Скачать или смотреть Quantum Photonic Entropy and Exascale Turbulence

NotebookLM: "...extreme-scale simulations of turbulence on the world’s first exascale supercomputer, Frontier. By utilizing a 32,768³ periodic grid, the authors successfully modeled the complex physics of small-scale turbulence at an unprecedentedly high Taylor-scale Reynolds number of 2500. To overcome the immense computational costs of such a high-fidelity model, the study employs the Multi-Resolution Independent Simulation (MRIS) technique, which uses ensemble averaging over short, high-resolution segments rather than continuous long-term simulations. The findings confirm that classical scaling laws and the zeroth law of turbulence remain valid at these extreme scales, while providing new data on the intermittency and probability distributions of energy dissipation and enstrophy. Ultimately, the work demonstrates that modern supercomputing can resolve fundamental questions about the dissipation range that were previously beyond reach."

"... entropy computing [is] a novel computational framework designed to solve complex NP-hard optimization problems by leveraging the physics of open photonic systems. The authors present a hybrid optoelectronic hardware called Dirac-3, which utilizes temporal photonic modes and a measurement-feedback loop to navigate difficult energy landscapes and avoid local minima. By mapping target Hamiltonians onto an effective dissipative operator, the system uses quantum fluctuations from single-photon detection as a source of "entropy" to stochastically search for global optima. Unlike traditional solvers, this architecture naturally supports higher-order interaction terms and continuous variables, demonstrating superior performance in benchmarks such as non-convex quadratic programming and combinatorial graph cuts. Ultimately, the study positions entropy computing as a scalable and energy-efficient paradigm that bridges the gap between classical CMOS devices and fully realized quantum computers."

https://www.cambridge.org/core/servic...
https://doi.org/10.1038/s42005-025-02...