Computer Physics Communications最新文献

{"title":"Improvements in charged lepton and photon propagation for the software PROPOSAL","authors":"Jean-Marco Alameddine ,&nbsp;Johannes Albrecht ,&nbsp;Hans Dembinski ,&nbsp;Pascal Gutjahr ,&nbsp;Karl-Heinz Kampert ,&nbsp;Wolfgang Rhode ,&nbsp;Maximilian Sackel ,&nbsp;Alexander Sandrock ,&nbsp;Jan Soedingrekso","doi":"10.1016/j.cpc.2024.109243","DOIUrl":"10.1016/j.cpc.2024.109243","url":null,"abstract":"<div><p>Accurate particle simulations are essential for the next generation of experiments in astroparticle physics. The Monte Carlo simulation library PROPOSAL is a flexible tool to efficiently propagate high-energy leptons and photons through large volumes of media, for example in the context of underground observatories. It is written as a <span>C++</span> library, including a Python interface. In this paper, the most recent updates of PROPOSAL are described, including the addition of electron, positron, and photon propagation, for which new interaction types have been implemented. This allows the usage of PROPOSAL to simulate electromagnetic particle cascades, for example in the context of air shower simulations. The precision of the propagation has been improved by including rare interaction processes, new photonuclear parametrizations, deflections in stochastic interactions, and the possibility of propagating in inhomogeneous density distributions. Additional technical improvements regarding the interpolation routine and the propagation algorithm are described.</p></div><div><h3>New version program summary</h3><p><em>Program Title:</em> PROPOSAL.</p><p><em>CPC Library link to program files:</em> <span>https://doi.org/10.17632/g478pjdcxy.2</span><svg><path></path></svg>.</p><p><em>Developer's repository link:</em> <span>https://github.com/tudo-astroparticlephysics/PROPOSAL</span><svg><path></path></svg>.</p><p><em>Licensing provisions:</em> LGPL.</p><p><em>Programming language:</em> C++, Python.</p><p><em>Journal reference of previous version:</em> Comput. Phys. Commun. 242 (2019) 132.</p><p><em>Does the new version supersede the previous version?:</em> Yes.</p><p><em>Reasons for the new version:</em> Substantial addition of features. Various bugfixes.</p><p><em>Summary of revisions:</em> The library now also treats photons and has the corresponding processes implemented. New parametrizations for photonuclear interaction have been implemented. The angular deflection in stochastic energy losses has been implemented in addition to the already existing multiple scattering implementation, which has been improved to reduce the runtime. The implementation of the Landau-Pomeranchuk-Migdal effect has been corrected. The propagation algorithm has been improved, including the support of inhomogeneous density distributions.</p><p><em>Nature of problem:</em> Three-dimensional propagation of charged leptons and photons through different media. Particles lose energy stochastically by ionization, bremsstrahlung, pair production, and photonuclear interaction for charged leptons (including annihilation with atomic electrons for positrons) and Compton scattering, pair production, photoelectric effect and photohadronic interaction for photons. Additionally, they are deflected while propagating through the medium due to both multiple elastic Coulomb scattering as well as deflections in individual stochastic interactions. Unstable particles eventually decay, pro","PeriodicalId":285,"journal":{"name":"Computer Physics Communications","volume":null,"pages":null},"PeriodicalIF":6.3,"publicationDate":"2024-05-15","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://www.sciencedirect.com/science/article/pii/S0010465524001668/pdfft?md5=8a6f84ccba67ed031fe5ffc276e87899&pid=1-s2.0-S0010465524001668-main.pdf","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"141025818","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"物理与天体物理","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"FlexibleSUSY extended to automatically compute physical quantities in any beyond the standard model theory: Charged lepton flavor violation processes, Higgs decays, and user-defined observables","authors":"Uladzimir Khasianevich ,&nbsp;Wojciech Kotlarski ,&nbsp;Dominik Stöckinger ,&nbsp;Alexander Voigt","doi":"10.1016/j.cpc.2024.109244","DOIUrl":"https://doi.org/10.1016/j.cpc.2024.109244","url":null,"abstract":"<div><p><span>FlexibleSUSY</span> is a framework for the automated computation of physical quantities (observables) in models beyond the Standard Model (BSM). This paper describes an extension of <span>FlexibleSUSY</span> which allows to define and add new observables that can be enabled and computed in applicable user-defined BSM models. The extension has already been used to include Charged Lepton Flavor Violation (CLFV) observables, but further observables can now be added straightforwardly. The paper is split into two parts. The first part is non-technical and describes from the user's perspective how to enable the calculation of predefined observables, in particular CLFV observables. The second part of the paper explains how to define new observables such that their automatic computation in any applicable BSM model becomes possible. A key ingredient is the new <span>NPointFunctions</span> extension which allows to use tree-level and loop calculations in the model-independent setup of observables. Three examples of increasing complexity are fully worked out. This illustrates the features and provides code snippets that may be used as a starting point for implementation of further observables.</p></div><div><h3>Program summary</h3><p><em>Program title:</em> <span>NPointFunctions</span></p><p><em>CPC Library link to program files:</em> <span>https://doi.org/10.17632/kf7m8gn8vp.2</span><svg><path></path></svg></p><p><em>Developer's repository link:</em> <span>https://github.com/FlexibleSUSY/FlexibleSUSY</span><svg><path></path></svg></p><p><em>Licensing provisions:</em> GPLv3</p><p><em>Programming language:</em> <span>C++</span>, <span>Wolfram Language</span>, <span>Fortran</span>, <span>Bourne shell</span></p><p><em>Journal reference of previous version::</em> Comput. Phys. Commun. 230 (2018) 145–217; PoS CompTools2021 (2022) 036</p><p><em>Does the new version supersede the previous version?:</em> Yes</p><p><em>Reasons for the new version:</em> Program extension including new observables and file structures</p><p><em>Nature of problem:</em> Determining observables for an arbitrary extension of the Standard Model supported by <span>FlexibleSUSY</span>, input by the user.</p><p><em>Solution method:</em> Generation of the code from automated algebraic manipulations. Automatic filling and compiling of predefined template files.</p><p><em>Additional comments including restrictions and unusual features:</em> Vertices with a direct product of Lorentz and color structures are supported. Settings of the advanced <span>NPointFunctions</span> mode rely on explicit specification of topologies.</p></div>","PeriodicalId":285,"journal":{"name":"Computer Physics Communications","volume":null,"pages":null},"PeriodicalIF":6.3,"publicationDate":"2024-05-15","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://www.sciencedirect.com/science/article/pii/S001046552400167X/pdfft?md5=d90f11286a31b2404401b33e36a8386d&pid=1-s2.0-S001046552400167X-main.pdf","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"141077923","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"物理与天体物理","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Method for scalable and performant GPU-accelerated simulation of multiphase compressible flow","authors":"Anand Radhakrishnan ,&nbsp;Henry Le Berre ,&nbsp;Benjamin Wilfong ,&nbsp;Jean-Sebastien Spratt ,&nbsp;Mauro Rodriguez Jr. ,&nbsp;Tim Colonius ,&nbsp;Spencer H. Bryngelson","doi":"10.1016/j.cpc.2024.109238","DOIUrl":"https://doi.org/10.1016/j.cpc.2024.109238","url":null,"abstract":"<div><p>Multiphase compressible flows are often characterized by a broad range of space and time scales, entailing large grids and small time steps. Simulations of these flows on CPU-based clusters can thus take several wall-clock days. Offloading the compute kernels to GPUs appears attractive but is memory-bound for many finite-volume and -difference methods, damping speedups. Even when realized, GPU-based kernels lead to more intrusive communication and I/O times owing to lower computation costs. We present a strategy for GPU acceleration of multiphase compressible flow solvers that addresses these challenges and obtains large speedups at scale. We use OpenACC for directive-based offloading of all compute kernels while maintaining low-level control when needed. An established Fortran preprocessor and metaprogramming tool, Fypp, enables otherwise hidden compile-time optimizations. This strategy exposes compile-time optimizations and high memory reuse while retaining readable, maintainable, and compact code. Remote direct memory access realized via CUDA-aware MPI and GPUDirect reduces halo-exchange communication time. We implement this approach in the open-source solver MFC <span>[1]</span>. Metaprogramming results in an 8-times speedup of the most expensive kernels compared to a statically compiled program, reaching 46% of peak FLOPs on modern NVIDIA GPUs and high arithmetic intensity (about 10 FLOPs/byte). In representative simulations, a single NVIDIA A100 GPU is 7-times faster compared to an Intel Xeon Cascade Lake (6248) CPU die, or about 300-times faster compared to a single such CPU core. At the same time, near-ideal (97%) weak scaling is observed for at least 13824 GPUs on OLCF Summit. A strong scaling efficiency of 84% is retained for an 8-times increase in GPU count. Collective I/O, implemented via MPI3, helps ensure the negligible contribution of data transfers (<span><math><mo>&lt;</mo><mn>1</mn><mtext>%</mtext></math></span> of the wall time for a typical, large simulation). Large many-GPU simulations of compressible (solid-)liquid-gas flows demonstrate the practical utility of this strategy.</p></div>","PeriodicalId":285,"journal":{"name":"Computer Physics Communications","volume":null,"pages":null},"PeriodicalIF":6.3,"publicationDate":"2024-05-13","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"140951033","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"物理与天体物理","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Optimized thread-block arrangement in a GPU implementation of a linear solver for atmospheric chemistry mechanisms","authors":"Christian Guzman Ruiz ,&nbsp;Mario Acosta ,&nbsp;Oriol Jorba ,&nbsp;Eduardo Cesar Galobardes ,&nbsp;Matthew Dawson ,&nbsp;Guillermo Oyarzun ,&nbsp;Carlos Pérez García-Pando ,&nbsp;Kim Serradell","doi":"10.1016/j.cpc.2024.109240","DOIUrl":"10.1016/j.cpc.2024.109240","url":null,"abstract":"<div><p>Earth system models (ESM) demand significant hardware resources and energy consumption to solve atmospheric chemistry processes. Recent studies have shown improved performance from running these models on GPU accelerators. Nonetheless, there is room for improvement in exploiting even more GPU resources.</p><p>This study proposes an optimized distribution of the chemical solver's computational load on the GPU, named Block-cells. Additionally, we evaluate different configurations for distributing the computational load in an NVIDIA GPU.</p><p>We use the linear solver from the Chemistry Across Multiple Phases (CAMP) framework as our test bed. An intermediate-complexity chemical mechanism under typical atmospheric conditions is used. Results demonstrate a 35× speedup compared to the single-CPU thread reference case. Even using the full resources of the node (40 physical cores) on the reference case, the Block-cells version outperforms them by 50%. The Block-cells approach shows promise in alleviating the computational burden of chemical solvers on GPU architectures.</p></div>","PeriodicalId":285,"journal":{"name":"Computer Physics Communications","volume":null,"pages":null},"PeriodicalIF":6.3,"publicationDate":"2024-05-13","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"141032446","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"物理与天体物理","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"RHDLPP: A multigroup radiation hydrodynamics code for laser-produced plasmas","authors":"Qi Min ,&nbsp;Ziyang Xu ,&nbsp;Siqi He ,&nbsp;Haidong Lu ,&nbsp;Xingbang Liu ,&nbsp;Ruizi Shen ,&nbsp;Yanhong Wu ,&nbsp;Qikun Pan ,&nbsp;Chongxiao Zhao ,&nbsp;Fei Chen ,&nbsp;Maogen Su ,&nbsp;Chenzhong Dong","doi":"10.1016/j.cpc.2024.109242","DOIUrl":"https://doi.org/10.1016/j.cpc.2024.109242","url":null,"abstract":"<div><p>In this paper, we introduce the RHDLPP, a flux-limited multigroup radiation hydrodynamics numerical code designed for simulating laser-produced plasmas in diverse environments. The code bifurcates into two packages: RHDLPP-LTP for low-temperature plasmas generated by moderate-intensity nanosecond lasers, and RHDLPP-HTP for high-temperature, high-density plasmas formed by high-intensity laser pulses. The core radiation hydrodynamic equations are resolved in the Eulerian frame, employing an operator-split method. This method decomposes the solution into two substeps: first, the explicit resolution of the hyperbolic subsystems integrating radiation and fluid dynamics; second, the implicit treatment of the parabolic part comprising stiff radiation diffusion, heat conduction, and energy exchange. Laser propagation and energy deposition are modeled through a hybrid approach, combining geometrical-optics ray-tracing in sub-critical plasma regions with a one-dimensional solution of the Helmholtz wave equation in super-critical areas. The thermodynamic states are ascertained using an equation of state, based on either the real gas approximation or the quotidian equation of state (QEOS). For ionization calculations, the code employs a steady-state collisional-radiation (CR) model using the screened-hydrogenic approximation. Additionally, RHDLPP includes RHDLPP-SpeIma3D, a three-dimensional spectral simulation post-processing module, for generating both temporally-spatially resolved and time-integrated spectra and imaging, facilitating direct comparisons with experimental data. The paper showcases a series of verification tests to establish the code's accuracy and efficiency, followed by application cases, including simulations of laser-produced aluminium (Al) plasmas, pre-pulse-induced target deformation of tin (Sn) microdroplets relevant to extreme ultraviolet lithography light sources, and varied imaging and spectroscopic simulations. These simulations highlight RHDLPP's effectiveness and applicability in fields such as laser-induced breakdown spectroscopy, extreme ultraviolet lithography sources, and high-energy-density physics.</p></div>","PeriodicalId":285,"journal":{"name":"Computer Physics Communications","volume":null,"pages":null},"PeriodicalIF":6.3,"publicationDate":"2024-05-11","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"141068718","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"物理与天体物理","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Direct/split invariant-preserving Fourier pseudo-spectral methods for the rotation-two-component Camassa–Holm system with peakon solitons","authors":"Qifeng Zhang,&nbsp;Tong Yan,&nbsp;Dinghua Xu,&nbsp;Yong Chen","doi":"10.1016/j.cpc.2024.109237","DOIUrl":"10.1016/j.cpc.2024.109237","url":null,"abstract":"<div><p>The Fourier pseudo-spectral method is well suited to solve PDEs under the periodic boundary condition due to its high-order accuracy and easy-to-implement feature. In this paper, we explore as well as comparatively study four classes of Fourier pseudo-spectral schemes for solving the rotation-two-component Camassa–Holm system which possibly owns peakon solitons. Via exploiting inherent structural properties of the system, we reformulate it into two kinds of different equivalent forms and then apply the Fourier pseudo-spectral method to derive two spatial semi-discrete systems, both of which are proved to preserve the corresponding invariants including mass, momentum and energy. Subsequently, we construct two linearly implicit schemes based on Strang splitting technique and two nonlinear schemes, respectively, for both semi-discrete systems. Owing to the different equivalent forms in the structure, one of the nonlinear schemes preserves discrete mass and momentum, while the other one is shown to preserve all three invariants. Numerical results under the situation of smooth/nonsmooth initial values are provided for distinct types of solutions to test the accuracy in long time simulation and to verify the capacity of predicting water wave propagation, as well as advantages in preserving these invariants. For instance, the present schemes are shown to be at least 14 significant digits, improving upon 10 from ones in previous references.</p></div>","PeriodicalId":285,"journal":{"name":"Computer Physics Communications","volume":null,"pages":null},"PeriodicalIF":6.3,"publicationDate":"2024-05-10","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"141028116","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"物理与天体物理","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"A new high-order shock-capturing TENO scheme combined with skew-symmetric-splitting method for compressible gas dynamics and turbulence simulation","authors":"Tian Liang ,&nbsp;Lin Fu","doi":"10.1016/j.cpc.2024.109236","DOIUrl":"10.1016/j.cpc.2024.109236","url":null,"abstract":"<div><p>The high-order shock-capturing scheme is one of the main building blocks for the simulation of the compressible fluid characterized by strong shockwaves and broadband length scales. However, the classical shock-capturing scheme fails to perform long-time stable and non-dissipative simulations since the quadratic invariants associated with the conservation equations cannot be conserved as a result of the inherent numerical dissipation. Additionally, the overall computational cost for classical shock-capturing schemes is quite expensive as a result of the time-consuming local characteristic decomposition and the nonlinear-weights computing process. In this work, based on a new efficient discontinuity indicator, which distinguishes the non-smooth high-wavenumber fluctuations and discontinuities from smooth scales in the wavenumber space, a paradigm of high-order shock-capturing scheme by recasting the non-dissipative skew-symmetric-splitting method with newly optimized dispersion property for smooth flow scales and invoking the nonlinear targeted ENO (TENO) schemes for non-smooth discontinuities is proposed. The resulting TENO-S scheme not only successfully performs long-time stable computations for smooth flows without numerical dissipation, but also recovers the robust shock-capturing capabilities with adaptive numerical dissipation. Without the necessity of parameter tuning case by case, extensive benchmark simulations involving a wide range of flow length scales and strong discontinuities demonstrate that the proposed TENO-S scheme performs significantly better than the straightforward deployment of WENO/TENO-family schemes with better spectral property and higher computational efficiency.</p></div>","PeriodicalId":285,"journal":{"name":"Computer Physics Communications","volume":null,"pages":null},"PeriodicalIF":6.3,"publicationDate":"2024-05-10","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"141040048","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"物理与天体物理","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"The TDHF code Sky3D version 1.2","authors":"Abhishek ,&nbsp;Paul Stevenson ,&nbsp;Yue Shi ,&nbsp;Esra Yüksel ,&nbsp;A.S. Umar","doi":"10.1016/j.cpc.2024.109239","DOIUrl":"https://doi.org/10.1016/j.cpc.2024.109239","url":null,"abstract":"<div><p>The Sky3D code has been widely used to describe nuclear ground states, collective vibrational excitations, and heavy-ion collisions. The approach is based on Skyrme forces or related energy density functionals. The static and dynamic equations are solved on a three-dimensional grid, and pairing is been implemented in the BCS approximation. This updated version of the code aims to facilitate the calculation of nuclear strength functions in the regime of linear response theory, while retaining all existing functionality and use cases. The strength functions are benchmarked against available RPA codes, and the user has the freedom of choice when selecting the nature of external excitation (from monopole to hexadecapole and more). Some utility programs are also provided that calculate the strength function from the time-dependent output of the dynamic calculations of the Sky3D code.</p></div><div><h3>New version program summary</h3><p><em>Program Title:</em> Sky3D</p><p><em>CPC Library link to program files:</em> <span>https://doi.org/10.17632/vzbrzvyrn4.2</span><svg><path></path></svg></p><p><em>Developer's repository link:</em> <span>https://github.com/manybody/sky3d</span><svg><path></path></svg></p><p><em>Licensing provisions:</em> GPLv3</p><p><em>Programming language:</em> Fortran, with one post-processing utility in Python.</p><p><em>Journal reference of previous version::</em> Schuetrumpf, B., Reinhard, P.G., Stevenson, P.D., Umar, A.S., and Maruhn, J.A. (2018). The TDHF code Sky3D version 1.1. <span>Comput. Phys. Commun. 229 (2018) 211–213.</span><svg><path></path></svg></p><p><em>Does the new version supersede the previous version?:</em> Yes.</p><p><em>Reasons for the new version:</em> The capability of reproducing the nuclear strength function for a variety of newly-coded external boosts has been added.</p><p><em>Nature of problem:</em> Calculating nuclear multipole strength functions is a crucial probe that can help model the nuclear system and its structure properties. A variety of models exist for this task, such as QRPA (Quasiparticle Random Phase Approximation) and its variants, but such approaches are often limited due to symmetry constraints. Time-dependent Hartree Fock (TDHF) has been used to simulate nuclear vibrations and collisions between nuclei for low energies without assuming any symmetry in the system. This code extends the TDHF to calculate the multi-pole strength functions of atomic nuclei. We showcase its reliability by comparing it with the established RPA codes for the calculation of such strength functions.</p><p><em>Solution method:</em> We extended previous versions of the Sky3D code [1,2] to include an external boost of multipole type where the user can provide custom input that decides the nature of the multipole (monopole, quadrupole, octupole, and so on) boost. The principal aim is to calculate the multipole strength function, which is the Fourier transform of the time-dependent expectation value of the m","PeriodicalId":285,"journal":{"name":"Computer Physics Communications","volume":null,"pages":null},"PeriodicalIF":6.3,"publicationDate":"2024-05-10","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://www.sciencedirect.com/science/article/pii/S0010465524001620/pdfft?md5=4a533df4e435e10a51c66bd31cf8a2bd&pid=1-s2.0-S0010465524001620-main.pdf","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"140918547","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"物理与天体物理","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"A differentiable programming framework for spin models","authors":"Tiago S. Farias ,&nbsp;Vitor V. Schultz ,&nbsp;José C.M. Mombach ,&nbsp;Jonas Maziero","doi":"10.1016/j.cpc.2024.109234","DOIUrl":"10.1016/j.cpc.2024.109234","url":null,"abstract":"<div><p>We introduce a novel framework for simulating spin models using differentiable programming, an approach that leverages the advancements in machine learning and computational efficiency. We focus on three distinct spin systems: the Ising model, the Potts model, and the Cellular Potts model, demonstrating the practicality and scalability of our framework in modeling these complex systems. Additionally, this framework allows for the optimization of spin models, which can adjust the parameters of a system by a defined objective function. In order to simulate these models, we adapt the Metropolis-Hastings algorithm to a differentiable programming paradigm, employing batched tensors for simulating spin lattices. This adaptation not only facilitates the integration with existing deep learning tools but also significantly enhances computational speed through parallel processing capabilities, as it can be implemented on different hardware architectures, including GPUs and TPUs.</p></div>","PeriodicalId":285,"journal":{"name":"Computer Physics Communications","volume":null,"pages":null},"PeriodicalIF":6.3,"publicationDate":"2024-05-10","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"141047424","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"物理与天体物理","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Fast-QSGS: A GPU accelerated program for structure generation of granular disordered media","authors":"Guang Yang,&nbsp;Tong Liu,&nbsp;Xukang Lu,&nbsp;Moran Wang","doi":"10.1016/j.cpc.2024.109241","DOIUrl":"10.1016/j.cpc.2024.109241","url":null,"abstract":"<div><p>We present Fast-QSGS, a GPU-accelerated program for granular disordered media generation. Based on vectorization, Fast-QSGS is accelerated by modern GPU thanks to the NumPy-compatible API provided by CuPy. We also introduce a variable growth probability function and seed spacing control to improve the speed and accuracy of the original QSGS method. Computational performance benchmarks are conducted on both consumer-grade and professional-grade GPUs. Generation of disordered media of size 400³ can be completed in 30 s on A100 and 110 s on RTX4060, achieving a speedup of over 400 compared with the serial version. Physical benchmarks on the reconstruction of Fontainebleau sandstone and hydrated cement are conducted. Our results demonstrate that the permeability of the reconstructed Fontainebleau sandstone falls within the range of experimental values. Additionally, the average relative error of the volume fraction of the unhydrated cement and capillary porosity of hydrated cement is 1.9 % and 3.4 % compared with Powers’ law, respectively.</p></div>","PeriodicalId":285,"journal":{"name":"Computer Physics Communications","volume":null,"pages":null},"PeriodicalIF":6.3,"publicationDate":"2024-05-09","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"141031119","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"物理与天体物理","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}