Parallel Computational Technologies: 16th International Conference, PCT 2022, Dubna, Russia, March 29–31, 2022, Revised Selected Papers (Communications in Computer and Information Science, 1618)

This book constitutes the refereed proceedings of the 16th International Conference on Parallel Computational Technologies, PCT 2022, held in Dubna, Russia, during March 29–31, 2022. The 22 full papers included in this book were carefully reviewed and selected from 60 submissions. They were organized in topical sections as follows: high performance architectures, tools and technologies; parallel numerical algorithms; supercomputer simulation. Preface Organization Contents High Performance Architectures, Tools and Technologies VGL Rating: A Novel Benchmarking Suite for Modern Supercomputing Architectures 1 Introduction 2 Related Work 3 Proposed Benchmarking Method 3.1 Selecting Graph Algorithms 3.2 Selecting Input Graphs 3.3 Principles Used to Form the Rating 4 Using VGL as a Benchmarking Core 5 Developed Benchmarking System 6 Using the Developed Benchmarking System to Rank Modern Supercomputing Platforms 7 Conclusion References HPC TaskMaster – Task Efficiency Monitoring System for the Supercomputer Center 1 Introduction 2 Related Work 3 System Architecture 4 Detecting Inefficient Tasks 4.1 Collected Data 4.2 Data Processing 4.3 Inferences 4.4 Example 5 User Statistics 6 Conclusions References Constructing an Expert System for Solving Astrophysical Problems Based on the Ontological Approach 1 Introduction 2 Scheme Construction and Rules Formalization for Solving Compute-Intensive Astrophysics Problems 3 Using Inference Rules for Intelligent Support 4 Codes and Computational Experiments 5 Conclusion References HPC Resources of South Ural State University 1 Introduction 2 High Performance Resources 2.1 ``Tornado SUSU'' Supercomputer 2.2 ``Neurocomputer'' Complex 3 System Software 3.1 Monitoring Systems 3.2 Control Systems 4 Application Software 5 Conclusions References Parallel Numerical Algorithms Comparative Analysis of Parallel Methods for Solving SLAEs in Three-Dimensional Initial-Boundary Value Problems 1 Introduction 2 Continuous and Discrete Problem Setting 3 Methods for Solving SLAEs in Implicit Schemes 3.1 Choice of an Initial Approximation for Solving a SLAE 3.2 Iterative Algorithms for Implicit Schemes 3.3 Some Questions of Generalization of the Considered Approaches and Speed-up of Computations 4 Examples of Numerical Experiments 5 Conclusion References Optimization of the Computational Process for Solving Grid Equations on a Heterogeneous Computing System 1 Introduction 2 Method for Solving Grid Equations 3 Software Implementation of the Method for Solving Grid Equations 4 Conclusions References Parallel Methods for Solving Saddle Type Systems 1 Introduction 2 Iterative Algorithms in the Krylov Subspaces for Solving Saddle SLAEs 2.1 General Scheme of the Krylov Approaches for Symmetric and Non-symmetric Algebraic Systems 2.2 Generalized G-K-A – Bidiagonalization Method 3 Scalable Parallelization of Iterative Methods 4 Conclusion References Compact LRnLA Algorithms for Flux-Based Numerical Schemes 1 Introduction 2 Compact Streaming for Flux-Based Schemes 2.1 Problem Statement 2.2 Requirements for Compactness 2.3 Compact Update 3 Implementation with LRnLA Algorithms 3.1 LRnLA Algorithm ConeTorre 3.2 FArSh Data Structure 3.3 Sample Implementation in CUDA 3.4 Performance and Data Sizes 3.5 Performance Benchmarks 4 Results and Discussion References Analysis of Block Stokes-Algebraic Multigrid Preconditioners on GPU Implementations 1 Introduction 2 Problem Formulation 2.1 Discretization and Formulation of Preconditioners 2.2 Discretization and Considered Problems 3 AMGCL Framework Modifications 4 Numerical Results 4.1 Unit Cube Problem 4.2 Stokes Lid Driven Cavity 4.3 Flow in a Channel with Obstacles 4.4 Flow in a Porous Medium 5 Conclusion References Implementation of the Algebraic Multigrid Solver Designed for Graphics Processing Units Based on the AMGCL Framework 1 Introduction 2 Aggregation AMG on the GPU 3 Implementation 4 Numerical Experiments 5 Conclusion References Measuring the Effectiveness of SAT-Based Guess-and-Determine Attacks in Algebraic Cryptanalysis 1 Introduction 2 Preliminaries 3 Using Tree-Like Metrics to Estimate the Effectiveness of SAT-Based Guess-and-Determine Attacks 4 Class of Considered Functions 5 Computational Experiments 5.1 General Scenario 5.2 Implementation and Results 6 Conclusion References Tuning of a Matrix-Matrix Multiplication Algorithm for Several GPUs Connected by Fast Communication Links 1 Introduction 2 Related Work 3 Testing Platforms 4 Parallel Matrix-Matrix Multiplication Algorithm for Multiple GPUs 5 Theoretical Optimal Tile Size 6 Tile Size Tuning for Different Platforms 6.1 Expected tile Size for Experimental Environments 6.2 Experimental Results 7 Discussion 8 Conclusion References Visualizing Multidimensional Linear Programming Problems 1 Introduction 2 Mathematical Model of the LP Visual Representation 3 Parallel Algorithm for Building an LP Problem Image 4 Computational Experiments 5 Conclusion References Supercomputer Simulation Quantum-Chemical Calculations of the Enthalpy of Formation of Some Tetrazine Derivatives 1 Introduction 2 Calculation Method 3 Results and Discussion 3.1 Enthalpy of Formation 3.2 IR Spectra and Frequency Analysis 4 Computational Details 5 Conclusions References A New Approach to the Supercomputer Simulation of Carbon Burning Sub-grid Physics in Ia Type Supernovae Explosion 1 Introduction 2 Numerical Model 2.1 Hydrodynamic Equations 2.2 Stellar Equation of State 2.3 Initial Profile 2.4 Carbon Burning 3 Parallel & Distributed Code 3.1 Turbulence Model of Carbon Burning 3.2 Nested Grid 3.3 Performance 4 Numerical Simulation of SNeIa Explosion 4.1 Turbulence Carbon Burning 4.2 Hydrodynamics of SNeIa Explosion 5 Discussion 6 Conclusions References Parallel Simulations of Dynamic Interaction Between Train Pantographs and an Overhead Catenary Line 1 Introduction 2 Mathematical Models 2.1 Catenary Line Model 2.2 Non-linear Static Analysis 2.3 Dynamic Analysis 2.4 Pantograph Model. Contact Interaction Between Pantographs and the Contact Wire 3 Model Problems Setup 3.1 Problems 1,2: Free Oscillations Excited Statically and Dynamically (by Impact Interaction) 3.2 Problem 3: Oscillations Driven by a Constant Push Force Moving Along the Span 4 Numerical Time-Integration Methods 5 Software Implementation 6 Results 6.1 Model Problems: Free Oscillations 6.2 Model Problem: Driven Oscillations 6.3 Validation of the Model Against the Etalon Problem from EN 50318:2018 7 OpenMP Parallelization and Code Performance 8 Conclusions References Construction of a Parallel Algorithm for the Numerical Modeling of Coke Sediments Burning from the Spherical Catalyst Grain 1 Introduction 2 Mathematical Model of Coke Burning 3 Constructing a Parallel Algorithm and Efficiency 4 Results of a Computational Experiment 5 Conclusion References MPI-Based PFEM-2 Method Solver for Convection-Dominated CFD Problems 1 Introduction 2 PFEM-2 Method 2.1 Main Concepts and Solution Algorithm 2.2 FEM Solution on the Eulerian Mesh 2.3 Operations with Lagrangian Particles 3 Parallel Software Implementation of PFEM-2 3.1 Parallel Version of the FEM Solution in PFEM-2 3.2 Parallel Version of Operations with Particles in PFEM-2 4 Numerical Experiment 4.1 Test Problem 5 Conclusion References Modeling of Two-Phase Fluid Flow Processes in a Fractured-Porous Type Reservoir Using Parallel Computations 1 Introduction 2 Formulation of the Problem 3 Parallel Implementation 4 Calculation Results 5 Conclusion References Kinetic Modeling of Isobutane Alkylation with Mixed C4 Olefins and Sulfuric Acid as a Catalyst Using the Asynchronous Global Optimization Algorithm 1 Introduction 2 Problem Statement 3 Parallel Algorithm for Solving Global Optimization Problems 3.1 Global Optimization Problem 3.2 Parallel Asynchronous Global Search Algorithm 4 Numerical Experiments 4.1 Search for Activation Energies and Pre-exponential Multipliers of the Reaction 4.2 Searching for Rate Constants Separately for Each Temperature 5 Conclusions and Future Work References Simulation of Nonstationary Thermal Fields in Permafrost Using Multicore Processors 1 Introduction 2 Problem Statement and Mathematical Model 3 Numerical Method 4 Parallel Implementation and Numerical Experiments 5 Conclusion References High-Performance Calculations for Modeling the Propagation of Allergenic Plant Pollen in an Atmospheric Boundary Layer 1 Introduction 2 Medical Aspects for the Problem of the Propagation of Allergenic Plant Pollen 3 Analysis of the Methods of Propagating Allergenic Plant Pollen 4 Mathematical Model and Its Computer Implementation of the Problem 5 Results of Modeling 6 Conclusion References Author Index