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

arXiv:2607.09882

Benchmarking Zero-Setup Quantum Circuit Simulators

Arul Rhik Mazumder, Hovnatan Karapetyan, Mohammed Zuhair Mullath, Rudy H. Tanin, Hrant Gharibyan, Hayk Tepanyan

quant-phcs.DCcs.PFSubmitted Jul 10, 2026• Announced Jul 14, 2026• 0 citations• Revised Sep 2, 2026 (v4)
v1 · Jul 10, 2026v2 · Aug 29, 2026v3 · Sep 1, 2026v4 · Sep 2, 2026
arXivHTMLPDF

Abstract

Practitioners increasingly rely on hosted simulation environments, but their performance characteristics remain poorly documented. We present a systematic benchmarking study of GPU-accelerated approximate quantum simulation across two widely used methods: matrix product states (MPS) and Pauli path simulation (PPS), comparing BlueQubit (a hosted tool that handles hardware provisioning, simulator configuration, and job orchestration) against AWS Braket, Quantum Rings, Qiskit pauli-prop, and PauliPropagation (written in Julia). For MPS, we find that GPU runtime yields sub-quadratic scaling with bond dimension, with a growing advantage over CPU at increasing scale. For Pauli path simulation on IBM's 127-qubit kicked Ising benchmark, GPUs deliver up to ${\sim}1{,}700\times$ speedup at fine truncation thresholds ($δ= 2.5 \times 10^{-5}$, 27.6M Pauli terms), and are the only backends that reach accuracy regimes below $δ= 10^{-5}$, which remained inaccessible to the commodity CPU-based implementations and self-contained SDKs evaluated here. We also provide a reproducible characterization of these simulators across regimes, including tradeoffs that isolated evaluations do not show. All benchmarking code and configurations are in a public GitHub repository.

Comments
12 pages, 12 figures

Related papers

A Practical Guide to using Pauli Path Simulators for Utility-Scale Quantum Experiments

Hrant Gharibyan, Siddharth Hariprakash, Mohammed Zuhair Mullath +1 more

In this this paper we present an inexpensive protocol to perform runtime and memory estimation for large-scale experiments with Pauli Path simulators (PPS). Additionally, we propose a conceptually simple solution for studying whether PPS can be used as a scientific discovery tool, rather than reproducing existing answers. We start by analyzing the dynamics of the Pauli coefficients tracked in the Heisenberg picture. In addition to surprisingly generic convergence features of the Pauli coefficient distributions, we find certain regularities that allow for extrapolation of memory and runtime requirements for smaller and smaller coefficient truncation parameter $δ$. We then introduce a framework for understanding convergence in the absence of rigorous error guarantees on PPS. Combined with runtime analysis, we propose bifurcating quantum simulation problems broadly into two classes, based on whether there is apparent convergence of expectation values as a function of $δ$. This serves as a way for practitioners to understand where their problem falls on the frontier of classical simulability. In the case without apparent convergence, PPS may still serve useful as a Monte Carlo-like estimate. Applied to IBM's utility-scale experiments, we show parameter regimes where both behaviors are realized. Some of our key findings challenge conventional intuition: reducing $δ$ does not always improve accuracy, and deeper quantum circuits may actually be easier to simulate than shallower ones. The BlueQubit SDK implementing these methods has been released publicly, offering researchers a comprehensive toolkit for evaluating this frontier classical simulation approach. These results establish practical guidelines for when PPS can serve as a reliable verification tool versus when it should be used as a complementary estimate alongside quantum experiments.

arXiv:2507.10771•Jul 14, 2025•0 citations

SIMULATIONS OF QUANTUM ALGORITHMS USING THE QISKIT AER LIBRARY AND THE MATRIX PRODUCT STATE METHOD ON CPU AND GPU

Димитр Костадінович Йовчев

This article presents a comprehensive study of the performance of quantum circuit simulation using the Matrix Product State (MPS) method within Qiskit Aer. The primary focus is on comparing the computational efficiency of implementations on central processing unit (CPU) and graphics processing unit (GPU) across various simulation platforms. Four key quantum algorithms are experimentally examined: the Variational Quantum Eigensolver (VQE), the Quantum Fourier Transform (QFT), the Quantum Approximate Optimization Algorithm (QAOA), and the Greenberger– Horne–Zeilinger (GHZ) state. For each algorithm, graphs are constructed showing the dependence of execution time on the number of qubits and the type of simulator. A single diagram is also provided to compare all four algorithms using the MPS method in Qiskit Aer. The results indicate that although the MPS simulator in Qiskit Aer is currently implemented only for CPU, it outperforms the traditional statevector simulator for low-entanglement circuits. At the same time, the lack of GPU acceleration limits the scalability of the MPS simulator. Analysis shows that implementing hardware acceleration of SVD operations using NVIDIA's cuQuantum and cuTensorNet libraries can provide a speedup of up to seven times. A comparative analysis of GPU implementations in Qiskit Aer and CUDA-Q is also conducted, including libraries such as cuStateVec, LAPACK, and cuSOLVER. Graphs are created to demonstrate the strengths and weaknesses of the simulators and simulation methods in Qiskit Aer. Notably, the MPS method remains one of the fastest by computation speed on CPU, but is sometimes surpassed by other methods on GPU. Further optimization directions for MPS are proposed, including the development of a hybrid CPU/GPU architecture and the utilization of modern GPU libraries to accelerate critical operations. Keywords: quantum computing, quantum simulation, Qiskit Aer, MPS, statevector, cudaq.

Telecommunication and information technologies•Jan 1, 2025•0 citations

GPU-Accelerated Quantum Simulation: Empirical Backend Selection, Gate Fusion, and Adaptive Precision

Poornima Kumaresan, Pavithra Muruganantham, Lakshmi Rajendran +1 more

Classical simulation of quantum circuits remains indispensable for algorithm development, hardware validation, and error analysis in the noisy intermediate-scale quantum (NISQ) era. However, state-vector simulation faces exponential memory scaling, with an n-qubit system requiring O(2^n) complex amplitudes, and existing simulators often lack the flexibility to exploit heterogeneous computing resources at runtime. This paper presents a GPU-accelerated quantum circuit simulation framework that introduces three contributions: (1) an empirical backend selection algorithm that benchmarks CuPy, PyTorch-CUDA, and NumPy-CPU backends at runtime and selects the optimal execution path based on measured throughput; (2) a directed acyclic graph (DAG) based gate fusion engine that reduces circuit depth through automated identification of fusible gate sequences, coupled with adaptive precision switching between complex64 and complex128 representations; and (3) a memory-aware fallback mechanism that monitors GPU memory consumption and gracefully degrades to CPU execution when resources are exhausted. The framework integrates with Qiskit, Cirq, PennyLane, and Amazon Braket through a unified adapter layer. Benchmarks on an NVIDIA A100-SXM4 (40 GiB) GPU demonstrate speedups of 64x to 146x over NumPy CPU execution for state-vector simulation of circuits with 20 to 28 qubits, with speedups exceeding 5x from 16 qubits onward. Hardware validation on an IBM quantum processing unit (QPU) confirms Bell state fidelity of 0.939, a five-qubit Greenberger-Horne-Zeilinger (GHZ) state fidelity of 0.853, and circuit depth reduction from 42 to 14 gates through the fusion pipeline. The system is designed for portability across NVIDIA consumer and data-center GPUs, requiring no vendor-specific compilation steps.

quant-pharXiv:2604.03816•Apr 7, 2026•0 citations

Quantum Computer Simulations at Warp Speed: Assessing the Impact of GPU Acceleration: A Case Study with IBM Qiskit Aer, Nvidia Thrust & cuQuantum

Jennifer Faj, Ivy Bo Peng, Jacob Wahlgren +1 more

Quantum computer simulators are crucial for the development of quantum computing. This work investigates GPU and multi-GPU systems' suitability and performance impact on a widely used simulation tool – the state vector simulator Qiskit Aer. In particular, we evaluate the performance of both Qiskit's default Nvidia Thrust backend and the recent Nvidia cuQuantum backend on Nvidia A100 GPUs. We provide a benchmark suite of representative quantum applications for characterization. For simulations with a large number of qubits, the two GPU backends can provide up to 14× speedup over the CPU backend, with Nvidia cuQuantum providing a further 1.5–3× speedup over the default Thrust backend. Our evaluation on a single GPU identifies the most important functions in Nvidia Thrust and cuQuantum for different quantum applications and their compute and memory bottlenecks. We also evaluate the gate fusion and cache-blocking optimizations on different quantum applications. Finally, we evaluate large-number qubit quantum applications on multi-GPU and identify data movement between host and GPU as the limiting factor for the performance.

IEEE International Conference on eScience•Sep 25, 2023•17 citations

Quantum Computer Simulations at Warp Speed: Assessing the Impact of GPU Acceleration

Jennifer Faj, Ivy Bo Peng, Jacob Wahlgren +1 more

Quantum computer simulators are crucial for the development of quantum computing. In this work, we investigate the suitability and performance impact of GPU and multi-GPU systems on a widely used simulation tool - the state vector simulator Qiskit Aer. In particular, we evaluate the performance of both Qiskit's default Nvidia Thrust backend and the recent Nvidia cuQuantum backend on Nvidia A100 GPUs. We provide a benchmark suite of representative quantum applications for characterization. For simulations with a large number of qubits, the two GPU backends can provide up to 14x speedup over the CPU backend, with Nvidia cuQuantum providing further 1.5-3x speedup over the default Thrust backend. Our evaluation on a single GPU identifies the most important functions in Nvidia Thrust and cuQuantum for different quantum applications and their compute and memory bottlenecks. We also evaluate the gate fusion and cache-blocking optimizations on different quantum applications. Finally, we evaluate large-number qubit quantum applications on multi-GPU and identify data movement between host and GPU as the limiting factor for the performance.

arXiv:2307.14860•Jul 27, 2023•2 citations

Realizing Quantum Kernel Models at Scale with Matrix Product State Simulation

Mekena Metcalf, Pablo Andrés-Martí­nez, Nathan Fitzpatrick

Data representation in quantum state space offers an alternative function space for machine learning tasks. However, benchmarking these algorithms at a practical scale has been limited by ineffective simulation methods. We develop a quantum kernel framework using a Matrix Product State (MPS) simulator and employ it to perform a classification task with 165 features and 6400 training data points, well beyond the scale of any prior work. We make use of a circuit ansatz on a linear chain of qubits with increasing interaction distance between qubits. We assess the MPS simulator performance on CPUs and GPUs and, by systematically increasing the qubit interaction distance, we identify a crossover point beyond which the GPU implementation runs faster. We show that quantum kernel model performance improves as the feature dimension and training data increases, which is the first evidence of quantum model performance at scale.

arXiv:2411.09336•International Conference for High Performance Computing, Networking, Storage and Analysis (SC)•Nov 17, 2024•3 citations
Practitioners increasingly rely on hosted simulation environments, but their performance characteristics remain poorly documented. We present a systematic benchmarking study of GPU-accelerated approximate quantum simulation across two widely used methods: matrix product states (MPS) and Pauli path simulation (PPS), comparing BlueQubit (a hosted tool that handles hardware provisioning, simulator configuration, and job orchestration) against AWS Braket, Quantum Rings, Qiskit pauli-prop, and PauliPropagation (written in Julia). For MPS, we find that GPU runtime yields sub-quadratic scaling with bond dimension, with a growing advantage over CPU at increasing scale. For Pauli path simulation on IBM's 127-qubit kicked Ising benchmark, GPUs deliver up to ∼1,700×{\sim}1{,}700\times∼1,700× speedup at fine truncation thresholds (δ=2.5×10−5δ= 2.5 \times 10^{-5}δ=2.5×10−5, 27.6M Pauli terms), and are the only backends that reach accuracy regimes below δ=10−5δ= 10^{-5}δ=10−5, which remained inaccessible to the commodity CPU-based implementations and self-contained SDKs evaluated here. We also provide a reproducible characterization of these simulators across regimes, including tradeoffs that isolated evaluations do not show. All benchmarking code and configurations are in a public GitHub repository.
In this this paper we present an inexpensive protocol to perform runtime and memory estimation for large-scale experiments with Pauli Path simulators (PPS). Additionally, we propose a conceptually simple solution for studying whether PPS can be used as a scientific discovery tool, rather than reproducing existing answers. We start by analyzing the dynamics of the Pauli coefficients tracked in the Heisenberg picture. In addition to surprisingly generic convergence features of the Pauli coefficient distributions, we find certain regularities that allow for extrapolation of memory and runtime requirements for smaller and smaller coefficient truncation parameter δδδ. We then introduce a framework for understanding convergence in the absence of rigorous error guarantees on PPS. Combined with runtime analysis, we propose bifurcating quantum simulation problems broadly into two classes, based on whether there is apparent convergence of expectation values as a function of δδδ. This serves as a way for practitioners to understand where their problem falls on the frontier of classical simulability. In the case without apparent convergence, PPS may still serve useful as a Monte Carlo-like estimate. Applied to IBM's utility-scale experiments, we show parameter regimes where both behaviors are realized. Some of our key findings challenge conventional intuition: reducing δδδ does not always improve accuracy, and deeper quantum circuits may actually be easier to simulate than shallower ones. The BlueQubit SDK implementing these methods has been released publicly, offering researchers a comprehensive toolkit for evaluating this frontier classical simulation approach. These results establish practical guidelines for when PPS can serve as a reliable verification tool versus when it should be used as a complementary estimate alongside quantum experiments.
This article presents a comprehensive study of the performance of quantum circuit simulation using the Matrix Product State (MPS) method within Qiskit Aer. The primary focus is on comparing the computational efficiency of implementations on central processing unit (CPU) and graphics processing unit (GPU) across various simulation platforms. Four key quantum algorithms are experimentally examined: the Variational Quantum Eigensolver (VQE), the Quantum Fourier Transform (QFT), the Quantum Approximate Optimization Algorithm (QAOA), and the Greenberger– Horne–Zeilinger (GHZ) state. For each algorithm, graphs are constructed showing the dependence of execution time on the number of qubits and the type of simulator. A single diagram is also provided to compare all four algorithms using the MPS method in Qiskit Aer. The results indicate that although the MPS simulator in Qiskit Aer is currently implemented only for CPU, it outperforms the traditional statevector simulator for low-entanglement circuits. At the same time, the lack of GPU acceleration limits the scalability of the MPS simulator. Analysis shows that implementing hardware acceleration of SVD operations using NVIDIA's cuQuantum and cuTensorNet libraries can provide a speedup of up to seven times. A comparative analysis of GPU implementations in Qiskit Aer and CUDA-Q is also conducted, including libraries such as cuStateVec, LAPACK, and cuSOLVER. Graphs are created to demonstrate the strengths and weaknesses of the simulators and simulation methods in Qiskit Aer. Notably, the MPS method remains one of the fastest by computation speed on CPU, but is sometimes surpassed by other methods on GPU. Further optimization directions for MPS are proposed, including the development of a hybrid CPU/GPU architecture and the utilization of modern GPU libraries to accelerate critical operations. Keywords: quantum computing, quantum simulation, Qiskit Aer, MPS, statevector, cudaq.
Classical simulation of quantum circuits remains indispensable for algorithm development, hardware validation, and error analysis in the noisy intermediate-scale quantum (NISQ) era. However, state-vector simulation faces exponential memory scaling, with an n-qubit system requiring O(2^n) complex amplitudes, and existing simulators often lack the flexibility to exploit heterogeneous computing resources at runtime. This paper presents a GPU-accelerated quantum circuit simulation framework that introduces three contributions: (1) an empirical backend selection algorithm that benchmarks CuPy, PyTorch-CUDA, and NumPy-CPU backends at runtime and selects the optimal execution path based on measured throughput; (2) a directed acyclic graph (DAG) based gate fusion engine that reduces circuit depth through automated identification of fusible gate sequences, coupled with adaptive precision switching between complex64 and complex128 representations; and (3) a memory-aware fallback mechanism that monitors GPU memory consumption and gracefully degrades to CPU execution when resources are exhausted. The framework integrates with Qiskit, Cirq, PennyLane, and Amazon Braket through a unified adapter layer. Benchmarks on an NVIDIA A100-SXM4 (40 GiB) GPU demonstrate speedups of 64x to 146x over NumPy CPU execution for state-vector simulation of circuits with 20 to 28 qubits, with speedups exceeding 5x from 16 qubits onward. Hardware validation on an IBM quantum processing unit (QPU) confirms Bell state fidelity of 0.939, a five-qubit Greenberger-Horne-Zeilinger (GHZ) state fidelity of 0.853, and circuit depth reduction from 42 to 14 gates through the fusion pipeline. The system is designed for portability across NVIDIA consumer and data-center GPUs, requiring no vendor-specific compilation steps.
Quantum computer simulators are crucial for the development of quantum computing. This work investigates GPU and multi-GPU systems' suitability and performance impact on a widely used simulation tool – the state vector simulator Qiskit Aer. In particular, we evaluate the performance of both Qiskit's default Nvidia Thrust backend and the recent Nvidia cuQuantum backend on Nvidia A100 GPUs. We provide a benchmark suite of representative quantum applications for characterization. For simulations with a large number of qubits, the two GPU backends can provide up to 14× speedup over the CPU backend, with Nvidia cuQuantum providing a further 1.5–3× speedup over the default Thrust backend. Our evaluation on a single GPU identifies the most important functions in Nvidia Thrust and cuQuantum for different quantum applications and their compute and memory bottlenecks. We also evaluate the gate fusion and cache-blocking optimizations on different quantum applications. Finally, we evaluate large-number qubit quantum applications on multi-GPU and identify data movement between host and GPU as the limiting factor for the performance.
Quantum computer simulators are crucial for the development of quantum computing. In this work, we investigate the suitability and performance impact of GPU and multi-GPU systems on a widely used simulation tool - the state vector simulator Qiskit Aer. In particular, we evaluate the performance of both Qiskit's default Nvidia Thrust backend and the recent Nvidia cuQuantum backend on Nvidia A100 GPUs. We provide a benchmark suite of representative quantum applications for characterization. For simulations with a large number of qubits, the two GPU backends can provide up to 14x speedup over the CPU backend, with Nvidia cuQuantum providing further 1.5-3x speedup over the default Thrust backend. Our evaluation on a single GPU identifies the most important functions in Nvidia Thrust and cuQuantum for different quantum applications and their compute and memory bottlenecks. We also evaluate the gate fusion and cache-blocking optimizations on different quantum applications. Finally, we evaluate large-number qubit quantum applications on multi-GPU and identify data movement between host and GPU as the limiting factor for the performance.
Data representation in quantum state space offers an alternative function space for machine learning tasks. However, benchmarking these algorithms at a practical scale has been limited by ineffective simulation methods. We develop a quantum kernel framework using a Matrix Product State (MPS) simulator and employ it to perform a classification task with 165 features and 6400 training data points, well beyond the scale of any prior work. We make use of a circuit ansatz on a linear chain of qubits with increasing interaction distance between qubits. We assess the MPS simulator performance on CPUs and GPUs and, by systematically increasing the qubit interaction distance, we identify a crossover point beyond which the GPU implementation runs faster. We show that quantum kernel model performance improves as the feature dimension and training data increases, which is the first evidence of quantum model performance at scale.