Graphics Library Research Articles

TorchAmi: Generalized CPU/GPU implementation of algorithmic matsubara integration

We present torchami, an advanced implementation of algorithmic Matsubara integration (AMI) that utilizes pytorch as a backend to provide easy parallelization and GPU support. AMI is a tool for analytically resolving the sequence of nested Matsubara integrals that arise in virtually all Feynman perturbative expansions. In this implementation we present a new AMI algorithm that creates a more natural symbolic representation of the Feynman integrands. In addition, we include peripheral tools that allow for import and labeling of simple graph structures and conversion to torchami input. The code is written in c++ with python bindings provided. Program summaryProgram Title: torchamiCPC Library link to program files:https://doi.org/10.17632/m79hnngy8s.1Developer's repository link:https://github.com/mdburke11/torchami/releases/tag/v1.0Licensing provisions: GPLv3Programming language:C++, pythonNature of problem: Feynman diagrams are pictorial representations of perturbative expansions often formulated in the imaginary frequency/time axis and involve a high-dimensional sequence of nested integral over spatial and temporal degrees of freedom.Solution method:torchami provides a framework to symbolically generate and store the analytic solution to the temporal Matsubara sums through repeated application of multipole residue theorems. The solutions are stored using a tree structure for arbitrary products and sums of Fermi/Bose functions, and the evaluation functions provide both CPU and GPU support with automatic parallelization for batch sampling problems.Additional comments including restrictions and unusual features: Requires C++17 standard, the boost graph library, as well as pytorch.

3D Reconstruction of Geometries for Urban Areas Supported by Computer Vision or Procedural Generations

This work presents a numerical mesh generation method for 3D urban scenes that could be easily converted into any 3D format, different from most implementations which are limited to specific environments in their applicability. The building models have shaped roofs and faces with static colors, combining the buildings with a ground grid. The building generation uses geographic positions and shape names, which can be extracted from OpenStreetMap. Additional steps, like a computer vision method, can be integrated into the generation optionally to improve the quality of the model, although this is highly time-consuming. Its function is to classify unknown roof shapes from satellite images with adequate resolution. The generation can also use custom geographic information. This aspect was tested using information created by procedural processes. The method was validated by results generated for many realistic scenarios with multiple building entities, comparing the results between using computer vision and not. The generated models were attempted to be rendered under Graphics Library Transmission Format and Unity Engine. In future work, a polygon-covering algorithm needs to be completed to process the building footprints more effectively, and a solution is required for the missing height values in OpenStreetMap.

Open-Source Solutions for Real-Time 3D Geospatial Web Integration

Abstract. The recent development of Urban Digital Twin (UDT) structures and the spread of Internet of Things (IoT) technology have extended the concept of Digital Twin (DT) to urban environment representation, enabling real-time connections between geospatial representations and their real-world counterparts. Concurrently, advancements in surveying technologies and web geospatial services have facilitated the acquisition of extensive 2D and 3D geospatial data, making their spatial integration essential. While some proprietary software solutions have recently enabled initial examples of UDT by integrating various 3D geospatial data and real-time sensor acquisition, the development of open-source solutions for real-time web-based geospatial management remains a challenge. This work demonstrates a possible open-source solution for developing a UDT accessible via the web. The case study involves the web visualization platform of the new Faculty of ITC building at the University of Twente in Enschede (The Netherlands), which is connected in real-time with weather data services provided by Visual Crossing (1). The platform is based on an HTML5 framework and employs WebGL JavaScript graphic libraries to visualize different modules of 3D web visualization that integrate various local and remote 3D datasets. The framework adopted in this experiment can be used in the future for developing low-cost UDT web platforms beneficial for researchers, municipalities, and private companies interested in the real-time geospatial management of urban environments.

Multi-pose dress simulation of fully-fashioned knitted skirt based on loop structure

To realize the three-dimensional structure simulation and dress simulation of the fully-fashioned knitted skirt, based on the study of the structural characteristics of the fully-fashioned knitted skirt, the mathematical model of the knitted skirt pattern, the loop geometry mathematical model and the loop mesh model were established by the method of mathematical modelling. The spatial transformation relationship of each pair of triangles in the plate model and the three-dimensional surface model was calculated by using the spatial matrix. To draw a loop on the surface model's triangular patch, spatial transformation was used to calculate the grid point positions and obtain three-dimensional coordinates of the loop-type value points. Based on C# programming language, the human body model in the database was matched with the knitted skirt surface model, and the three-dimensional graphics library Open-GL and loop drawing method were used to realize the simulation of the knitted skirt dress based on the loop structure. The simulation method was tested by the dressing effects of hip skirts and umbrella skirts with different postures. The results showed that the spatial transformation could realize the fast transformation and calculate the coordinates of the loop-type value points. The C# programming language and Open-GL technologies realized the visual simulation of the fully-fashioned knitted skirt and the multi-pose dressing simulation based on the loop structure.

ГРАФИЧЕСКИЕ СРЕДСТВА ОПЕРАЦИОННЫХ СИСТЕМ

The methods of processing multimedia and graphical information in operating systems are studied using the example of MCI, API and GDI interfaces of Windows operating system. The interfaces of GDI and GDI+ graphics devices are examined in detail. The capabilities and features of the OpenGL graphics library, which is commonly used in the Windows operating system, are presented. The principles of operation of the GDI interface are given. Three types of drawing tools supported by the context of the GDI interface are examined in detail: pen, brush and font. The stages of coordination of logical and physical coordinate systems are considered. The OpenGL library graphics elements (bits arrays, polygons, segments, polylines, etc.), as well as basic geometric objects and a set of polyhedra: cube, cone, torus, octahedron, icosahedron, dodecahedron and tetrahedron are considered. As an example of OpenGL library usage, a three-dimensional graphics editor was studied. It works with three-dimensional objects of various types and allows reproduction of an object in two modes (lines and a solid surface), as well as performance of following operations with a graphic object: deformation, rotation, transpositionong along axes, loading and saving on hard drive.

TLPGNN: A Lightweight Two-level Parallelism Paradigm for Graph Neural Network Computation on Single and Multiple GPUs

Graph Neural Networks (GNNs) are an emerging class of deep learning models specifically designed for graph-structured data. They have been effectively employed in a variety of real-world applications, including recommendation systems, drug development, and analysis of social networks. The GNN computation includes regular neural network operations and general graph convolution operations, which take most of the total computation time. Though several recent works have been proposed to accelerate the computation for GNNs, they face the limitations of heavy pre-processing, low efficiency atomic operations, and unnecessary kernel launches. In this article, we designTLPGNN, a lightweight two-level parallelism paradigm for GNN computation. First, we conduct a systematic analysis of the hardware resource usage of GNN workloads to understand the characteristics of GNN workloads deeply. With the insightful observations, we then divide the GNN computation into two levels, i.e.,vertex parallelismfor the first level andfeature parallelismfor the second. Next, we employ a novel hybrid dynamic workload assignment to address the imbalanced workload distribution. Furthermore, we fuse the kernels to reduce the number of kernel launches and cache the frequently accessed data into registers to avoid unnecessary memory traffic. To scaleTLPGNNto multi-GPU environments, we propose an edge-aware row-wise 1-D partition method to ensure a balanced workload distribution across different GPU devices. Experimental results on various benchmark datasets demonstrate the superiority of our approach, achieving substantial performance improvement over state-of-the-art GNN computation systems, including Deep Graph Library (DGL), GNNAdvisor, and FeatGraph, with speedups of 6.1×, 7.7×, and 3.0×, respectively, on average. Evaluations of multiple-GPUTLPGNNalso demonstrate that our solution achieves both linear scalability and a well-balanced workload distribution.

SEISMIC, a Python-based code of the quantas package to calculate the phase and group acoustic velocities in crystals

The present work reports the theoretical background and the capabilities of SEISMIC, a Python code specifically developed to calculate the propagation of the sound waves inside crystalline materials. SEISMIC is a tool integrated in the Quantas package and provides a series of useful information for engineers and geophysicists, such as the phase and group velocities, the power flow angle, the enhancement factors, and the polarization vectors, using as input the elastic moduli and the density of the material. Numerical treatments of the derivatives were avoided, using analytical methods to obtain numerically stable results. The code relies only on Python numerical and graphical libraries to ensure a full cross-platform usability.

The plebeian Graph Library: A WebGL based network visualisation and diagnostics package

The plebeian Graph Library: A WebGL based network visualisation and diagnostics package

FROM 3D SURVEY TO DIGITAL REALITY OF A COMPLEX ARCHITECTURE: A DIGITAL WORKFLOW FOR CULTURAL HERITAGE PROMOTION

Abstract. In recent years, the digitalization and dissemination of historical heritage have become crucial nodes in the preservation and valorization of Cultural Heritage (CH). Technologies such as Unmanned Aerial Vehicle (UAV) and terrestrial photogrammetry, Terrestrial Laser Scanning (TLS) and handheld Simultaneous Localisation and Mapping (SLAM) laser scanning allow the generation of digital models of architecture that can be explored through interactive web platforms, such as those based on WebGL graphic library. These are considered one of the most promising innovations for digitizing and sharing CH site due to their application in a wide range of contexts, promoting new forms of interaction with architecture at different scales. Additionally, the use of geomatic tools allows for a more complete 3D reconstruction and evaluation of the results by comparing different techniques. The article focuses on digitization as a tool for documenting and sharing CH assets, with the aim of developing a replicable prototype platform for an immersive Virtual Tour (VT) of an art collection and the architectural complex in which it is resided. In addition, this paper presents the results of a case study conducted at the Ricci Oddi Gallery of Modern Art in Piacenza, Italy. The source code of the implemented application is available on GitHub to permit replicability for other case studies.

Dose-response modeling and treatment plan assessment with a python software toolkit

This software assistant aims at calculating the dose-response relations of tumors and normal tissues, or clinically assessing already determined values by other researchers. It can also indicate the optimal dose prescription by optimizing the expected treatment outcome. The software is developed solely in python programming language, and it employs PSFL license for its Graphical User Interface (GUI), NUMPY, MATPLOTLIB, and SCIPY libraries. It comprises of two components. The first is the Dose-response relations derivation component, which takes as input the dose volume histograms (DVHs) of patients and their recorded responses regarding a given clinical endpoint to determine the parameters of different tumor control probability (TCP) or normal tissue complication probability (NTCP) models. The second is the Treatment Plan Assessment component, which uses the DVHs of a plan and the dose-response parameters values of the involved tumors and organs at risk (OARs) to calculate their expected responses. Additionally, the overall probabilities of benefit (PB), injury (PI) and complication-free tumor control (P+) are calculated. The software calculates rapidly the corresponding generalized equivalent uniform doses (gEUD) and biologically effective uniform doses (D‾‾) for the Lyman-Kutcher-Burman (LKB), parallel volume (PV) and relative seriality (RS) models respectively, determining the model parameters. In the Dose-Response Relations Derivation component, the software plots the dose-response curves of the irradiated organ with the relevant confidence internals along with the data of the patients with and without toxicity. It also calculates the odds ratio (OR) and the area under the curve (AUC) of different dose metrics or model parameter values against the individual patient outcomes to determine their discrimination capacity. It also performs a goodness-of-fit evaluation of any model parameter set. The user has the option of viewing plots like Scatter, 3D surfaces, and Bootstrap plots. In the Treatment Plan Assessment part, the software calculates the TCP and NTCP values of the involved tumors and OARs, respectively. Furthermore, it plots the dose-response curves of the TCPs, NTCPs, PB, PI, and P+ for a range of prescription doses for different treatment plans. The presented software is ideal for efficiently conducting studies of radiobiological modeling. Furthermore, it is ideal for performing treatment plan assessment, comparison, and optimization studies.

Stereographic Projection of Theoretical Orientation Relationships Between Crystals of Any Types in Phase Transformation and Precipitation

AbstractStereographic projection is a classic technology to represent the angular relationships of lattice planes and directions. In some cases, it is necessary to project arbitrary lattice planes or directions of a crystal onto an arbitrary lattice plane of another crystal, e.g., when representing orientation relationships (ORs) in phase transformation or precipitation. Commercial EBSD (electron backscatter diffraction) software cannot illustrate theoretical orientation relationships based purely on the parallelism conditions. This work presents a generalized and unified formulation for stereographic projection of any lattice planes or directions onto any lattice planes of any crystals in the 32 point groups, which is utilized to represent the theoretical orientation relationships between any crystal types. Examples are given to illustrate the correlations of common orientation relationships in both cubic and hexagonal crystals. A procedure is also established to color code and simultaneously project all low‐index Miller/Bravais planes or directions of a crystal onto an arbitrary lattice plane of the same crystal, which can be utilized to visualize the crystal symmetry and determine the standard stereographic triangle (SST). All these illustrations are realized in a computer program written in C++ using the OpenGL graphic libraries.

HAZniCS – Software Components for Multiphysics Problems

We introduce the software toolbox HAZniCS for solving interface-coupled multiphysics problems. HAZniCS is a suite of modules that combines the well-known FEniCS framework for finite element discretization with solver and graph library HAZmath. The focus of this article is on the design and implementation of robust and efficient solver algorithms which tackle issues related to the complex interfacial coupling of the physical problems often encountered in applications in brain biomechanics. The robustness and efficiency of the numerical algorithms and methods is shown in several numerical examples, namely the Darcy-Stokes equations that model the flow of cerebrospinal fluid in the human brain and the mixed-dimensional model of electrodiffusion in the brain tissue.

CoGNN: An Algorithm-Hardware Co-Design Approach to Accelerate GNN Inference With Minibatch Sampling

As a new algorithm of graph embedding, Graph Neural Networks (GNNs) have been widely used in many fields. However, GNN computing has the characteristics of both sparse graph processing and dense neural network, which make it difficult to be deployed efficiently on the existing graph processing accelerators or neural network accelerators. Recently, some GNN accelerators have been proposed, but the following challenges have not been fully solved: 1) The mini-batch GNN inference scenario has the potential of software and hardware co-design, which can bring 30% computation amount reduction, and this is not well utilized. Besides, the cost of Message Flow Graph construction is large and may account for more than 50% of the total delay. 2) The feature aggregation has a large amount of data access and relatively small amount of computation, which leads to low on-chip data reuse, only 10% of dense computing. 3) Without the optimization of sparse computing units, simple memory bank and cross bar architecture can easily lead to bank access conflict and load imbalance, reducing the utilization of computing units to less than 60%. In order to solve the above problems, we propose a algorithm-hardware co-design scheme to accelerate GNN inference, which includes three technologies: 1) A reuse-aware sampling method is proposed for mini-batch inference scenarios, which reduces 30% of the calculation and improves the on-chip reusability of local data. 2) Through the node-wise parallelism-aware quantization, the features and weights are quantized to integers with eight or four bits, which reduces the amount of memory access by at least four times. 3) An accelerator supporting the above technologies is designed and evaluated, and different operations are supported by the sampling-inference integration architecture. The multi-bank on-chip memory pool is designed to support data reuse, and edge stream reordering is used to reduce data access conflicts, improving the utilization of computing units by 1.5x. Combined with the above technologies, the experiments show that our design achieves 9.2× speedup and 29× energy efficiency improvement compared with Deep Graph Library framework running on servers equiped with CPU and GPU.

A Scalable Data Structure for Efficient Graph Analytics and In-Place Mutations

The graph model enables a broad range of analyses; thus, graph processing (GP) is an invaluable tool in data analytics. At the heart of every GP system lies a concurrent graph data structure that stores the graph. Such a data structure needs to be highly efficient for both graph algorithms and queries. Due to the continuous evolution, the sparsity, and the scale-free nature of real-world graphs, GP systems face the challenge of providing an appropriate graph data structure that enables both fast analytical workloads and fast, low-memory graph mutations. Existing graph structures offer a hard tradeoff among read-only performance, update friendliness, and memory consumption upon updates. In this paper, we introduce CSR++, a new graph data structure that removes these tradeoffs and enables both fast read-only analytics, and quick and memory-friendly mutations. CSR++ combines ideas from CSR, the fastest read-only data structure, and adjacency lists (ALs) to achieve the best of both worlds. We compare CSR++ to CSR, ALs from the Boost Graph Library (BGL), and the following state-of-the-art update-friendly graph structures: LLAMA, STINGER, GraphOne, and Teseo. In our evaluation, which is based on popular GP algorithms executed over real-world graphs, we show that CSR++ remains close to CSR in read-only concurrent performance (within 10% on average) while significantly outperforming CSR (by an order of magnitude) and LLAMA (by almost 2×) with frequent updates. We also show that both CSR++’s update throughput and analytics performance exceed those of several state-of-the-art graph structures while maintaining low memory consumption when the workload includes updates.

Нечеткая динамическая онтологическая модель для поддержки принятия решений по управлению энергоемкими системами на основе прецедентов

The article discusses the features of applying the precedent approach when managing complex energy-intensive systems in the context of the need to take into account various energy, technical, environmental and operational indicators, as well as the uncertainty of many internal and external factors influence. This leads to the presence of a large amount of semi-structured information that can be presented using various scales, which determines the prospects of using the precedent approach. The proposed fuzzy ontological model for supporting decision support based on precedents is described, characterized by the use of dynamic concepts, as well as concepts in the form of different scale numerical and linguistic variables. An algorithm for assessing the proximity of precedents based on an ontological model is proposed, which differs by taking into account the dynamic aspects of changes in the state of controlled systems. The developed algorithms for fuzzy inference for decision support based on precedents are presented, which allow the use of both linguistic and numerical variables as input characteristics of the fuzzy production model, as well as using various logical connections between the rules pre-requisites. The software that implements the developed model and algorithms is described. Particular attention is paid to the modified fuzzy inference component, implemented using Python 3.8.7 language tools. To implement the user interface of the specified component, the cross-platform graphic library Tkinter was used. The results of computational experiments using real data obtained during the operation of an energy-intensive system for processing fine ore raw materials, including a conveyor-type roasting machine, are presented. Minimization of specific total costs for thermal and electrical energy was considered as a criterion for the effectiveness of management decisions. The outcome obtained showed that the proposed model and software make it possible to obtain a result comparable to the one of using complex analytical dependencies, while ensuring a reduction in time and financial costs.

Time-Varying Topology Formation Reconfiguration Control of the Multi-Agent System Based on the Improved Hungarian Algorithm

Distributed time-varying formation technology for multi-agent systems is recently become a research hotspot in formation control field. However, the formation reconfiguration control technology for agents that randomly appeared to fail during maneuvers is rarely studied. In this paper, the topological relations between intelligence are designed by graph theory to simplify the cooperative interaction between multi-agent systems. Moreover, this paper constructs the time-varying configuration of the target formation based on the rigidity graph theory and leader–follower strategy. Drawing on the establishment of the expert experience database in a collaborative process, we innovatively propose the establishment of a graphic library to help the multi-agent system quickly form an affine transformation as soon as it is disabled. Secondly, the improved Hungarian algorithm is adopted to allocate the target point when the first failure occurs. This algorithm incorporates a gradient weighting factor from the auction algorithm to improve the speed of system reconfiguration with minimum path cost. On this basis, a distributed multi-agent control law based on consistency theory is established, and the system’s stability can be guaranteed via Lyapunov functions. Finally, the simulation results demonstrate the feasibility and effectiveness of the proposed formation reconfiguration control algorithm in a collaborative environment.

Shooting Simulator «Inhibitor»: Mathematical Support of Target Layout Animation

A mathematical support of target layout animation on the image of the shooting range with shielding by nearby objects and without flickering for the optical-electronic shooting simulator'Inhibitor' developed at Institute of mechanics UdmFRC UB RAS and at Computer facilities department of Kalashnikov ISTU jointly with JSC «Kalashnikov» Concern» is described. The tactical and technical task for the implementation of realistic target layout animation is given: support for all exercises from the Shooting Course with emerging-disappearing and moving targets on the surface of the shooting range with local objects that shield them, as well as folds of the area and other targets. Movements should be realistic in speed and angular dimensions of targets, smooth without jerking and correspond to the exercise scenario from the Firing Course: targets up to 2 m/s, armored vehicles up to 4 m/s at an angle of 15-20 from maximum ranges to minimum ones. The animation frequency should be maintained at least 25 frames/s, with the priority of the target layout (at least up to 32 at the same time) over special effects and local objects. The conducted studies of the animated target positioning accuracy with strong reduction in their size at longer ranges to ensure the realism of aiming at such images confirmed compliance with the requirements of TDP (significantly less than 0.3 thousandth range or 1.5 mm). Based on the prospects for further research and development of electronic shooting simulators and thanks to the improvement of computing tools and the development of software graphic libraries, it is necessary to expand the functional capabilities of simulators and reduce the cost, so as to increase competitiveness, for example, due to the realism of the animation of the target environment.

YASARA Model-Interactive Molecular Modeling from Two Dimensions to Virtual Realities.

The industry's transition from three-dimensional (3D) glasses to virtual reality (VR) headsets has left modelers stranded without hardware supply, since walking around and waving arms in a virtual world is a great experience but also very tiring when doing time-intensive modeling work. We present a novel software implementation that uses a VR headset while sitting at a desk in front of the normal screen, which is beamed into the virtual reality together with keyboard, mouse, and chair using the headset's cameras and an extra tracker attached to the seat-back. Compared to 3D glasses, this yields a comparably relaxing but much more immersive workplace and provides additional possibilities such as taking molecules into one's hands, standing up, and walking or teleporting through the models. This VR functionality has been combined with a molecular graphics engine based on Vulkan, a next-generation cross-platform application programming interface (API) for GPUs and the successor of the widely used Open Graphics Library (OpenGL). It is built into the YASARA Model program, which includes many features like small and large molecule builders, electron densities, partial surfaces, contact analysis, coordinate manipulation, and animations. Interactive tutorials are provided to guide modelers into VR and familiarize them with the molecular modeling features. YASARA Model is available for Linux, Windows, Android, and MacOS (the latter without VR) with an introductory video at www.YASARA.org/vr.

基于生物量的玉米可视化模型构建

Maize is one of the characteristic species in Shanxi Province. It has strong drought and cold tolerance. The study of maize growth and development law through the construction of the model is helpful to accurately understand the growth characteristics of maize, improve the yield, and further provide an important theoretical basis for the realization of precision agriculture. In this paper, "Xianyu 335" maize was taken as the research object. The data of leaf length, leaf width, stem length, stem number and tassel length, as well as leaf, stem, tassel, ear, cob and root organ morphology and biomass of maize during the whole growth period were obtained by field experiment. The maize geometric form structure module was constructed. The dynamic simulation of maize growing process was realized by object-oriented programming in Visual Studio software platform. With the help of 3D OpenGL graphic library Nurbs (Non-Uniform Rational B-Splines) curve and surface simulation technology, the growth visualization of maize organs was realized based on the idea of curve and surface control point selection parameter modeling.

CyjShiny: A cytoscape.js R Shiny Widget for network visualization and analysis.

cyjShiny is an open-source R package that allows users to embed network visualization into Shiny apps and R Markdown documents. cyjShiny (https://github.com/cytoscape/cyjShiny) builds on the cytoscape.js Javascript graph library. Additionally, the package provides helper functions to convert common R data representations (e.g., data.frame) into forms compatible with cytoscape.js.