Pub Date : 2025-11-14DOI: 10.1016/j.cpc.2025.109946
Xin Li , Qinglin Tang , Xuanxuan Zhou
In this paper, a new class of high-order structure-preserving algorithms are developed for the dynamics of general nonlinear Schrödinger/Gross-Pitaevskii equations. We present an improved Lagrange multiplier method which provides a more relaxed environment for the development and implementation of numerical schemes. The optimization strategy facilitates the development of stable schemes and the whole method consists of two parts: first, the application of a high-order semi-implicit scheme to discretize the original model as a prediction, and second, the utilization of a similar scheme to approximate the optimization model as a correction. In addition, we present a modification that reduces the computational complexity and cost of the numerical implementation as well. Numerical results and some comparisons are provided to demonstrate the novelty of methods for simulating the equations with one or multiple components. The applications in Bose-Einstein condensation with various rotation speeds and strongly repulsive interaction indicate that our methods are efficient, accurate and robust.
{"title":"On a class of high-order structure-preserving methods for the dynamics of nonlinear Schrödinger/Gross-Pitaevskii equations","authors":"Xin Li , Qinglin Tang , Xuanxuan Zhou","doi":"10.1016/j.cpc.2025.109946","DOIUrl":"10.1016/j.cpc.2025.109946","url":null,"abstract":"<div><div>In this paper, a new class of high-order structure-preserving algorithms are developed for the dynamics of general nonlinear Schrödinger/Gross-Pitaevskii equations. We present an improved Lagrange multiplier method which provides a more relaxed environment for the development and implementation of numerical schemes. The optimization strategy facilitates the development of stable schemes and the whole method consists of two parts: first, the application of a high-order semi-implicit scheme to discretize the original model as a prediction, and second, the utilization of a similar scheme to approximate the optimization model as a correction. In addition, we present a modification that reduces the computational complexity and cost of the numerical implementation as well. Numerical results and some comparisons are provided to demonstrate the novelty of methods for simulating the equations with one or multiple components. The applications in Bose-Einstein condensation with various rotation speeds and strongly repulsive interaction indicate that our methods are efficient, accurate and robust.</div></div>","PeriodicalId":285,"journal":{"name":"Computer Physics Communications","volume":"319 ","pages":"Article 109946"},"PeriodicalIF":3.4,"publicationDate":"2025-11-14","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145576744","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}
Pub Date : 2025-11-14DOI: 10.1016/j.cpc.2025.109930
Hugo Gabrielidis , Filippo Gatti , Stéphane Vialle
In this study, we develop a Diffusion Transformer (referred as to DiT1D) for synthesizing realistic earthquake time histories. The DiT1D generates realistic broadband accelerograms (0–30Hz resolution), constrained at low frequency by 3-dimensional (3D) elastodynamics numerical simulations, ensuring the fulfillment of the minimum observable physics. The DiT1D architecture, successfully adopted in super-resolution image generation, is trained on recorded single-station 3-components (3C) accelerograms. Thanks to Multi-Head Cross-Attention (MHCA) layers, we guide the DiT1D inference by enforcing the low-frequency part of the accelerogram spectrum into it. The DiT1D learns the low-to-high frequency map from the recorded accelerograms, duly normalized, and successfully transfer it to synthetic time histories. The latter are low-frequency by nature, because of the lack of knowledge on the underground structure of the Earth, demanded to fully calibrate the numerical model. We developed a CNN-LSTM lightweight network in conjunction with the DiT1D, so to predict the peak amplitude of the broadband signal from its low-pass-filtered counterpart, and rescale the normalized accelerograms rendered by the DiT1D. Despite the DiT1D being agnostic to any earthquake event peculiarities (magnitude, site conditions, etc.), it showcases remarkable zero-shot prediction realism when applied to the output of validated earthquake simulations. The generated time histories are viable input accelerograms for earthquake-resistant structural design and the pre-trained DiT1D holds a huge potential to integrate full-scale fault-to-structure digital twins of earthquake-prone regions. The pretrained DiT1D is available at https://github.com/HugoGabrielidis16/Seismic_DiT1D.
{"title":"Physics-based super-resolved simulation of 3D elastic wave propagation adopting scalable diffusion transformer","authors":"Hugo Gabrielidis , Filippo Gatti , Stéphane Vialle","doi":"10.1016/j.cpc.2025.109930","DOIUrl":"10.1016/j.cpc.2025.109930","url":null,"abstract":"<div><div>In this study, we develop a Diffusion Transformer (referred as to DiT1D) for synthesizing realistic earthquake time histories. The DiT1D generates realistic broadband accelerograms (0–30Hz resolution), constrained at low frequency by 3-dimensional (3D) elastodynamics numerical simulations, ensuring the fulfillment of the minimum observable physics. The DiT1D architecture, successfully adopted in super-resolution image generation, is trained on recorded single-station 3-components (3C) accelerograms. Thanks to Multi-Head Cross-Attention (MHCA) layers, we guide the DiT1D inference by enforcing the low-frequency part of the accelerogram spectrum into it. The DiT1D learns the low-to-high frequency map from the recorded accelerograms, duly normalized, and successfully transfer it to synthetic time histories. The latter are low-frequency by nature, because of the lack of knowledge on the underground structure of the Earth, demanded to fully calibrate the numerical model. We developed a CNN-LSTM lightweight network in conjunction with the DiT1D, so to predict the peak amplitude of the broadband signal from its low-pass-filtered counterpart, and rescale the normalized accelerograms rendered by the DiT1D. Despite the DiT1D being agnostic to any earthquake event peculiarities (magnitude, site conditions, etc.), it showcases remarkable zero-shot prediction realism when applied to the output of validated earthquake simulations. The generated time histories are viable input accelerograms for earthquake-resistant structural design and the pre-trained DiT1D holds a huge potential to integrate full-scale fault-to-structure digital twins of earthquake-prone regions. The pretrained DiT1D is available at <span><span>https://github.com/HugoGabrielidis16/Seismic_DiT1D</span><svg><path></path></svg></span>.</div></div>","PeriodicalId":285,"journal":{"name":"Computer Physics Communications","volume":"320 ","pages":"Article 109930"},"PeriodicalIF":3.4,"publicationDate":"2025-11-14","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145616213","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}
Pub Date : 2025-11-14DOI: 10.1016/j.cpc.2025.109933
Joshua Davies , Kay Schönwald , Matthias Steinhauser , Daniel Stremmer
We present the library ggxy, written in C++, which can be used to compute partonic and hadronic cross sections for gluon-induced processes with at least one closed heavy quark loop. It is based on analytic ingredients which avoids, to a large extent, expensive numerical integration. This results in significantly shorter run-times than other similar tools. Modifying input parameters, changing the renormalization scheme and varying renormalization and factorization scales is straightforward. In Version 1 of ggxy we implement all routines which are needed to compute partonic and hadronic cross sections for Higgs boson pair production up to next-to-leading order in QCD. We provide flexible interfaces and allow the user to interact with the built-in amplitudes at various levels. PROGRAM SUMMARYProgram title:ggxyDeveloper’s repository link:https://gitlab.com/ggxy/ggxy-releaseLicensing provisions: GNU General Public License Version 3 Programming language: C++ and Fortran External routines/libraries used:avhlib, boost, Collier, CuTtools, eigen, LHAPDF, lievaluate, OneLOop, Recola, CRunDecNature of problem: The computation of partonic and hadronic cross sections for gluon-induced processes. In Version 1, the Higgs boson pair production process is implemented at next-to-leading order in Quantum Chromodynamics. Solution method: For the virtual corrections, deep expansions around the forward and high energy limit are used. Restrictions: The run-times depend crucially on the requested precision. Results at the per-mille level can be obtained in about 30 minutes using a single core on a AMD Ryzen Threadripper PRO 3955WX processor. References and Links: are provided in the paper
{"title":"ggxy: A flexible library to compute gluon-induced cross sections","authors":"Joshua Davies , Kay Schönwald , Matthias Steinhauser , Daniel Stremmer","doi":"10.1016/j.cpc.2025.109933","DOIUrl":"10.1016/j.cpc.2025.109933","url":null,"abstract":"<div><div>We present the library <span>ggxy</span>, written in <span>C++</span>, which can be used to compute partonic and hadronic cross sections for gluon-induced processes with at least one closed heavy quark loop. It is based on analytic ingredients which avoids, to a large extent, expensive numerical integration. This results in significantly shorter run-times than other similar tools. Modifying input parameters, changing the renormalization scheme and varying renormalization and factorization scales is straightforward. In Version 1 of <span>ggxy</span> we implement all routines which are needed to compute partonic and hadronic cross sections for Higgs boson pair production up to next-to-leading order in QCD. We provide flexible interfaces and allow the user to interact with the built-in amplitudes at various levels. <strong>PROGRAM SUMMARY</strong> <em>Program title:</em> <span>ggxy</span> <em>Developer’s repository link:</em> <span><span>https://gitlab.com/ggxy/ggxy-release</span><svg><path></path></svg></span> <em>Licensing provisions:</em> GNU General Public License Version 3 <em>Programming language:</em> C++ and Fortran <em>External routines/libraries used:</em> <span>avhlib</span>, <span>boost</span>, <span>Collier</span>, <span>CuTtools</span>, <span>eigen</span>, <span>LHAPDF</span>, <span>lievaluate</span>, <span>OneLOop</span>, <span>Recola</span>, <span>CRunDec</span> <em>Nature of problem:</em> The computation of partonic and hadronic cross sections for gluon-induced processes. In Version 1, the Higgs boson pair production process is implemented at next-to-leading order in Quantum Chromodynamics. <em>Solution method:</em> For the virtual corrections, deep expansions around the forward and high energy limit are used. <em>Restrictions:</em> The run-times depend crucially on the requested precision. Results at the per-mille level can be obtained in about 30 minutes using a single core on a AMD Ryzen Threadripper PRO 3955WX processor. <em>References and Links:</em> are provided in the paper</div></div>","PeriodicalId":285,"journal":{"name":"Computer Physics Communications","volume":"320 ","pages":"Article 109933"},"PeriodicalIF":3.4,"publicationDate":"2025-11-14","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145555178","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}
Pub Date : 2025-11-12DOI: 10.1016/j.cpc.2025.109952
Siarhei Dzianisau , Deokjung Lee
Deterministic neutron transport codes are a vital modern technology for high-fidelity reactor simulations. Many versions of such codes currently employ a method of characteristics (MOC) for solving the neutron transport equation. With the rise of graphics processing unit (GPU) computing power, offloading the most time-consuming parts of codes to GPU became possible. In this study, we present our novel GPU-enabled code, STREAM3D-GPU, with an MOC module offloaded to the GPU. It became possible due to a newly introduced, axially decomposed GPU-enabled three-dimensional (3D) MOC/diamond difference (DD) solver. This solver allowed us to reduce the GPU memory burden from 50 GB per MPI process to below 4.5 GB, thus enabling modern consumer-grade GPUs for large reactor calculations. We confirmed it for a 3D OPR-1000 quarter-core model with thermal-hydraulic feedback. We found that the axially decomposed GPU-enabled 3D MOC/DD solver in STREAM3D-GPU was faster by 22.1 times when using 8 GPU cards and by 40.5 times when using 16 GPU cards compared to a 128-core reference solution. Therefore, we estimated that one GPU card was equal to 324-353 parallel CPU cores. CPU nodes of comparable performance would cost 2.77-3.54 times more than a GPU system and consume 5.4-6.5 times more energy.
{"title":"Axially decomposed GPU-enabled three-dimensional method of characteristics/diamond difference solver in neutron transport code STREAM3D-GPU","authors":"Siarhei Dzianisau , Deokjung Lee","doi":"10.1016/j.cpc.2025.109952","DOIUrl":"10.1016/j.cpc.2025.109952","url":null,"abstract":"<div><div>Deterministic neutron transport codes are a vital modern technology for high-fidelity reactor simulations. Many versions of such codes currently employ a method of characteristics (MOC) for solving the neutron transport equation. With the rise of graphics processing unit (GPU) computing power, offloading the most time-consuming parts of codes to GPU became possible. In this study, we present our novel GPU-enabled code, STREAM3D-GPU, with an MOC module offloaded to the GPU. It became possible due to a newly introduced, axially decomposed GPU-enabled three-dimensional (3D) MOC/diamond difference (DD) solver. This solver allowed us to reduce the GPU memory burden from 50 GB per MPI process to below 4.5 GB, thus enabling modern consumer-grade GPUs for large reactor calculations. We confirmed it for a 3D OPR-1000 quarter-core model with thermal-hydraulic feedback. We found that the axially decomposed GPU-enabled 3D MOC/DD solver in STREAM3D-GPU was faster by 22.1 times when using 8 GPU cards and by 40.5 times when using 16 GPU cards compared to a 128-core reference solution. Therefore, we estimated that one GPU card was equal to 324-353 parallel CPU cores. CPU nodes of comparable performance would cost 2.77-3.54 times more than a GPU system and consume 5.4-6.5 times more energy.</div></div>","PeriodicalId":285,"journal":{"name":"Computer Physics Communications","volume":"319 ","pages":"Article 109952"},"PeriodicalIF":3.4,"publicationDate":"2025-11-12","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145576716","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}
Pub Date : 2025-11-11DOI: 10.1016/j.cpc.2025.109943
Liang Wan , Qingsong Mei , Haowen Liu , Huafeng Zhang , Jun-Ping Du , Shigenobu Ogata , Wen Tong Geng
Despite its widespread use in materials science, conventional molecular dynamics (MD) simulations are severely constrained by timescale limitations. To address this shortcoming, we propose an empirical formulation of accelerated MD method, adapted from a collective-variable-based extended system dynamics framework. While this framework is originally developed for efficient free energy sampling and reaction pathway determination of specific rare events in condensed matter, we have modified it to enable accelerated MD simulation and prediction of microstructure evolution of materials across a broad range of scenarios. In essence, the nearest neighbor off-centering absolute displacement (NNOAD), which quantifies the deviation of an atom from the geometric center of its nearest neighbors in materials, is introduced. We propose that the collection of NNOADs of all atoms can serve as a generalized reaction coordinate for various structural transitions in materials. The NNOAD of each atom, represented by its three components, is coupled with three additional dynamic variables assigned to the atom. Time evolution of the additional dynamic variables follows Langevin equation, while Nosé-Hoover dynamics is employed to thermostat the system. Through careful analysis and benchmark simulations, we established appropriate parameter ranges for the equations in our method. Application of this method to several test cases demonstrates its capability to accelerate MD simulations by several orders of magnitude while maintaining kinetic consistency and good accuracy in predicting long timescale microstructure evolutions of materials. We also provide some preliminary thoughts on theoretical justification of the method, offering insights into its underlying principles.
{"title":"An empirical formulation of accelerated molecular dynamics for simulating and predicting microstructure evolution in materials","authors":"Liang Wan , Qingsong Mei , Haowen Liu , Huafeng Zhang , Jun-Ping Du , Shigenobu Ogata , Wen Tong Geng","doi":"10.1016/j.cpc.2025.109943","DOIUrl":"10.1016/j.cpc.2025.109943","url":null,"abstract":"<div><div>Despite its widespread use in materials science, conventional molecular dynamics (MD) simulations are severely constrained by timescale limitations. To address this shortcoming, we propose an empirical formulation of accelerated MD method, adapted from a collective-variable-based extended system dynamics framework. While this framework is originally developed for efficient free energy sampling and reaction pathway determination of specific rare events in condensed matter, we have modified it to enable accelerated MD simulation and prediction of microstructure evolution of materials across a broad range of scenarios. In essence, the nearest neighbor off-centering absolute displacement (NNOAD), which quantifies the deviation of an atom from the geometric center of its nearest neighbors in materials, is introduced. We propose that the collection of NNOADs of all atoms can serve as a generalized reaction coordinate for various structural transitions in materials. The NNOAD of each atom, represented by its three components, is coupled with three additional dynamic variables assigned to the atom. Time evolution of the additional dynamic variables follows Langevin equation, while Nosé-Hoover dynamics is employed to thermostat the system. Through careful analysis and benchmark simulations, we established appropriate parameter ranges for the equations in our method. Application of this method to several test cases demonstrates its capability to accelerate MD simulations by several orders of magnitude while maintaining kinetic consistency and good accuracy in predicting long timescale microstructure evolutions of materials. We also provide some preliminary thoughts on theoretical justification of the method, offering insights into its underlying principles.</div></div>","PeriodicalId":285,"journal":{"name":"Computer Physics Communications","volume":"320 ","pages":"Article 109943"},"PeriodicalIF":3.4,"publicationDate":"2025-11-11","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145681708","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}
Pub Date : 2025-11-11DOI: 10.1016/j.cpc.2025.109944
Muhammad Hassan , Sunghyun Yoon , Yu Chen , Pilseok Kim , Hongryeol Yun , Hyuk Taek Kwon , Youn-Sang Bae , Chung-Yul Yoo , Dong-Yeun Koh , Chang Seop Hong , Ki Bong Lee , Yongchul G. Chung
Adsorption breakthrough modeling often requires complex software environments and scripting, limiting accessibility for many practitioners. We present AIM, a MATLAB-based graphical user interface (GUI) application that streamlines fixed-bed adsorption modeling and analysis through an integrated workflow, which includes isotherm fitting, enthalpy of adsorption estimation, mixture prediction, and multicomponent breakthrough simulations. AIM supports 13 isotherm models for isotherm fitting and includes Ideal Adsorbed Solution Theory (IAST) implementation (FastIAS) and extended Langmuir models for mixture isotherm predictions. Moreover, the isotherm models can be used to run non-isothermal breakthrough simulations along with isosteric enthalpies of adsorption from the Clausius-Clapeyron and Virial equations. Users can export detailed column and outlet profiles (e.g., composition, temperature) in multiple formats, enhancing reproducibility and data sharing among practitioners. We compared the breakthrough simulation results from AIM workflow with the experimental data in the literature for ternary gas mixture (CO2/H2/N2) and found excellent agreement for outlet compositions and temperature profiles.
{"title":"AIM: A user-friendly GUI workflow program for isotherm fitting, mixture prediction, isosteric heat of adsorption estimation, and breakthrough simulation","authors":"Muhammad Hassan , Sunghyun Yoon , Yu Chen , Pilseok Kim , Hongryeol Yun , Hyuk Taek Kwon , Youn-Sang Bae , Chung-Yul Yoo , Dong-Yeun Koh , Chang Seop Hong , Ki Bong Lee , Yongchul G. Chung","doi":"10.1016/j.cpc.2025.109944","DOIUrl":"10.1016/j.cpc.2025.109944","url":null,"abstract":"<div><div>Adsorption breakthrough modeling often requires complex software environments and scripting, limiting accessibility for many practitioners. We present AIM, a MATLAB-based graphical user interface (GUI) application that streamlines fixed-bed adsorption modeling and analysis through an integrated workflow, which includes isotherm fitting, enthalpy of adsorption estimation, mixture prediction, and multicomponent breakthrough simulations. AIM supports 13 isotherm models for isotherm fitting and includes Ideal Adsorbed Solution Theory (IAST) implementation (FastIAS) and extended Langmuir models for mixture isotherm predictions. Moreover, the isotherm models can be used to run non-isothermal breakthrough simulations along with isosteric enthalpies of adsorption from the Clausius-Clapeyron and Virial equations. Users can export detailed column and outlet profiles (e.g., composition, temperature) in multiple formats, enhancing reproducibility and data sharing among practitioners. We compared the breakthrough simulation results from AIM workflow with the experimental data in the literature for ternary gas mixture (CO<sub>2</sub>/H<sub>2</sub>/N<sub>2</sub>) and found excellent agreement for outlet compositions and temperature profiles.</div></div>","PeriodicalId":285,"journal":{"name":"Computer Physics Communications","volume":"319 ","pages":"Article 109944"},"PeriodicalIF":3.4,"publicationDate":"2025-11-11","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145576745","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}
Pub Date : 2025-11-11DOI: 10.1016/j.cpc.2025.109945
Johanna Langner, Isabelle Weber, Henryk A. Witek, Yuan-Pern Lee
SpectraMatcher is a cross-platform desktop application for interactive comparison of experimental and computed vibronic spectra, designed to assist in the recognition and assignment of spectral patterns. It provides an intuitive graphical interface — with no coding or scripting required — for importing experimental spectra, visualizing them alongside the corresponding theoretical spectra constructed from Gaussian frequency calculations, and adjusting key parameters such as peak width, intensity scaling factors, and vibration-type-specific anharmonic corrections. SpectraMatcher features an automated peak-matching algorithm that assigns experimental and computed peaks based on their intensity ratio and proximity. Assignments and spectra can be exported in multiple formats for publication or for further analysis. The software remains responsive even for large datasets, and supports efficient and reproducible interpretation of vibronic spectra.
{"title":"SpectraMatcher: A python program for interactive analysis and peak assignment of vibronic spectra","authors":"Johanna Langner, Isabelle Weber, Henryk A. Witek, Yuan-Pern Lee","doi":"10.1016/j.cpc.2025.109945","DOIUrl":"10.1016/j.cpc.2025.109945","url":null,"abstract":"<div><div><em>SpectraMatcher</em> is a cross-platform desktop application for interactive comparison of experimental and computed vibronic spectra, designed to assist in the recognition and assignment of spectral patterns. It provides an intuitive graphical interface — with no coding or scripting required — for importing experimental spectra, visualizing them alongside the corresponding theoretical spectra constructed from Gaussian frequency calculations, and adjusting key parameters such as peak width, intensity scaling factors, and vibration-type-specific anharmonic corrections. <em>SpectraMatcher</em> features an automated peak-matching algorithm that assigns experimental and computed peaks based on their intensity ratio and proximity. Assignments and spectra can be exported in multiple formats for publication or for further analysis. The software remains responsive even for large datasets, and supports efficient and reproducible interpretation of vibronic spectra.</div></div>","PeriodicalId":285,"journal":{"name":"Computer Physics Communications","volume":"320 ","pages":"Article 109945"},"PeriodicalIF":3.4,"publicationDate":"2025-11-11","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145616217","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}
Pub Date : 2025-11-10DOI: 10.1016/j.cpc.2025.109934
M. Dovale-Álvarez
<div><div>Deep Frequency Modulation Interferometry (DFMI) is an emerging laser interferometry technique for high-precision metrology, offering picometer-level displacement measurements and the potential for absolute length determination with sub-wavelength accuracy. However, the design and optimization of DFMI systems involve a complex interplay between interferometer physics, laser technology, multiple noise sources, and the choice of data processing algorithm. To address this, we present <span>DeepFMKit</span>, a new open-source Python library for the end-to-end simulation and analysis of DFMI systems. The framework features a high-fidelity physics engine that rigorously models key physical effects such as time-of-flight delays in dynamic interferometers, arbitrary laser modulation waveforms, and colored noise from user-defined 1/<em>f<sup>α</sup></em> spectral densities. This engine is coupled with a suite of interchangeable parameter estimation algorithms, including a highly-optimized, frequency-domain Non-linear Least Squares (NLS) for high-throughput batch processing of experimental data, and multiple time-domain Extended Kalman Filter (EKF) implementations for sample-by-sample state tracking, featuring both random walk and integrated random walk (constant velocity) process models. Furthermore, <span>DeepFMKit</span> includes a high-throughput experimentation framework for automating large-scale parameter sweeps and Monte Carlo analyses, enabling systematic characterization of system performance. <span>DeepFMKit</span>’s modular, object-oriented architecture facilitates the rapid configuration of virtual experiments, providing a powerful computational tool for researchers to prototype designs, investigate systematic errors, and accelerate the development of precision interferometry.</div><div><em>Program Title:</em> DeepFMKit</div><div><em>CPC Library link to program files:</em> (to be added by Technical Editor)</div><div><em>Developer’s repository link:</em> <span><span>github.com/mdovale/DeepFMKit</span><svg><path></path></svg></span></div><div><em>Licensing provisions:</em> BSD 3-clause</div><div><em>Programming language:</em> Python 3.9 and higher</div><div><em>Nature of problem:</em> Deep Frequency Modulation Interferometry (DFMI) is an increasingly important technique for high-precision optical metrology. The learning curve for modeling DFMI systems is very steep, as even a basic simulation requires implementing complex, non-linear physics, advanced digital signal processing for time-delays, and specialized parameter estimation algorithms. There are many researchers who must develop bespoke, time-consuming, and error-prone simulation toolchains by combining disparate libraries. The existing general-purpose optical design packages or specialized interferometry simulators are not tailored for DFMI’s unique signal generation and signal processing needs.</div><div><em>Solution method:</em> <span>DeepFMKit</span> serves as a complete compu
{"title":"The DeepFMKit python package: A toolbox for simulating and analyzing deep frequency modulation interferometers","authors":"M. Dovale-Álvarez","doi":"10.1016/j.cpc.2025.109934","DOIUrl":"10.1016/j.cpc.2025.109934","url":null,"abstract":"<div><div>Deep Frequency Modulation Interferometry (DFMI) is an emerging laser interferometry technique for high-precision metrology, offering picometer-level displacement measurements and the potential for absolute length determination with sub-wavelength accuracy. However, the design and optimization of DFMI systems involve a complex interplay between interferometer physics, laser technology, multiple noise sources, and the choice of data processing algorithm. To address this, we present <span>DeepFMKit</span>, a new open-source Python library for the end-to-end simulation and analysis of DFMI systems. The framework features a high-fidelity physics engine that rigorously models key physical effects such as time-of-flight delays in dynamic interferometers, arbitrary laser modulation waveforms, and colored noise from user-defined 1/<em>f<sup>α</sup></em> spectral densities. This engine is coupled with a suite of interchangeable parameter estimation algorithms, including a highly-optimized, frequency-domain Non-linear Least Squares (NLS) for high-throughput batch processing of experimental data, and multiple time-domain Extended Kalman Filter (EKF) implementations for sample-by-sample state tracking, featuring both random walk and integrated random walk (constant velocity) process models. Furthermore, <span>DeepFMKit</span> includes a high-throughput experimentation framework for automating large-scale parameter sweeps and Monte Carlo analyses, enabling systematic characterization of system performance. <span>DeepFMKit</span>’s modular, object-oriented architecture facilitates the rapid configuration of virtual experiments, providing a powerful computational tool for researchers to prototype designs, investigate systematic errors, and accelerate the development of precision interferometry.</div><div><em>Program Title:</em> DeepFMKit</div><div><em>CPC Library link to program files:</em> (to be added by Technical Editor)</div><div><em>Developer’s repository link:</em> <span><span>github.com/mdovale/DeepFMKit</span><svg><path></path></svg></span></div><div><em>Licensing provisions:</em> BSD 3-clause</div><div><em>Programming language:</em> Python 3.9 and higher</div><div><em>Nature of problem:</em> Deep Frequency Modulation Interferometry (DFMI) is an increasingly important technique for high-precision optical metrology. The learning curve for modeling DFMI systems is very steep, as even a basic simulation requires implementing complex, non-linear physics, advanced digital signal processing for time-delays, and specialized parameter estimation algorithms. There are many researchers who must develop bespoke, time-consuming, and error-prone simulation toolchains by combining disparate libraries. The existing general-purpose optical design packages or specialized interferometry simulators are not tailored for DFMI’s unique signal generation and signal processing needs.</div><div><em>Solution method:</em> <span>DeepFMKit</span> serves as a complete compu","PeriodicalId":285,"journal":{"name":"Computer Physics Communications","volume":"319 ","pages":"Article 109934"},"PeriodicalIF":3.4,"publicationDate":"2025-11-10","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145576711","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}
Pub Date : 2025-11-10DOI: 10.1016/j.cpc.2025.109936
Maksymilian Wojnar
Recent advances in generative neural networks, particularly flow matching (FM), have enabled the generation of high-fidelity samples while significantly reducing computational costs. A promising application of these models is accelerating simulations in high-energy physics (HEP), address research institutions meet their increasing computational demands. In this work, we leverage FM to develop surrogate models for fast simulations of zero degree calorimeters in the ALICE experiment. We present an effective training strategy that enables the training of fast generative models with an exceptionally small number of parameters. This approach achieves state-of-the-art simulation fidelity for both neutron (ZN) and proton (ZP) detectors, while offering substantial reductions in computational costs compared to existing methods. Our FM model achieves a Wasserstein distance of 1.27 for the ZN simulation with an inference time of 0.46 ms per sample, compared to the current best of 1.20 with an inference time of approximately 109 ms. The latent FM model further improves the inference speed, reducing the sampling time to 0.026 ms per sample, with a minimal trade-off in accuracy. Similarly, our approach achieves a Wasserstein distance of 1.30 for the ZP simulation, outperforming the current best of 2.08. The source code is available at https://github.com/m-wojnar/faster_zdc.
{"title":"Even faster simulations with flow matching: A study of zero degree calorimeter responses","authors":"Maksymilian Wojnar","doi":"10.1016/j.cpc.2025.109936","DOIUrl":"10.1016/j.cpc.2025.109936","url":null,"abstract":"<div><div>Recent advances in generative neural networks, particularly flow matching (FM), have enabled the generation of high-fidelity samples while significantly reducing computational costs. A promising application of these models is accelerating simulations in high-energy physics (HEP), address research institutions meet their increasing computational demands. In this work, we leverage FM to develop surrogate models for fast simulations of zero degree calorimeters in the ALICE experiment. We present an effective training strategy that enables the training of fast generative models with an exceptionally small number of parameters. This approach achieves state-of-the-art simulation fidelity for both neutron (ZN) and proton (ZP) detectors, while offering substantial reductions in computational costs compared to existing methods. Our FM model achieves a Wasserstein distance of 1.27 for the ZN simulation with an inference time of 0.46 ms per sample, compared to the current best of 1.20 with an inference time of approximately 109 ms. The latent FM model further improves the inference speed, reducing the sampling time to 0.026 ms per sample, with a minimal trade-off in accuracy. Similarly, our approach achieves a Wasserstein distance of 1.30 for the ZP simulation, outperforming the current best of 2.08. The source code is available at <span><span>https://github.com/m-wojnar/faster_zdc</span><svg><path></path></svg></span>.</div></div>","PeriodicalId":285,"journal":{"name":"Computer Physics Communications","volume":"319 ","pages":"Article 109936"},"PeriodicalIF":3.4,"publicationDate":"2025-11-10","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145576742","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}
Pub Date : 2025-11-10DOI: 10.1016/j.cpc.2025.109920
V.I. Korobov , J. Buša
A Fortran package SWEXPHC is presented, designed to calculate nonrelativistic energies for the bound-state three-body problem with Coulomb interaction. The implementation is based on MPFUN2020 package written by D.H. Bailey, which allows calculations with arbitrary precision, where the number of working digits can be adjusted by the user. The approximate wave function is chosen in the form of a variational exponential expansion, which has proven itself over many years as an effective method for obtaining highly accurate solutions for various three-particle systems such as the helium atom and/or the molecular hydrogen ion.
PROGRAM SUMMARY/NEW VERSION PROGRAM SUMMARY
Program Title: SWEXPHC
CPC Library link to program files: (to be added by Technical Editor)
Licensing provisions: GPLv3
Programming language: Fortran 90
Nature of problem(approx. 50–250 words):
The quantum nonrelativistic three-body problem with Coulomb interaction for bound states of arbitrary total orbital angular momenta L is solved.
Solution method(approx. 50–250 words): To solve the problem a variational method based on exponential expansion [1] is used. In the program we utilize the thread-safe multiprecision code [2] along with parallelization based on the OpenMP Fortran program inteface.
References
[1] V.I. Korobov, Coulomb three-body bound-state problem: Variational calculations of nonrelativistic energies. Phys. Rev. A 61, 064503 (2000).
[2] MPFUN2020: A thread-safe arbitrary precision package with special functions, D. Bailey, http://www.davidhbailey.com/dhbsoftware/.
{"title":"SWEXPHC: Variational bound state solution for the three-body nonrelativistic Schrödinger equation","authors":"V.I. Korobov , J. Buša","doi":"10.1016/j.cpc.2025.109920","DOIUrl":"10.1016/j.cpc.2025.109920","url":null,"abstract":"<div><div>A Fortran package SWEXPHC is presented, designed to calculate nonrelativistic energies for the bound-state three-body problem with Coulomb interaction. The implementation is based on MPFUN2020 package written by D.H. Bailey, which allows calculations with arbitrary precision, where the number of working digits can be adjusted by the user. The approximate wave function is chosen in the form of a variational exponential expansion, which has proven itself over many years as an effective method for obtaining highly accurate solutions for various three-particle systems such as the helium atom and/or the molecular hydrogen ion.</div><div><strong>PROGRAM SUMMARY/NEW VERSION PROGRAM SUMMARY</strong></div><div><em>Program Title: SWEXPHC</em></div><div><em>CPC Library link to program files:</em> (to be added by Technical Editor)</div><div><em>Licensing provisions:</em> GPLv3</div><div><em>Programming language:</em> Fortran 90</div><div><em>Nature of problem(approx. 50–250 words):</em></div><div>The quantum nonrelativistic three-body problem with Coulomb interaction for bound states of arbitrary total orbital angular momenta <em>L</em> is solved.</div><div><em>Solution method(approx. 50–250 words):</em> To solve the problem a variational method based on exponential expansion [1] is used. In the program we utilize the thread-safe multiprecision code [2] along with parallelization based on the OpenMP Fortran program inteface.</div><div><strong>References</strong></div><div>[1] V.I. Korobov, Coulomb three-body bound-state problem: Variational calculations of nonrelativistic energies. Phys. Rev. A <strong>61</strong>, 064503 (2000).</div><div>[2] MPFUN2020: A thread-safe arbitrary precision package with special functions, D. Bailey, <span><span>http://www.davidhbailey.com/dhbsoftware/</span><svg><path></path></svg></span>.</div></div>","PeriodicalId":285,"journal":{"name":"Computer Physics Communications","volume":"319 ","pages":"Article 109920"},"PeriodicalIF":3.4,"publicationDate":"2025-11-10","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145576741","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}
扫码关注我们
求助内容：
应助结果提醒方式：
京公网安备 11010802042870号