Supercomputing: 7th Russian Supercomputing Days, RuSCDays 2021, Moscow, Russia, September 27–28, 2021, Revised Selected Papers (Communications in Computer and Information Science Book 1510)

This book constitutes the refereed post-conference proceedings of the 7th Russian Supercomputing Days, RuSCDays 2021, held in Moscow, Russia, in September 2021. The 37 revised full papers and 3 short papers presented were carefully reviewed and selected from 99 submissions. The papers are organized in the following topical sections: supercomputer simulation; HPC, BigData, AI: architectures, technologies, tools; and distributed and cloud computing. Preface Organization Contents Supercomputer Simulation 3D Simulation of the Reactive Transport at Pore Scale 1 Introduction 2 Statement of the Problem 2.1 Level-Set Method 3 Numerical Solution 3.1 Solution of the Stokes Equation 3.2 Solving the Convection-Diffusion Equation 3.3 Movement of the Interface 4 Implementation of the Algorithm 5 Numerical Experiments 6 Conclusions References Application of Docking and Quantum Chemistry to the Search for Inhibitors of SARS-CoV-2 Main Protease Abstract 1 Introduction 2 Materials and Methods 2.1 Protein Model 2.2 Preparing the Database for Virtual Screening 2.3 SOL Docking Program 2.4 Protein-Ligand Binding Enthalpy 3 Results 4 Conclusions Acknowledgements References Calculation of Integral Coefficients for Correlation Magnetodynamics and Verification of the Theory 1 Introduction 2 CMD Equations 3 Calculation of Integral CMD Coefficients 4 CMD Equations Verification 5 Conclusion References Clumps and Filaments Generated by Collisions Between Rotating and Non-rotating Molecular Clouds Abstract 1 Introduction 2 Modeling and Parallel Code Realization 3 Results and Discussion 4 Conclusion Acknowledgements References Exploiting Structural Constraints of Proteolytic Catalytic Triads for Fast Supercomputer Scaffold Probing in Enzyme Design Studies Abstract 1 Introduction 2 Materials and Methods 2.1 Backbone-Based Vectorization of Triads 2.2 Protease Triads Dataset Construction 2.3 Clusterization of Triad Vectors 2.4 Visualization of Clusterization Results 2.5 Scaffold Preprocessing and Placement Search Procedure 2.6 Scaffold Library Construction 2.7 Rosetta Match Assessment 3 Results and Discussion 3.1 Backbone-Based Vectorization of Triads 3.2 Space of Architectures of Proteases’ Catalytic Triads 3.3 Scaffold Searching 4 Conclusions Acknowledgements References Fast Computation of Electromagnetic Field in the Far Field Zone Based on Spherical Coordinates Splitting 1 Introduction 2 Method Theory 3 Far Field Calculation Algorithms 3.1 Traditional Algorithms 3.2 Spherical Coordinates Splitting Algorithms 4 Calculation Results 5 Conclusion References Greedy Dissection Method for Shared Parallelism in Incomplete Factorization Within INMOST Platform 1 Introduction 2 INMOST Linear Solvers 3 Greedy Dissection 4 Solution Strategy 5 Numerical Experiment 5.1 Poisson Problem Matrix 5.2 Black-Oil Problem Matrix 6 Conclusions References Heterogeneous Parallel Algorithm for Compressible Flow Simulations on Adaptive Mixed Meshes Abstract 1 Introduction 2 Mathematical Model and Numerical Method 3 Parallel Algorithm and Implementation 3.1 Sequential Time Integration Step Algorithm 3.2 Distributed-Memory MPI Parallelization 3.3 Heterogeneous Execution and Overlap Mode 4 Parallel Efficiency and Scalability 4.1 MPI Parallelization and Overlap Mode 4.2 Heterogeneous Execution 4.3 Parallel Program Scaling 5 Conclusions Acknowledgments References High-Performance Atomistic Modeling of Evaporation of Thin Films Under Intense Laser Irradiation Abstract 1 Introduction 2 Method 3 Results and Discussion 4 Conclusions References High-Performance Implementation of 3D Seismic Target-Oriented Imaging Abstract 1 Introduction 2 Target-Oriented Imaging Operator 3 Implementation of Imaging Operator 4 Numerical Examples 4.1 Angle Domain Imaging vs. Acquisition Domain Migration 4.2 Synthetic Data Tests 4.3 Real Data Tests 5 Conclusions Acknowledgments References Implementation of Elliptic Solvers Within ParCS Parallel Framework 1 Introduction 2 Model Elliptic Equation 3 Solvers Description 3.1 BICGstab 3.2 Geometric Multigrid 3.3 Implementation 4 Performance and Scalability 5 Conclusion and Outlook References JAD-Based SpMV Kernels Using Multiple-Precision Libraries for GPUs 1 Introduction 2 Background 2.1 Multiple-Precision GPU Libraries 2.2 Jagged Diagonal Format 3 JAD Kernels Using Multiple-Precision Libraries 4 Performance and Memory Evaluation 5 Conclusions References Modeling the Process of Mixing Water in the Estuarine Area in the Presence of a Density Gradient of the Aquatic Environment Using Hybrid MPI/OpenMP Technology 1 Introduction 2 Problem Setup 3 Method for Solving Grid Equations 4 Parallel Implementation of the Algorithm 5 Results of Numerical Experiments 6 Conclusions References Multicomp: Software Package for Multiscale Simulations Abstract 1 Introduction 2 Multicomp Package 3 Examples of Simulations Performed in Multicomp and Productivity Evaluation 4 Summary Acknowledgments References Parallel Algorithm for Calculating the Radius of Stability in Multicriteria Optimization Conditions for Catalytic Reforming of Gasoline Abstract 1 Introduction 2 Multicriteria Optimization in Catalytic Reforming of Gasoline Based on a Kinetic Model 3 Mathematical Formulation of the Problem of Pareto Set Stability Analysis 4 Parallel Algorithm for Analyzing the Stability of the Pareto Frontier 5 Research Results 6 Conclusion Acknowledgement References Parallel Global Search Algorithm for Optimization of the Kinetic Parameters of Chemical Reactions 1 Introduction 2 Mathematical Model 3 Parallel Global Search Algorithm 4 Numerical Experiments 5 Conclusions and Future Work References Parameters Optimization of Linear and Nonlinear Solvers in GeRa Code 1 Introduction 2 Description of Models 3 Numerical Experiments for Different Linear Solvers 4 Run-Time Parameters Optimization 5 Adaptive Relaxation for Newton's Method 6 Conclusion References Quantum-Chemical Calculations of the Structure and Thermochemical Properties of 3,6-bis(2,2,2-trinitroethylnitramino)-1,2,4,5-tetrazine 1 Introduction 2 Methods 3 Results and Discussion 3.1 Structure 3.2 Enthalpy of Formation 3.3 IR Absorption Spectra 4 Computational Details 5 Conclusions References Solution of Large-Scale Black Oil Recovery Problem in Parallel Using INMOST Platform 1 Introduction 2 INMOST Functionality 3 Problem Setting 4 Wells 5 Numerical Experiment 6 Conclusions References Supercomputer Simulation of Turbulent Flow Around Isolated UAV Rotor and Associated Acoustic Fields 1 Introduction 2 Mathematical Model and Numerical Method 3 Parallel Implementation 4 Problem Setup 5 Numerical Case Setup 6 Results 7 Conclusion References Using 16-th Order Multioperators-Based Scheme for Supercomputer Simulation of the Initial Stage of Laminar-Turbulent Transitions 1 Introduction 2 Computational Details 2.1 Problem Formulation 2.2 Multioperators-Based Scheme 3 Parallel Implementation 4 Examples of Numerical Solutions 5 Conclusions References Wave Spectrum of Flowing Drops 1 Introduction 2 Model and Methods 2.1 Lattice Boltzmann Method 2.2 Immersed Boundary Method 2.3 Combined IB-LBM Algorithm 2.4 Drop Motion 3 Simulation 3.1 Simulation Setup 3.2 Details of Simulation 3.3 Spectrum Calculation 4 Results 5 Conclusion References HPC, BigData, AI: Architectures, Technologies, Tools Advanced Genetic Algorithm in the Problem of Linear Solver Parameters Optimization 1 Introduction 2 Optimization Approach 2.1 Genetic Algorithm 2.2 Introducing the Neural Network 2.3 Generalized Optimization Algorithm 3 Neural Network Accuracy Evaluation 4 Software and Numerical Methods 5 Optimization Algorithm Evaluation 5.1 NN Model Training 5.2 Examples of Neural Network Predictions 5.3 Multiple Launches: Stability Test 5.4 CFD SLAEs Optimization 6 Conclusions References Analysis of Software Package Usage Based on Methods for Identifying Similar HPC Applications 1 Introduction 2 Background and Related Work 3 Methods for Software Package Detection 3.1 Overview of Previously Developed Method 3.2 Assessing Real-Life Accuracy of Static Method 4 Applying Proposed Methods for Detecting Similar HPC Applications for Package Version Detection 4.1 Prebuilt Package Version Detection 4.2 Unknown Package Version Detection 5 Conclusions and Future Work References Comparative Efficiency Analysis of MPI Blocking and Non-blocking Communications with Coarray Fortran 1 Introduction 2 Statement of the Problem 3 Numerical Approach 4 Domain Decomposition 5 Classical MPI Point-to-Point Communications 6 Coarray Fortran 7 Testing and Analysis 8 Conclusion References Investigating Performance of the XAMG Library for Solving Linear Systems with Multiple Right-Hand Sides 1 Introduction 2 Capabilities of the XAMG Library 3 Basic Design Features 3.1 Program Code Elements of XAMG Library 4 Code Optimizations 4.1 Reduced-Precision Floating-Point Calculations 4.2 Compression of Indices 4.3 Vector Status Flags 4.4 Per Level Hierarchy Construction 4.5 Vectorization of Basic Operations 5 Performance Evaluation 5.1 Test Application 5.2 Testing Methodology 5.3 Single-Node Performance Evaluation Results 5.4 Multi-node Performance Evaluation Results 6 Conclusions References ``Mini-Benchmarking'' Approach to Optimize Evolutionary Methods of Neural Architecture Search 1 Introduction 2 Neural Architecture Search Problems 3 Techniques for Tuning Hyperparameters 4 NAS Search Space Optimization 4.1 ``Mini-Benchmarking'' Approach 5 Automatic System Architecture 6 Experiments 6.1 Experimental Setup and Configuration 7 Conclusion References On the Prospects of Quantum Computing in Models of Social Behavior 1 Introduction 2 Finite-Dimensional Quantum Models 3 The Simplest Model of Community Behavior 4 Conclusion References Qualitative and Quantitative Study of Modern GPU Synchronization Approaches 1 Introduction 2 Hardware and Software Parallelism in CUDA 3 Problem Statement and Implementation 3.1 Asynchronous Implementation 3.2 Synchronous Implementation with Cooperative Groups 3.3 Synchronous Implementation with Semaphores 4 Benchmarks and Tau Model 5 Conclusion References Rust Language for Supercomputing Applications 1 Introduction 2 Related Work 3 Simple Code and Mathematics Libraries 3.1 Benchmarks 3.2 Matrix Multiplication 3.3 Linear Algebra Libraries 4 Shared-Memory Parallelism 4.1 Benchmarks 5 MPI 5.1 Memory Safety 5.2 Absence of Data Races 5.3 Absence of Undefined Behavior 5.4 Incorporating Verbosity into Types 5.5 Benchmarks 6 Conclusion References Study of the Algorithms Information Structure as the Basis of a Training Workshop 1 Introduction 2 Practical Assignment 2020 3 Algolang Language 3.1 General Approach 3.2 Language Description 4 Algoload System 5 AlgoView Visualization System 6 Practical Assignment Results 7 Conclusions References The Impact of Compiler Level Optimization on the Performance of Iterative Poisson Solver for Numerical Modeling of Protostellar Disks 1 Introduction 2 Mathematical Model and Numerical Simulation 3 Performance Analysis 4 Conclusion References Universal Assessment System for Analyzing the Quality of Supercomputer Resources Usage 1 Introduction 2 Background and Related Work 3 Description of the Assessment System 3.1 General Approach 3.2 Proposed Assessments 3.3 Possible Benefits of Using the Proposed System 4 Details on the Implementation of the Assessments 5 Evaluation of the Developed Assessment System 5.1 Assessing CPU and Memory Usage 5.2 Assessing Network Usage 6 Conclusions and Future Work References Distributed and Cloud Computing Desktop Grid and Cloud Computing: Short Survey 1 Introduction: Desktop Grid and Cloud Computing 2 Survey 2.1 Extension of a Desktop Grid Using a Cloud 2.2 Extension of a Cloud Using a Desktop Grid Resources 2.3 Implementation of a Cloud Basing on Resources of a Desktop Grid 2.4 Desktop Grid as a Service 2.5 Other Ideas 3 Conclusion References Multi-cloud Privacy-Preserving Logistic Regression Abstract 1 Introduction 2 Logistic Regression and Gradient Descent 3 Related Work 4 Multi-cloud Logistic Regression with Residue Number System 4.1 Residue Number System 4.2 Logistic Regression with Residue Number System 4.3 Multi-cloud Logistic Regression 4.4 Logistic Function Approximation 5 Performance Analysis 6 Conclusion Acknowledgment References National Research Computer Network of Russia: Development Plans, Role and Facilities in Ensuring the Functioning of the National Supercomputer Infrastructure Abstract 1 Introduction 2 NIKS’2021: Telecommunications Infrastructure, Network Connectivity and User Base 2.1 An Overview of the Telecommunications Infrastructure of NIKS 2.2 General Characteristics of the User Base of NIKS 3 Directions for the Development of the Infrastructure and Service Platform of NIKS 3.1 Criteria and Principles for Selecting Organizations to Connect to NIKS 3.2 Main Directions of Development of the Telecommunications Infrastructure 3.3 NIKS as the Telecommunications Infrastructure for High-Speed Connectivity of SCCs of NSI 4 Conclusion Acknowledgment References Project Progress Forecasting in a Desktop Grid 1 Introduction 2 Desktop Grid and BOINC 3 Mathematical Model 3.1 Forecasting Model 3.2 Statistical Model 4 Numerical Experiments 5 Conclusion and Discussion References Searching for Orthogonal Latin Squares via Cells Mapping and BOINC-Based Cube-and-Conquer 1 Introduction 2 Orthogonal Diagonal Latin Squares 3 Finding Orthogonal Latin Squares via Cells Mapping Schemes 3.1 Cells Mapping Schema 3.2 ESODLS CMS 3.3 Search for ESODLS CMS of Order 10 4 Finding MODLS via CMS, Cube-and-Conquer, and BOINC 4.1 SAT Encoding 4.2 Solving in RakeSearch via Cube-and-Conquer 5 Related Works 6 Conclusions References Toward Crowdsourced Drug Discovery: Start-Up of the Volunteer Computing Project SiDock@home 1 Introduction 2 Preliminaries 2.1 Virtual Drug Screening 2.2 BOINC-Based Volunteer Computing 2.3 The Project SiDock@home 3 Computational Experiments 3.1 Server 3.2 Clients 4 Conclusion References Using Mobile Devices in a Voluntary Distributed Computing Project to Solve Combinatorial Problems Abstract 1 Introduction 2 BOINC Software 3 Integration of Mobile Devices into the Desktop Grid 4 Computational Application 4.1 Algorithm 4.2 Development 5 Setup Project Parameters 6 Results 7 Conclusion Acknowledgements References Author Index