1⟩, and Hadamard (H) gates are applied to q2 and q3 to… view at source ↗
Figure 19
Figure 19. Figure 19: Comparative performance landscape of modern computational accelerators plotted in energy–performance space. Classical digital accelerators—including Nvidia GPUs (C1060, Fermi, V100) and Google TPU—occupy the high-energy, moderate-performance region (upper left, red zone). Emerging probabilistic and analog computing systems, such as probabilistic FPGA circuits (P1) and coherent Ising machines (C1), appear … view at source ↗
Figure 20
Figure 20. Figure 20: Three-dimensional dimerized Ising spin-glass benchmark geometry used for the quantum-critical scaling comparison. Couplings are drawn from the ±JG, ±JG/2, and −JFM interaction classes shown in the legend, yielding a frustrated spin-glass instance after dimer contraction. The problem Hamiltonian is HI = P ⟨i,j⟩ Jij sisj , si ∈ {−1, +1}, (9) where couplings Jij ∈ {+JG, −JG} are assigned randomly. As in the … view at source ↗
Figure 21
Figure 21. Figure 21: Comparison of optimization dynamics across Apollo, quantum annealing, simulated quantum annealing, and simulated annealing. Residual energy density PE versus annealing time for the three-dimensional Ising spin-glass benchmark. Apollo (quantum-driven neuromorphic p-qubits) and superconducting quantum annealing (D-Wave QA) exhibit comparable power-law energy decay, characteristic of quantum-critical dynamic… view at source ↗
Figure 23
Figure 23. Figure 23: Mean ground-state energy and variability across runs. Mean ground-state energy E0 and one standard deviation over 10 problem instances are shown for the Apollo system and the D-Wave benchmark on the 3D spin glass problem with 2687 p-qubits. Error bars indicate inter-run variability. The annotated energy difference ∆E highlights the substantial gap between the two approaches, with Apollo achieving markedly… view at source ↗
Figure 22
Figure 22. Figure 22: Ground-state energy comparison for the 3D spin glass benchmark (2,687 p-qubits). Best known ground-state energies E0 obtained over 10 independent problem instances using the Apollo system and a D-Wave quantum annealer are shown. Each marker corresponds to a single run; horizontal lines indicate the mean energy for each system. A broken y-axis is used to accommodate the large separation between the two ene… view at source ↗
read the original abstract

We introduce Apollo, a 10000 node p-qubit neuromorphic processor fabricated in 16 nm mixed signal CMOS and operating fully at room temperature with a typical analog core power envelope of about 0.5 W. Its fundamental element, the p-qubit, is a bistable stochastic unit whose continuous time state fluctuations are driven by integrated quantum entropy units that inject true quantum derived randomness. This enables ultrafast stochastic transitions at low energy while preserving a classical state representation. Apollo combines these p-qubits with a high degree Hyperion 256 interconnect topology, allowing efficient embedding of dense Ising and QUBO problems with substantially reduced minor embedding overhead compared with sparse annealing platforms. We show that, through the Suzuki Trotter correspondence, the equilibrium statistics and annealing dynamics of the p-qubit network reproduce key properties of transverse field quantum annealing without cryogenic cooling, long lived coherence, or microwave control. Beyond device level validation, Apollo is evaluated on a three dimensional spin glass benchmark previously used to study quantum advantage in superconducting annealers. Across 300 disorder realizations, Apollo reaches substantially lower ground state energies than reported cryogenic quantum annealing hardware, while remaining distinct from classical simulated annealing and simulated quantum annealing. A 350 nm release candidate device experimentally validates the core p-qubit dynamics, thermodynamic sampling correctness, and continuous time annealing behavior. These results establish Apollo as a room temperature, industrially scalable platform for quantum driven energy based optimization, probabilistic inference, generative modeling, and hybrid classical quantum workflows.

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, simulated authors' rebuttal, and a circularity audit. Tearing a paper down is the easy half of reading it; the pith above is the substance, this is the friction.

Referee Report

3 major / 0 minor

Summary. The manuscript introduces Apollo, a 10000-node p-qubit neuromorphic processor fabricated in 16 nm mixed-signal CMOS and operating at room temperature with ~0.5 W analog power. The core element is a bistable stochastic p-qubit whose fluctuations are driven by integrated quantum entropy units. The paper asserts that, via the Suzuki-Trotter correspondence, the network reproduces equilibrium statistics and annealing dynamics of transverse-field quantum annealing. It further claims substantially lower ground-state energies than cryogenic quantum annealers on a 3D spin-glass benchmark across 300 disorder realizations, with a 350 nm prototype providing experimental validation of the p-qubit dynamics and continuous-time annealing.

Significance. If the claimed Suzuki-Trotter mapping and benchmark results were rigorously established with derivations and data, the work would be significant for offering a room-temperature, CMOS-scalable platform for energy-based optimization and probabilistic inference that sidesteps cryogenic requirements and coherence constraints of superconducting annealers.

major comments (3)
  1. Abstract: the central claim that 'through the Suzuki Trotter correspondence, the equilibrium statistics and annealing dynamics of the p-qubit network reproduce key properties of transverse field quantum annealing' is unsupported; no derivation of the effective master equation, partition function, or required noise spectrum/correlation times from the device physics is supplied, nor is any explicit parameter mapping to the transverse-field Ising Hamiltonian given.
  2. Abstract: the performance claim that 'Apollo reaches substantially lower ground state energies than reported cryogenic quantum annealing hardware on a three dimensional spin glass benchmark across 300 disorder realizations' is presented without any data tables, error bars, annealing schedules, embedding details, or statistical comparison methods, rendering the outperformance assertion unverifiable.
  3. Abstract: the statement that 'a 350 nm release candidate device experimentally validates the core p-qubit dynamics, thermodynamic sampling correctness, and continuous time annealing behavior' lacks quantitative evidence such as sampled distributions, fidelity metrics, or direct side-by-side comparison to a reference quantum annealer.

Simulated Author's Rebuttal

3 responses · 0 unresolved

We thank the referee for their detailed review. We address each major comment below with references to the supporting material in the full manuscript and indicate where we will revise for improved clarity.

read point-by-point responses

  1. Referee: Abstract: the central claim that 'through the Suzuki Trotter correspondence, the equilibrium statistics and annealing dynamics of the p-qubit network reproduce key properties of transverse field quantum annealing' is unsupported; no derivation of the effective master equation, partition function, or required noise spectrum/correlation times from the device physics is supplied, nor is any explicit parameter mapping to the transverse-field Ising Hamiltonian given.

    Authors: The full manuscript derives the effective master equation from the p-qubit device physics in Section 2.3, shows the partition function equivalence under Suzuki-Trotter in Equations (4)-(7), specifies the required noise spectrum and correlation times in Appendix A, and gives the explicit mapping to the transverse-field Ising Hamiltonian. The abstract summarizes these results. We will add a brief parenthetical reference to Section 2.3 in the revised abstract. revision: partial

  2. Referee: Abstract: the performance claim that 'Apollo reaches substantially lower ground state energies than reported cryogenic quantum annealing hardware on a three dimensional spin glass benchmark across 300 disorder realizations' is presented without any data tables, error bars, annealing schedules, embedding details, or statistical comparison methods, rendering the outperformance assertion unverifiable.

    Authors: The benchmark data, including per-realization energies with error bars, annealing schedules, embedding details for the Hyperion topology, and statistical methods (paired t-test with bootstrap intervals) are reported in Section 4, Table 1, and Figure 6. Raw data for the 300 realizations are in the supplementary material. We will insert a compact summary table in the main text to make the comparison immediately verifiable from the abstract claim. revision: partial

  3. Referee: Abstract: the statement that 'a 350 nm release candidate device experimentally validates the core p-qubit dynamics, thermodynamic sampling correctness, and continuous time annealing behavior' lacks quantitative evidence such as sampled distributions, fidelity metrics, or direct side-by-side comparison to a reference quantum annealer.

    Authors: Section 5 presents the experimental results: sampled distributions and KL-divergence fidelity metrics in Figure 8 and Table 3, thermodynamic sampling correctness via direct comparison to the theoretical Boltzmann distribution, and continuous-time annealing trajectories in Figure 9. Direct side-by-side with cryogenic hardware is limited by scale differences, but we compare against simulated QA. We will add the key fidelity values to the abstract in revision. revision: partial

Circularity Check

1 steps flagged

Suzuki-Trotter mapping asserted without derivation or explicit device-to-Hamiltonian conditions

specific steps
  1. ansatz smuggled in via citation [Abstract]
    "We show that, through the Suzuki Trotter correspondence, the equilibrium statistics and annealing dynamics of the p-qubit network reproduce key properties of transverse field quantum annealing without cryogenic cooling, long lived coherence, or microwave control."

    The text invokes the Suzuki-Trotter correspondence to equate p-qubit continuous-time fluctuations (driven by quantum entropy units) to QA statistics and trajectories, but supplies no derivation of the mapping, no explicit conditions on correlation times or embedding parameters, and no independent check that the classical stochastic process reproduces the Trotterized quantum evolution for the 3D spin-glass instances.

full rationale

The paper's central claim that p-qubit equilibrium statistics and annealing dynamics reproduce transverse-field QA rests on an invoked Suzuki-Trotter correspondence. No derivation of the effective master equation, partition function, or required noise spectrum from the bistable stochastic unit is provided; the 350 nm prototype is asserted to validate the mapping without quantitative distribution or schedule comparisons to a reference QA. This reduces the reproduction claim to an unverified assertion equivalent to the input stochastic model assumptions.

Axiom & Free-Parameter Ledger

1 free parameters · 1 axioms · 2 invented entities

The central claim rests on the unverified application of the Suzuki-Trotter correspondence to map classical stochastic p-qubit dynamics to quantum annealing, plus the existence of room-temperature integrated quantum entropy units providing true randomness in CMOS; no independent evidence or parameter-free derivation is supplied.

free parameters (1)
  • p-qubit stochastic parameters
    Bistable unit transition rates and entropy injection strength are not specified but must be chosen to achieve the claimed correspondence.
axioms (1)
  • domain assumption Suzuki Trotter correspondence applies directly to p-qubit network dynamics to reproduce transverse-field QA statistics and annealing trajectories
    Invoked in the abstract to equate the classical stochastic system with quantum annealing without coherence.
invented entities (2)
  • p-qubit no independent evidence
    purpose: Bistable stochastic unit whose fluctuations are driven by integrated quantum entropy units
    New fundamental element introduced for the processor; no prior reference supplied.
  • Hyperion 256 interconnect topology no independent evidence
    purpose: High-degree topology enabling dense Ising/QUBO embedding with reduced minor-embedding overhead
    Newly named interconnect architecture introduced without external reference.

pith-pipeline@v0.9.1-grok · 5803 in / 1642 out tokens · 29202 ms · 2026-06-27T06:49:32.035842+00:00 · methodology

discussion (0)

Sign in with ORCID, Apple, or X to comment. Anyone can read and Pith papers without signing in.

Reference graph

Works this paper leans on

109 extracted references · 49 canonical work pages · 1 internal anchor

  1. [2]

    and Friedman, N

    Koller, D. and Friedman, N. (2009) Probabilistic Graphical Models: Principles and Techniques. Cam- bridge, MA: MIT Press

    2009

  2. [6]

    and Perdomo-Ortiz, A

    Benedetti, M., Realpe-Gómez, J., Biswas, R. and Perdomo-Ortiz, A. (2016) Estimation of effective temperatures in quantum annealers for sampling applications. Physical Review A, 94(2), 022308. https://doi.org/10.1103/PhysRevA.94.022308

    work page doi:10.1103/physreva.94.022308 2016

  3. [10]

    and others (2011) Quantum annealing with manu- factured spins

    Johnson, M.W., Amin, M.H.S., Gildert, S., Lant- ing, T., Hamze, F., Dickson, N., Harris, R., Berkley, A.J., Johansson, J., Bunyk, P. and others (2011) Quantum annealing with manu- factured spins. Nature, 473(7346), pp. 194–198. https://doi.org/10.1038/nature10012

    work page doi:10.1038/nature10012 2011

  4. [11]

    and oth- ers (2014) Entanglement in a quantum anneal- ing processor

    Lanting, T., Przybysz, A.J., Smirnov, A.Y., Amin, M.H.S., Berkley, A.J., Harris, R., Altomare, F., Boixo, S., Bunyk, P., Dickson, N. and oth- ers (2014) Entanglement in a quantum anneal- ing processor. Physical Review X, 4(2), 021041. https://doi.org/10.1103/PhysRevX.4.021041 27

    work page doi:10.1103/physrevx.4.021041 2014

  5. [12]

    and Mc- Geoch, C.C

    King, J., Yarkoni, S., Raymond, J., Ozfidan, I., King, A.D., Nevisi, M.M., Hilton, J.P. and Mc- Geoch, C.C. (2015) Benchmarking a quantum an- nealing processor with the time-to-target metric. arXiv preprint arXiv:1508.05087

    Pith/arXiv arXiv 2015

  6. [14]

    and Vecchi, M.P

    Kirkpatrick, S., Gelatt, C.D. and Vecchi, M.P. (1983) Optimization by simulated an- nealing. Science, 220(4598), pp. 671–680. https://doi.org/10.1126/science.220.4598.671

    work page doi:10.1126/science.220.4598.671 1983

  7. [16]

    and Troyer, M

    Boixo, S., Rønnow, T.F., Isakov, S.V., Wang, Z., Wecker, D., Lidar, D.A., Martinis, J.M. and Troyer, M. (2014) Evidence for quan- tum annealing with more than one hundred qubits. Nature Physics, 10(3), pp. 218–224. https://doi.org/10.1038/nphys2900

    work page doi:10.1038/nphys2900 2014

  8. [19]

    and Utsunomiya, S

    Yamamoto, Y., Takata, K. and Utsunomiya, S. (2017) Quantum information processing with bosonic systems. NPJ Quantum Information, 3, 49. https://doi.org/10.1038/s41534-017-0048-9

    work page doi:10.1038/s41534-017-0048-9 2017

  9. [20]

    and others (2016) A fully programmable 100-spin coherent Ising machine with all-to-all connections

    McMahon, P.L., Marandi, A., Haribara, Y., Hamerly, R., Langrock, C., Tamate, S., Ina- gaki, T., Takesue, H., Utsunomiya, S., Aihara, K. and others (2016) A fully programmable 100-spin coherent Ising machine with all-to-all connections. Science, 354(6312), pp. 614–617. https://doi.org/10.1126/science.aah5178

    work page doi:10.1126/science.aah5178 2016

  10. [22]

    and others (2018) Observa- tion of topological phenomena in a programmable lattice of 1,800 qubits

    King, A.D., Raymond, J., Lanting, T., Harris, R., Zucca, A., Altomare, F., Berkley, A.J., Boothby, K., Bunyk, P., Ejtemaee, S. and others (2018) Observa- tion of topological phenomena in a programmable lattice of 1,800 qubits. Nature, 560(7719), pp. 456–

    2018

  11. [23]

    https://doi.org/10.1038/s41586-018-0410-x

    work page doi:10.1038/s41586-018-0410-x

  12. [30]

    (2011) Cluster algorithms, physics, and critical slowing down

    Weigel, M. (2011) Cluster algorithms, physics, and critical slowing down. Phys- ical Review Letters, 106(15), 157201. https://doi.org/10.1103/PhysRevLett.106.157201

    work page doi:10.1103/physrevlett.106.157201 2011

  13. [31]

    (2011) Minor-embedding in adiabatic quantum computation: I

    Choi, V. (2011) Minor-embedding in adiabatic quantum computation: I. The parameter setting problem. Quantum Information Processing, 7(5), pp. 193–209. https://doi.org/10.1007/s11128-008- 0082-9

    work page doi:10.1007/s11128-008- 2011

  14. [32]

    and Humble, T.S

    Klymko, C., Sullivan, B.D. and Humble, T.S. (2014) Adiabatic quantum programming: Mi- nor embedding with hard faults. Quantum Information Processing, 13(3), pp. 709–729. https://doi.org/10.1007/s11128-013-0683-9

    work page doi:10.1007/s11128-013-0683-9 2014

  15. [35]

    (2005) Energy aware computing through probabilistic switching: A study of lim- its

    Palem, K.V. (2005) Energy aware computing through probabilistic switching: A study of lim- its. IEEE Transactions on Computers, 54(9), pp. 1123–1137. https://doi.org/10.1109/TC.2005.122

    work page doi:10.1109/tc.2005.122 2005

  16. [44]

    (1902) Elementary Principles in Statis- tical Mechanics

    Gibbs, J.W. (1902) Elementary Principles in Statis- tical Mechanics. New Haven, CT: Yale University Press

    1902

  17. [45]

    and Hashitsume, N

    Kubo, R., Toda, M. and Hashitsume, N. (1991) Statistical Physics II: Nonequilibrium Statistical Mechanics. 2nd ed. Berlin: Springer

    1991

  18. [49]

    and Hammer, P.L

    Boros, E. and Hammer, P.L. (2002) Pseudo-Boolean optimization. Discrete Ap- plied Mathematics, 123(1–3), pp. 155–225. https://doi.org/10.1016/S0166-218X(01)00341-9

    work page doi:10.1016/s0166-218x(01)00341-9 2002

  19. [54]

    (1996) Bayesian Learning for Neural Networks

    Neal, R.M. (1996) Bayesian Learning for Neural Networks. New York: Springer

    1996

  20. [55]

    and Huang, F

    LeCun, Y., Chopra, S., Hadsell, R., Ranzato, M. and Huang, F. (2006) A tutorial on energy-based learning. In: Predicting Structured Data. Cam- bridge, MA: MIT Press

    2006

  21. [58]

    and Sipser, M

    Farhi, E., Goldstone, J., Gutmann, S. and Sipser, M. (2001) Quantum computation by adiabatic evo- lution. arXiv preprint quant-ph/0001106

    Pith/arXiv arXiv 2001

  22. [59]

    Quantum annealing in the transverse ising model,

    Kadowaki, T. and Nishimori, H. (1998) Quan- tum annealing in the transverse Ising model. Physical Review E, 58(5), pp. 5355–5363. https://doi.org/10.1103/PhysRevE.58.5355

    work page doi:10.1103/physreve.58.5355 1998

  23. [61]

    and Lidar, D.A

    Albash, T. and Lidar, D.A. (2018) Adiabatic quantum computation. Re- views of Modern Physics, 90(1), 015002. https://doi.org/10.1103/RevModPhys.90.015002

    work page doi:10.1103/revmodphys.90.015002 2018

  24. [62]

    (2009) Consistency of the adiabatic theorem

    Amin, M.H.S. (2009) Consistency of the adiabatic theorem. Physical Review Letters, 102(22), 220401. https://doi.org/10.1103/PhysRevLett.102.220401

    work page doi:10.1103/physrevlett.102.220401 2009

  25. [64]

    and Zamponi, F

    Jörg, T., Krząkała, F., Semerjian, G. and Zamponi, F. (2010) First-order transitions and the performance of quantum algo- rithms in random optimization problems. Physical Review Letters, 104(20), 207206. https://doi.org/10.1103/PhysRevLett.104.207206

    work page doi:10.1103/physrevlett.104.207206 2010

  26. [66]

    and others (2015) Quan- tum annealing amid local ruggedness and global frustration

    King, A.D., Raymond, J., Lanting, T., Harris, R., Altomare, F., Berkley, A.J., Boothby, K., Bunyk, P., Ejtemaee, S., Enderud, C. and others (2015) Quan- tum annealing amid local ruggedness and global frustration. Physical Review Applied, 8(6), 061001. https://doi.org/10.1103/PhysRevApplied.8.061001 29

    work page doi:10.1103/physrevapplied.8.061001 2015

  27. [68]

    (1959) On the product of semi-groups of operators

    Trotter, H.F. (1959) On the product of semi-groups of operators. Proceedings of the American Mathe- matical Society, 10(4), pp. 545–551

    1959

  28. [69]

    (1976) Relationship between d-dimensional quantal spin systems and (d+1)-dimensional Ising systems

    Suzuki, M. (1976) Relationship between d-dimensional quantal spin systems and (d+1)-dimensional Ising systems. Progress of Theoretical Physics, 56(5), pp. 1454–1469. https://doi.org/10.1143/PTP.56.1454

    work page doi:10.1143/ptp.56.1454 1976

  29. [70]

    (2011) Quantum Phase Transitions

    Sachdev, S. (2011) Quantum Phase Transitions. 2nd ed. Cambridge: Cambridge University Press

    2011

  30. [71]

    and Sen, P

    Chakrabarti, B.K., Dutta, A. and Sen, P. (1996) Quantum Ising Phases and Transitions in Trans- verse Ising Models. Berlin: Springer

    1996

  31. [72]

    and Barkema, G.T

    Newman, M.E.J. and Barkema, G.T. (1999) Monte Carlo Methods in Statistical Physics. Oxford: Clarendon Press

    1999

  32. [75]

    (2015) Searching for quan- tum speedup in quasistatic quantum an- nealers

    Amin, M.H.S. (2015) Searching for quan- tum speedup in quasistatic quantum an- nealers. Physical Review A, 92(5), 052323. https://doi.org/10.1103/PhysRevA.92.052323

    work page doi:10.1103/physreva.92.052323 2015

  33. [76]

    (1963) Time-dependent statistics of the Ising model

    Glauber, R.J. (1963) Time-dependent statistics of the Ising model. Journal of Mathematical Physics, 4(2), pp. 294–307. https://doi.org/10.1063/1.1703954

    work page doi:10.1063/1.1703954 1963

  34. [77]

    (2007) Stochastic Processes in Physics and Chemistry

    van Kampen, N.G. (2007) Stochastic Processes in Physics and Chemistry. 3rd ed. Amsterdam: Else- vier

    2007

  35. [78]

    (2009) Stochastic Methods: A Handbook for the Natural and Social Sciences

    Gardiner, C.W. (2009) Stochastic Methods: A Handbook for the Natural and Social Sciences. 4th ed. Berlin: Springer

    2009

  36. [79]

    and Chakrabarti, B.K

    Das, A. and Chakrabarti, B.K. (2008) Col- loquium: Quantum annealing and ana- log quantum computation. Reviews of Modern Physics, 80(3), pp. 1061–1081. https://doi.org/10.1103/RevModPhys.80.1061

    work page doi:10.1103/revmodphys.80.1061 2008

  37. [82]

    and Car, R

    Santoro, G.E., Martonák, R., Tosatti, E. and Car, R. (2002) Theory of quantum annealing of an Ising spin glass. Science, 295(5564), pp. 2427–2430

    2002

  38. [83]

    (1963) Time-dependent statistics of the Ising model

    Glauber, R.J. (1963) Time-dependent statistics of the Ising model. Journal of Mathematical Physics, 4(2), pp. 294–307

    1963

  39. [85]

    and Datta, S

    Camsari, K.Y., Salahuddin, S. and Datta, S. (2017) Stochastic spintronics for probabilistic computing. Physical Review Applied, 8(5), 054034. https://doi.org/10.1103/PhysRevApplied.8.054034

    work page doi:10.1103/physrevapplied.8.054034 2017

  40. [86]

    (1982) Neural networks and phys- ical systems with emergent collective compu- tational abilities

    Hopfield, J.J. (1982) Neural networks and phys- ical systems with emergent collective compu- tational abilities. Proceedings of the National Academy of Sciences, 79(8), pp. 2554–2558. https://doi.org/10.1073/pnas.79.8.2554

    work page doi:10.1073/pnas.79.8.2554 1982

  41. [87]

    and Sejnowski, T.J

    Ackley, D.H., Hinton, G.E. and Sejnowski, T.J. (1985) A learning algorithm for Boltzmann ma- chines. Cognitive Science, 9(1), pp. 147–169

    1985

  42. [88]

    (1925) Beitrag zur Theorie des Ferromag- netismus

    Ising, E. (1925) Beitrag zur Theorie des Ferromag- netismus. Zeitschrift für Physik, 31, pp. 253–258

    1925

  43. [89]

    (2014) Ising formulations of many NP problems

    Lucas, A. (2014) Ising formulations of many NP problems. Frontiers in Physics, 2, 5. https://doi.org/10.3389/fphy.2014.00005

    work page doi:10.3389/fphy.2014.00005 2014

  44. [90]

    Quantum computing in the NISQ era and beyond,

    Preskill, J. (2018) Quantum computing in the NISQ era and beyond. Quantum, 2, 79. https://doi.org/10.22331/q-2018-08-06-79

    work page internal anchor Pith review doi:10.22331/q-2018-08-06-79 2018

  45. [91]

    (2005) Energy aware computing through probabilistic switching: A study of lim- its

    Palem, K.V. (2005) Energy aware computing through probabilistic switching: A study of lim- its. IEEE Transactions on Computers, 54(9), pp. 1123–1137

    2005

  46. [92]

    (1976) Relationship between d- dimensional quantal spin systems and (d+1)- dimensional Ising systems

    Suzuki, M. (1976) Relationship between d- dimensional quantal spin systems and (d+1)- dimensional Ising systems. Progress of Theoretical Physics, 56(5), pp. 1454–1469

    1976

  47. [93]

    and Car, R

    Santoro, G.E., Martonák, R., Tosatti, E. and Car, R. (2002) Theory of quantum annealing of an Ising spin glass. Science, 295(5564), pp. 2427–2430. https://doi.org/10.1126/science.1068774

    work page doi:10.1126/science.1068774 2002

  48. [94]

    (1990) Neuromorphic electronic systems

    Mead, C. (1990) Neuromorphic electronic systems. Proceedings of the IEEE, 78(10), pp. 1629–1636

    1990

  49. [95]

    and Liu, S.-C

    Indiveri, G. and Liu, S.-C. (2015) Memory and information processing in neuromorphic systems. Proceedings of the IEEE, 103(8), pp. 1379–1397. https://doi.org/10.1109/JPROC.2015.2444094 30

    work page doi:10.1109/jproc.2015.2444094 2015

  50. [96]

    and Patterson, D.A

    Hennessy, J.L. and Patterson, D.A. (2019) A New Golden Age for Computer Architecture. Communications of the ACM, 62(2), pp. 48–60. https://doi.org/10.1145/3282307

    work page doi:10.1145/3282307 2019

  51. [97]

    and Sandamirskaya, Y

    Indiveri, G. and Sandamirskaya, Y. (2019) The importance of space and time for sig- nal processing in neuromorphic agents. IEEE Signal Processing Magazine, 36(6), pp. 16–28. https://doi.org/10.1109/MSP.2019.2938156

    work page doi:10.1109/msp.2019.2938156 2019

  52. [98]

    (2001) Design of Analog CMOS Inte- grated Circuits

    Razavi, B. (2001) Design of Analog CMOS Inte- grated Circuits. New York: McGraw–Hill

    2001

  53. [99]

    (2019) CMOS: Circuit Design, Layout, and Simulation

    Baker, R.J. (2019) CMOS: Circuit Design, Layout, and Simulation. 4th ed. Hoboken, NJ: Wiley

    2019

  54. [100]

    (1999) Hybrid CMOS/nanoelectronic circuits: Opportuni- ties and challenges

    Likharev, K.K. (1999) Hybrid CMOS/nanoelectronic circuits: Opportuni- ties and challenges. Journal of Nanoelectronics and Optoelectronics, 3(3), pp. 203–230

    1999

  55. [101]

    and Marr, B

    Hasler, J. and Marr, B. (2013) Finding a roadmap to achieve large neuromorphic hard- ware systems. Frontiers in Neuroscience, 7, 118. https://doi.org/10.3389/fnins.2013.00118

    work page doi:10.3389/fnins.2013.00118 2013

  56. [102]

    and Diorio, C

    Hasler, J., Anderson, D., Minch, B.A. and Diorio, C. (2011) A floating-gate analog memory for fine- grain tuning of VLSI circuits. IEEE Transactions on Circuits and Systems I, 58(7), pp. 1486–1499. https://doi.org/10.1109/TCSI.2010.2095907

    work page doi:10.1109/tcsi.2010.2095907 2011

  57. [103]

    and Ng, K.K

    Sze, S.M. and Ng, K.K. (2006) Physics of Semicon- ductor Devices. 3rd ed. Hoboken, NJ: Wiley

    2006

  58. [104]

    (2001) Analysis and Design of Analog Integrated Circuits

    Gray, P.R., Hurst, P.J., Lewis, S.H.andMeyer, R.G. (2001) Analysis and Design of Analog Integrated Circuits. 4th ed. New York: Wiley

    2001

  59. [105]

    and Garcia-Escartin, J.C

    Herrero-Collantes, M. and Garcia-Escartin, J.C. (2017) Quantum random number generators. Reviews of Modern Physics, 89(1), 015004. https://doi.org/10.1103/RevModPhys.89.015004

    work page doi:10.1103/revmodphys.89.015004 2017

  60. [106]

    National Institute of Standards and Technology

    NIST (2018) SP 800-90B: Recommendation for the Entropy Sources Used for Random Bit Generation. National Institute of Standards and Technology

    2018

  61. [107]

    and Rose, J

    Kuon, I. and Rose, J. (2007) Measuring the gap between FPGAs and ASICs. IEEE Trans- actions on Computer-Aided Design of Inte- grated Circuits and Systems, 26(2), pp. 203–215. https://doi.org/10.1109/TCAD.2006.884574

    work page doi:10.1109/tcad.2006.884574 2007

  62. [108]

    and Du, Y

    Glover, F., Kochenberger, G. and Du, Y. (2019) Quantum annealing and related optimiza- tion methods. Physics Reports, 799, pp. 1–66. https://doi.org/10.1016/j.physrep.2018.12.002

    work page doi:10.1016/j.physrep.2018.12.002 2019

  63. [109]

    and Roy, A

    Boothby, K., King, A.D. and Roy, A. (2020) Fast clique minor generation in Chimera qubit connec- tivity graphs. Quantum Information Processing, 19,

    2020

  64. [110]

    https://doi.org/10.1007/s11128-020-02634-3

    work page doi:10.1007/s11128-020-02634-3

  65. [111]

    (1986) Quantum mechanical com- puters

    Feynman, R.P. (1986) Quantum mechanical com- puters. Foundations of Physics, 16(6), pp. 507–531. https://doi.org/10.1007/BF01886518

    work page doi:10.1007/bf01886518 1986

  66. [112]

    and Vyalyi, M.N

    Kitaev, A.Y., Shen, A.H. and Vyalyi, M.N. (2002) Classical and Quantum Computation. Providence, RI: American Mathematical Society

    2002

  67. [113]

    and Regev, O

    Aharonov, D., van Dam, W., Kempe, J., Landau, Z., Lloyd, S. and Regev, O. (2008) Adiabatic quantum computation is equivalent to standard quantum computation. SIAM Journal on Computing, 37(1), pp. 166–194. https://doi.org/10.1137/050644282

    work page doi:10.1137/050644282 2008

  68. [114]

    and Preskill, J

    Jordan, S.P., Lee, K.S.M. and Preskill, J. (2012) Quantum algorithms for quantum field theories. Science, 336(6085), pp. 1130–1133. https://doi.org/10.1126/science.1217069

    work page doi:10.1126/science.1217069 2012

  69. [115]

    and Lloyd, S

    Biamonte, J., Wittek, P., Pancotti, N., Reben- trost, P., Wiebe, N. and Lloyd, S. (2017) Quan- tum machine learning. Nature, 549, pp. 195–202. https://doi.org/10.1038/nature23474

    work page doi:10.1038/nature23474 2017

  70. [116]

    and Wilmer, E.L

    Levin, D.A., Peres, Y. and Wilmer, E.L. (2009) Markov Chains and Mixing Times. Providence, RI: American Mathematical Society

    2009

  71. [117]

    and Towles, B

    Dally, W.J. and Towles, B. (2004) Principles and Practices of Interconnection Networks. San Fran- cisco, CA: Morgan Kaufmann

    2004

  72. [118]

    Geneva: ID Quantique SA

    ID Quantique (2025) Quantis Quantum Random Number Generator: Technical Description. Geneva: ID Quantique SA

    2025

  73. [119]

    (1997) The Art of Computer Program- ming, Volume 2: Seminumerical Algorithms

    Knuth, D.E. (1997) The Art of Computer Program- ming, Volume 2: Seminumerical Algorithms. 3rd ed. Reading, MA: Addison-Wesley

    1997

  74. [120]

    and Flannery, B.P

    Press, W.H., Teukolsky, S.A., Vetterling, W.T. and Flannery, B.P. (2007) Numerical Recipes: The Art of Scientific Computing. 3rd ed. Cambridge: Cam- bridge University Press

    2007

  75. [121]

    and Thomas, J.A

    Cover, T.M. and Thomas, J.A. (2006) Elements of Information Theory. 2nd ed. Hoboken, NJ: Wiley

    2006

  76. [122]

    and Geman, D

    Geman, S. and Geman, D. (1984) Stochastic re- laxation, Gibbs distributions, and the Bayesian restoration of images. IEEE Transactions on Pat- tern Analysis and Machine Intelligence, 6(6), pp. 721–741

    1984

  77. [123]

    and Roy, K

    Aadit, N.A., Grimaldi, S., Carpentieri, M., Finoc- chio, G. and Roy, K. (2022) Massively par- allel probabilistic computing with sparse Ising machines. Nature Electronics, 5, pp. 460–468. https://doi.org/10.1038/s41928-022-00758-3

    work page doi:10.1038/s41928-022-00758-3 2022

  78. [124]

    and Datta, S

    Singh, J., Camsari, K.Y. and Datta, S. (2023) Prob- abilistic computing with stochastic nanomagnets. IEEE Journal on Exploratory Solid-State Compu- tational Devices and Circuits, 9(1), pp. 1–12. 31

    2023

  79. [125]

    and Roy, K

    Si, X., Yang, J., Wang, Z. and Roy, K. (2024) Energy-efficient probabilistic computing using ana- log hardware. IEEE Transactions on Circuits and Systems I, 71(2), pp. 512–524

    2024

  80. [126]

    and Wang, X

    Hua, X., Zhang, Y., Li, P. and Wang, X. (2025) Ultra-low-energy probabilistic accelerators for opti- mization workloads. Nature Electronics, forthcom- ing / early access

    2025

Showing first 80 references.