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

Scaling Quantum State-Vector Simulation to 36 Qubits on Google Cloud TPU: A Pure-JAX Architecture Without Out-of-Memory Failures

Ashitesh Singh

Published Jun 15, 2026• 0 citations
Publisher

Abstract

Classical simulation of quantum circuits scales exponentially: an n-qubit state vector requires $2^{n}$ complex amplitudes, consuming $2^{n}\times8$ bytes of memory for single-precision complex numbers. On a single accelerator this wall is typically hit at 29-30 qubits $(\approx4.3\text{ GB})$. We describe an architecture built entirely in pure JAX with zero dependency on Qiskit, PennyLane, Cirq, or any classical quantum framework that scales quantum state-vector simulation to 36 qubits (549.76 GB state-vector footprint) across a 64-chip Google Cloud TPU v6e mesh, and to 33 qubits (64 GB) across a 16-chip TPU v5e cluster. The key insight is that JAX's PositionalSharding distributes the $2^{n}$-amplitude state vector uniformly across chip HBM pools as a first-class XLA buffer, while jax.lax.fori_loop keeps the XLA computation graph at $\mathcal{O}(1)$ size regardless of circuit depth, preventing the host-side Out-of-Memory (OOM) errors that plague compiler-unrolled deep circuits. We demonstrate five experiments (GHZ state preparation, VQC, VQE H2, QAOA MaxCut, Monte Carlo noise simulation) running at sub-millisecond gate latency after JIT compilation, with full reverse-mode automatic differentiation through the entire circuit.

DOI
10.5281/zenodo.20704989

Related papers

JAX Quantum Research Suite: A Unified, Hardware-Accelerated, Differentiable Simulator for NISQ-Era Algorithm Research Across GPU and Cloud TPU Clusters

Ashitesh Singh

We present the JAX Quantum Research Suite, a high-performance, differentiable quantum circuit simulator spanning two co-existing hardware acceleration layers: a GPU division targeting NVIDIA RTX-class consumer GPUs via CUDA and a Cloud TPU division distributed across Google Cloud TPU v5e-16 and v6e-64 VM clusters. The suite is implemented in pure JAX (jax.numpy, jax.lax) for all statevector primitives, enabling seamless composition with jax.grad, jax.vmap, and jax.sharding.Positional Sharding.

•Jun 15, 2026•0 citations

Simulation of Quantum Many-Body Dynamics with Tensor Processing Units: Floquet Prethermalization

Alan Morningstar, Markus Hauru, Jackson Beall +4 more

Tensor processing units (TPUs) are specialized hardware accelerators developed by Google to support large-scale machine-learning tasks but they can also be leveraged to accelerate and scale other linearalgebra-intensive computations.In this paper, we demonstrate the usage of TPUs for massively parallel classical simulations of quantum many-body dynamics on long time scales.We apply our methods to study the phenomenon of Floquet prethermalization, i.e., exponentially slow heating in quantum spin chains subject to high-frequency periodic driving.We simulate the dynamics of L = 34 qubits for over 10 5 Floquet periods, corresponding to circuits with 4 × 10 6 nearest-neighbor two-qubit gates.The circuits simulated have no additional symmetries and represent a pure-state evolution in the full 2 L -dimensional Hilbert space.This is achieved by distributing the computation over 128 TPU cores.On that size TPU cluster, we find speed-ups in wall-clock run time of 230 times and 15 times when compared to reference CPU and single-graphics-processing-unit (GPU) simulations, respectively, for shorter-time 30-qubit simulations that can be handled by all three platforms.We study the computational cost of the simulations, as a function of both the number of qubits and the number of TPU cores used, up to our maximum capacity of L = 40 qubits, which requires a "full pod" of 2048 TPU cores with tens of terabytes of memory in total.For these simulations, an eight-TPU-core machine is comparable to a single A100 GPU and thus the full TPU pod is comparable to a machine with hundreds of top-of-the-line GPUs.However, the TPU pod is more energy and cost efficient and readily accessible (via Google Cloud), unlike such large many-GPU configurations.We also study the accumulation of numerical error as a function of circuit depth in very deep circuits.Our work demonstrates that TPUs can offer significant advantages for state-of-the-art simulations of quantum many-body dynamics.

arXiv:2111.08044•PRX Quantum•May 11, 2022•38 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•3 citations

JAQSI: Just Another Quantum Simulator

Melvin Strobl, Maja Franz, Lukas Scheller +3 more

Classical simulation supports the development and validation of quantum algorithms through full-state inspection, controlled noise models and backpropagation, albeit for small quantum systems. Existing frameworks offer gate-level execution, differentiable pulse-level dynamics or optimised hardware execution. Just another quantum simulator (JAQSI) combines gate- and pulse-level simulation in one differentiable circuit model built upon JAX. Circuits are represented as pure functions of numerical inputs which support vectorised batches and JIT compilation. Within a single circuit, matrix-defined operations, gates synthesised from Hamiltonian time evolution and noise channels can be composed freely. For a fast simulation, gate and channel fusion reduces passes over the state, while adjoint differentiation computes expectation-value gradients with a constant number of statevectors. At pulse level, JAQSI reduces evolution under single-term Hamiltonians to a scalar pulse-area integral and reuses cached execution plans and compiled solvers. In a benchmark we compare JAQSI with PennyLane, Qiskit, Qibo, Qulacs, QuTiP and dynamiqs. At gate level, we observe performance either on par or an order of magnitude faster across statevector, density, noise and gradient simulation than the compared simulators up to 12 qubits. Beyond 18 qubits, JAQSI is on par with PennyLane Lightning in both statevector and gradient simulation. In the statevector and noisy density-matrix pulse benchmarks, it is at least 5 times faster than the other simulators, even with its closed-form solve disabled. Disabling the latter increases its runtime by at most a factor of 1.5. We provide JAQSI as open-source software on GitHub and PyPI.

quant-pharXiv:2610.05086•Oct 6, 2026•0 citations

Breaking the Quantum Memory Wall on Consumer Silicon: Exact 28-to-53 Qubit Sub-Manifold Contraction and Paraconsistent Thermodynamics on a Single 4GB GPU

Maycol Jhonatan Benavides Sánchez

The classical emulation of quantum supremacy circuits and large-scale industrial quantum optimization is fundamentally bottlenecked by the exponential Hilbert-space "Memory Wall" (O(2^N)). Standard simulation frameworks (such as IBM Qiskit Aer and NVIDIA cuQuantum) suffer immediate Out-of-Memory (OOM) crashes on commercial GPUs (4GB VRAM) when attempting universal state-vector simulation at N = 28 qubits, while brute-force supercomputing approaches—exemplified by the 41.9-million-core New Sunway system—require 1.3 Petabytes of RAM and 15 Megawatts of power to resolve the 53-qubit Google Sycamore circuit (m = 20) in 304.4 seconds. In this work, we present the unified Chronos-Quantum Engine, a proprietary GPU microarchitecture operating within a strict 3.58 GB VRAM ceiling on a single 60W laptop GPU (NVIDIA RTX 3050 Ti). We demonstrate three verified operational regimes on physical silicon: (1) Exact 28-Qubit Universal State-Vector Evolution (268,435,456 complex states confined to 2.147 GB VRAM) achieving IEEE-754 exactness (L2 error = 8.32 × 10^-12 against FP64) for real-time industrial QAOA optimization and long-time Floquet-MBL discrete time-crystal dynamics; (2) Full 53-Qubit Sycamore Circuit Resolution (S = 128 coherent phase partitions across all 20 layers and all 1,555 gates with zero omissions), achieving global unitarity ||Ψ||^2_2 = 0.99998384, unbiased Porter-Thomas second moment ⟨u2⟩/⟨u⟩2 = 2.014852 (0.038% relative error from theoretical ideal), and Linear Cross-Entropy Benchmarking fidelity FXEB = 1.014852 (1.039321 under Born quantile sampling) in 372.34 seconds—consuming 204,383× less energy and 338,540× less memory than Sunway; and (3) Supremacy Threshold Crossing in 2.96 seconds (Slice 1, FXEB = 0.0235 > 0.002) alongside 783 ms Paraconsistent L4 Thermodynamic Screening at the 95% Overlap Gap Property (OGP) barrier.

•Oct 9, 2026•0 citations

Prototype of a Batched Quantum Circuit Simulator for the Vector Engine

Keichi Takahashi, Toshio Mōri, Hiroyuki Takizawa

State-of-the-art quantum circuit simulators have mostly focused on scaling the number of qubits. However, we argue that studying current noisy quantum computers and variational quantum algorithms benefits from high-throughput simulation of intermediate-scale quantum circuits. We present the first implementation and evaluation of a batched quantum simulator on the NEC Vector Engine (VE), a vector processor with massive memory bandwidth that is ideal for memory-intensive state vector simulation. To take advantage of the long-vector architecture of VE, we design a parallelization strategy and memory layout suited for batched state vector simulation. Our evaluation shows that the latest-generation VE Type 30A achieves equal performance as the NVIDIA A100 80 GB model for single-qubit gates, but slightly underperforms A100 for two-qubit gates. Furthermore, our vector-friendly noise gate implementation offers a large speedup compared to that of cuStateVec. When simulating a noisy 16-qubit random circuit with our simulator, VE30 is up to 1.95 × faster than A100.

International Conference for High Performance Computing, Networking, Storage and Analysis (SC)•Nov 10, 2023•4 citations
Classical simulation of quantum circuits scales exponentially: an n-qubit state vector requires 2n2^{n}2n complex amplitudes, consuming 2n×82^{n}\times82n×8 bytes of memory for single-precision complex numbers. On a single accelerator this wall is typically hit at 29-30 qubits (≈4.3 GB)(\approx4.3\text{ GB})(≈4.3 GB). We describe an architecture built entirely in pure JAX with zero dependency on Qiskit, PennyLane, Cirq, or any classical quantum framework that scales quantum state-vector simulation to 36 qubits (549.76 GB state-vector footprint) across a 64-chip Google Cloud TPU v6e mesh, and to 33 qubits (64 GB) across a 16-chip TPU v5e cluster. The key insight is that JAX's PositionalSharding distributes the 2n2^{n}2n-amplitude state vector uniformly across chip HBM pools as a first-class XLA buffer, while jax.lax.fori_loop keeps the XLA computation graph at O(1)\mathcal{O}(1)O(1) size regardless of circuit depth, preventing the host-side Out-of-Memory (OOM) errors that plague compiler-unrolled deep circuits. We demonstrate five experiments (GHZ state preparation, VQC, VQE H2, QAOA MaxCut, Monte Carlo noise simulation) running at sub-millisecond gate latency after JIT compilation, with full reverse-mode automatic differentiation through the entire circuit.
Tensor processing units (TPUs) are specialized hardware accelerators developed by Google to support large-scale machine-learning tasks but they can also be leveraged to accelerate and scale other linearalgebra-intensive computations.In this paper, we demonstrate the usage of TPUs for massively parallel classical simulations of quantum many-body dynamics on long time scales.We apply our methods to study the phenomenon of Floquet prethermalization, i.e., exponentially slow heating in quantum spin chains subject to high-frequency periodic driving.We simulate the dynamics of L = 34 qubits for over 10 5 Floquet periods, corresponding to circuits with 4 × 10 6 nearest-neighbor two-qubit gates.The circuits simulated have no additional symmetries and represent a pure-state evolution in the full 2 L -dimensional Hilbert space.This is achieved by distributing the computation over 128 TPU cores.On that size TPU cluster, we find speed-ups in wall-clock run time of 230 times and 15 times when compared to reference CPU and single-graphics-processing-unit (GPU) simulations, respectively, for shorter-time 30-qubit simulations that can be handled by all three platforms.We study the computational cost of the simulations, as a function of both the number of qubits and the number of TPU cores used, up to our maximum capacity of L = 40 qubits, which requires a "full pod" of 2048 TPU cores with tens of terabytes of memory in total.For these simulations, an eight-TPU-core machine is comparable to a single A100 GPU and thus the full TPU pod is comparable to a machine with hundreds of top-of-the-line GPUs.However, the TPU pod is more energy and cost efficient and readily accessible (via Google Cloud), unlike such large many-GPU configurations.We also study the accumulation of numerical error as a function of circuit depth in very deep circuits.Our work demonstrates that TPUs can offer significant advantages for state-of-the-art simulations of quantum many-body dynamics.
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.
Classical simulation supports the development and validation of quantum algorithms through full-state inspection, controlled noise models and backpropagation, albeit for small quantum systems. Existing frameworks offer gate-level execution, differentiable pulse-level dynamics or optimised hardware execution. Just another quantum simulator (JAQSI) combines gate- and pulse-level simulation in one differentiable circuit model built upon JAX. Circuits are represented as pure functions of numerical inputs which support vectorised batches and JIT compilation. Within a single circuit, matrix-defined operations, gates synthesised from Hamiltonian time evolution and noise channels can be composed freely. For a fast simulation, gate and channel fusion reduces passes over the state, while adjoint differentiation computes expectation-value gradients with a constant number of statevectors. At pulse level, JAQSI reduces evolution under single-term Hamiltonians to a scalar pulse-area integral and reuses cached execution plans and compiled solvers. In a benchmark we compare JAQSI with PennyLane, Qiskit, Qibo, Qulacs, QuTiP and dynamiqs. At gate level, we observe performance either on par or an order of magnitude faster across statevector, density, noise and gradient simulation than the compared simulators up to 12 qubits. Beyond 18 qubits, JAQSI is on par with PennyLane Lightning in both statevector and gradient simulation. In the statevector and noisy density-matrix pulse benchmarks, it is at least 5 times faster than the other simulators, even with its closed-form solve disabled. Disabling the latter increases its runtime by at most a factor of 1.5. We provide JAQSI as open-source software on GitHub and PyPI.
The classical emulation of quantum supremacy circuits and large-scale industrial quantum optimization is fundamentally bottlenecked by the exponential Hilbert-space "Memory Wall" (O(2^N)). Standard simulation frameworks (such as IBM Qiskit Aer and NVIDIA cuQuantum) suffer immediate Out-of-Memory (OOM) crashes on commercial GPUs (4GB VRAM) when attempting universal state-vector simulation at N = 28 qubits, while brute-force supercomputing approaches—exemplified by the 41.9-million-core New Sunway system—require 1.3 Petabytes of RAM and 15 Megawatts of power to resolve the 53-qubit Google Sycamore circuit (m = 20) in 304.4 seconds. In this work, we present the unified Chronos-Quantum Engine, a proprietary GPU microarchitecture operating within a strict 3.58 GB VRAM ceiling on a single 60W laptop GPU (NVIDIA RTX 3050 Ti). We demonstrate three verified operational regimes on physical silicon: (1) Exact 28-Qubit Universal State-Vector Evolution (268,435,456 complex states confined to 2.147 GB VRAM) achieving IEEE-754 exactness (L2 error = 8.32 × 10^-12 against FP64) for real-time industrial QAOA optimization and long-time Floquet-MBL discrete time-crystal dynamics; (2) Full 53-Qubit Sycamore Circuit Resolution (S = 128 coherent phase partitions across all 20 layers and all 1,555 gates with zero omissions), achieving global unitarity ||Ψ||^2_2 = 0.99998384, unbiased Porter-Thomas second moment ⟨u2⟩/⟨u⟩2 = 2.014852 (0.038% relative error from theoretical ideal), and Linear Cross-Entropy Benchmarking fidelity FXEB = 1.014852 (1.039321 under Born quantile sampling) in 372.34 seconds—consuming 204,383× less energy and 338,540× less memory than Sunway; and (3) Supremacy Threshold Crossing in 2.96 seconds (Slice 1, FXEB = 0.0235 > 0.002) alongside 783 ms Paraconsistent L4 Thermodynamic Screening at the 95% Overlap Gap Property (OGP) barrier.
State-of-the-art quantum circuit simulators have mostly focused on scaling the number of qubits. However, we argue that studying current noisy quantum computers and variational quantum algorithms benefits from high-throughput simulation of intermediate-scale quantum circuits. We present the first implementation and evaluation of a batched quantum simulator on the NEC Vector Engine (VE), a vector processor with massive memory bandwidth that is ideal for memory-intensive state vector simulation. To take advantage of the long-vector architecture of VE, we design a parallelization strategy and memory layout suited for batched state vector simulation. Our evaluation shows that the latest-generation VE Type 30A achieves equal performance as the NVIDIA A100 80 GB model for single-qubit gates, but slightly underperforms A100 for two-qubit gates. Furthermore, our vector-friendly noise gate implementation offers a large speedup compared to that of cuStateVec. When simulating a noisy 16-qubit random circuit with our simulator, VE30 is up to 1.95 × faster than A100.