Matrix-product-state simulations reproduce exact results to N=48

Gerhard Hellstern of Duale Hochschule Baden-Württemberg (DHBW) Stuttgart and Marc Maussner have demonstrated a digital quantum simulation benchmark using instanton-mediated tunnelling, achieving accurate results up to 48 lattice sites. Their work verifies the dynamics of a lattice model against an exact classical solution, proving that a specific quantum circuit structure maintains a Schmidt rank of at most four at every gate boundary, regardless of the system’s size. Matrix-product-state simulations corroborate these digital quantum simulations, while measurements on superconducting processors at N = 8 recover lattice splitting to within a few percent.

Instanton Tunnelling as Benchmark: Classical-Quantum Correspondence

The admissible Trotter step for this simulation shrinks as N–2.28 over 16 to 64 lattice sites, meaning computational cost increases with scale, but remains manageable for verifying quantum dynamics against classical solutions. This scaling indicates that simulating one period of tunnelling requires approximately N3.3 two-qubit gates, a figure determined by splitting the kinetic term bond by bond. A kinetic/potential splitting strategy, however, maintains a step size near 0.065 up to 128 lattice sites, suggesting a pathway to mitigate the gate count increase.

These calculations establish a clear relationship between system size and computational demands for this specific quantum simulation approach. Researchers measured full transfer at a lattice size of eight using two superconducting processors, achieving lattice splitting recovery to within a few percent.

This hardware validation confirms the dependence of the splitting on the lattice structure itself, rather than being limited by the resolution of the instanton calculation at this scale. The researchers used the transformation from a real-time to an imaginary-time formulation to recast the tunnelling problem as a classical propagation problem within an inverted potential. This allowed them to identify a well-defined “path”, the instanton, which captures the non-perturbative nature of vacuum fluctuations.

Unlike perturbative effects, which scale with Planck’s constant, instanton-induced effects are proportional to exp(-S_0/ℏ), introducing an essential singularity at zero Planck’s constant. This approach provides a robust framework for benchmarking quantum simulations against an exact classical solution at every step, offering a unique opportunity to verify the accuracy and scalability of quantum hardware.

Unary Encoding Maps Lattice Sites to Qubits

A unary encoding directly maps lattice site number to qubit number, simplifying the representation of the quantum system and enabling classically inexpensive reference calculations at any scale. This approach assigns each spatial location on the lattice to a single qubit, confining the dynamics to a single-excitation sector that streamlines the simulation process.

The structure of the quantum circuit employed further constrains the complexity of the simulation; the Schmidt rank, a measure of entanglement, is mathematically limited to a maximum of four at every gate boundary. This constraint, independent of the number of lattice sites, suggests a degree of controllability within the system and simplifies the verification process.

The admissibility of the Trotter step decreases roughly as the inverse square of the lattice size (N^(-2)). However, the researchers found this scaling to be N^(-2. 20) to N^(-2. 12) for N ≥ 24, depending on the accuracy criterion used.

Matrix-Product-State Simulation Reproduces Exact Results to N=48

Matrix-product-state (MPS) simulations have successfully mirrored the results of digital quantum simulations up to a lattice size of 48 qubits, validating the accuracy of classical methods against increasingly complex quantum systems. Beyond this scale, the research team rigorously tested the accuracy of their approach across multiple computational levels, labelled tiers A through E. Tiers A, B, and E focused on classically verifiable calculations of the double-well Hamiltonian, allowing for direct comparison with known exact answers.

Tier C validation involved comparing the simulated survival probability, the likelihood of a quantum state remaining in a specific configuration over time, with the exact cosine squared function, achieving agreement to within 6. 6 x 10-5 across lattice sizes ranging from 12 to 48. Finite-window extrapolations of the tunnel splitting from both gate-level MPS simulations and exact diagonalisation differed by only 6. 3 x 10-5, demonstrating a high degree of consistency between the two methods. Reducing the Trotter step size by a factor of two consistently reduced the error in the calculated energy splitting as O(δt2), confirming the expected scaling behavior.

The excitation number, a measure of the system’s deviation from its ground state, remained below 10-12, indicating a high level of precision. The authors acknowledge the presence of higher-order corrections.

Trotter Step Scaling with Lattice Size (N)

The admissible Trotter step for these simulations decreases proportionally to N–2, meaning the computational effort required for accurate modelling scales with the cube of the lattice size, specifically N^(3. This scaling arises from a bond-by-bond approach to splitting the kinetic term, a method that maintains a step size near 0.

This achievement extends beyond previous demonstrations of full transfers on small quantum registers by incorporating a repeated, session-resolved analysis that isolates uncertainty stemming from differences between experimental runs rather than random noise. The team regards this reproducibility-qualified measurement, alongside the benchmark itself, as the primary contribution of their research.

Reproducibility of Quantum Runs Across Time & Processors

At intermediate system sizes, simulating quantum dynamics demands substantial computational resources. This highlights a critical constraint: maintaining accuracy as system size increases necessitates careful management of gate errors and fidelity loss. Researchers calculated the number of Trotter steps needed to span one tunnelling period, comparing it to the steps permissible within a fidelity budget, revealing the trade-offs between simulation depth and accuracy.

Over 9 repetitions, runs taken minutes apart agree to shot noise, while runs hours to weeks apart differ by up to 6.1 %, also on a fixed set of qubits; with sessions as the sampling unit (4 and 3 sessions), the processors differ in the damping ( p≈ 0.02), while a difference in the splitting is not established ( p≈ 0.10). The mathematical structure of the quantum circuit itself imposes constraints on entanglement growth.

Normalising by the measured single-excitation population extends the usable depth about 4-fold but does not improve the splitting. A Gaussian start state, which needs no diagonalisation, reproduces the splitting of the eigenvector-based start to within 1.6 percentage points.

👉 More information
🗞 Instanton tunnelling as a verifiable benchmark for digital quantum simulation: circuit structure, classical simulability and hardware reproducibility
✍️ Gerhard Hellstern (Duale Hochschule Baden-Württemberg (DHBW) Stuttgart, Stuttgart, Germany); Marc Maussner (infoteam Software AG, Bubenreuth, Germany)
🧠 ArXiv: https://arxiv.org/abs/2611.24624v1

Stay current

See today’s quantum computing news on Quantum Zeitgeist for the latest breakthroughs in qubits, hardware, algorithms, and industry deals.

Avatar of Rusty Flint

Rusty Flint

Rusty is a quantum science nerd. He's been into academic science all his life, but spent his formative years doing less academic things. Now he turns his attention to write about his passion, the quantum realm. He loves all things Quantum Physics especially. Rusty likes the more esoteric side of Quantum Computing and the Quantum world. Everything from Quantum Entanglement to Quantum Physics. Rusty thinks that we are in the 1950s quantum equivalent of the classical computing world. While other quantum journalists focus on IBM's latest chip or which startup just raised $50 million, Rusty's over here writing 3,000-word deep dives on whether quantum entanglement might explain why you sometimes think about someone right before they text you. (Spoiler: it doesn't, but the exploration is fascinating)

Latest Posts by Rusty Flint: