Supercomputing: 9th Russian Supercomputing Days, RuSCDays 2023, Moscow, Russia, September 25–26, 2023, Revised Selected Papers, Part I (Lecture Notes in Computer Science)

✍ Scribed by Vladimir Voevodin (editor), Sergey Sobolev (editor), Mikhail Yakobovskiy (editor), Rashit Shagaliev (editor)

• Publisher: Springer
• Year: 2024
• Tongue: English
• Leaves: 333
• Category: Library

✦ Synopsis

The two-volume set LNCS 14388 and 14389 constitutes the refereed proceedings of the 9th Russian Supercomputing Days International Conference (RuSCDays 2023) held in Moscow, Russia, during September 25-26, 2023.

The 44 full papers and 1 short paper presented in these proceedings were carefully reviewed and selected from 104 submissions. The papers have been organized in the following topical sections: supercomputer simulation; distributed computing; and HPC, BigData, AI: algorithms, technologies, evaluation.

✦ Table of Contents

Preface
Organization
Contents – Part I
Contents – Part II
Supercomputer Simulation
Application of HPC for Simulation of Sealant Influence on the Aircraft Assembly Process
1 Introduction
2 Mathematical Model
3 Numerical Results
3.1 Influence of Assembly Parameters on the Assembly Result
3.2 Efficiency of Task Parallelization
4 Conclusion
References
Block Algebraic Multigrid Method for Saddle-Point Problems of Various Physics
1 Introduction
2 Block Algebraic Multigrid Method
2.1 Coarse Space Selection
2.2 Block Gauss–Seidel Smoother
2.3 Implementation with S3M Facilities
3 Numerical Experiments
3.1 Problems Considered
3.2 Convergence Results
4 Conclusions
References
Computational Efficiency of Iterative Methods for Solving Inverse Problems
1 Introduction
2 Problem Statement
3 Multistage Iterative Method
4 Computational Efficiency of the Multistage Method
4.1 Parameters of the Multistage Method
4.2 Multi-core CPU Performance
4.3 GPU Performance
5 Conclusion
References
Computer Memory Requirements for Matrix-Forming Approach to Global Stability Analysis of Fluid Flows
1 Introduction
2 Global Stability Analysis Methodology
3 Solution of the Generalized Eigenvalue Problem
4 Considered Stability Problems
4.1 GSA-Based Prediction of Forming and Evolution of Tollmien-Schlichting Waves in Laminar Flow Over a Flat Plate with Rectangular Cavity
4.2 GSA-Based Prediction of Onset of Transonic Buffet in Turbulent Flow Past Infinite Swept Wing
4.3 GSA-Based Prediction of Onset of Transonic Buffet in Turbulent Flow Past Simplified NASA Common Research Model (CRM) of Airplane
5 Estimates of RAM Requirements
6 Conclusions
References
CubicEoS.jl: Extensible, Open-Source Isothermal Phase Equilibrium Calculations for Fluids
1 Introduction
2 Problem Statement
2.1 Phase Stability
2.2 Phase Split
2.3 Models of Fluid
3 Numerical Approach
3.1 Phase Stability
3.2 Phase Split
4 Implementation Notes
5 Results
5.1 Example 1
5.2 Example 2
5.3 Example 3
5.4 Example 4
5.5 Multi-Threaded Computation of Phase Diagram
6 Conclusion
References
Efficiency and Accuracy of High-Performance Calculations of the Electrostatic Energy of Thin Films Atomistic Clusters
1 Introduction
2 Method of the Simulation
3 Results and Discussion
4 Conclusions
References
Field-Split Iterative Solver Vs Direct One for Quasi-Static Biot Equation
1 Introduction
2 Problem Overview
2.1 Biot Equations in Quasistatic State
2.2 Discretization the Biot Equations
3 Algorithms of SLAE Solution
3.1 Direct Solver
3.2 Iterative Solver
3.3 Construct SLAE Preconditioner
4 Implementation of the Direct and Iterative Algorithms
5 Numerical Experiments to Estimate Complexity of the Direct and Iterative Approaches
6 Conclusion
References
GPU-Accelerated Matrix Exponent for Solving 1D Time-Dependent Schrödinger Equation
1 Introduction
2 Related Works
3 1D TDSE Model of a H2+ Molecular Ion
4 Remarks About the Matrix Exponent Calculation
5 Testing Platforms
6 Analysis of Numerical Experiments
7 Results
8 Discussion
9 Conclusions
References
How to Make Lanczos-Montgomery Fast on Modern Supercomputers?
1 RSA-232 Factorization
1.1 GNFS Stages
2 Parallel Block Lanczos-Montgomery Method
2.1 The Algorithm
2.2 Implementation Details
2.3 Analysis of INM RAS Implementation Performance
3 How to Make the Method Fast?
3.1 Optimizations on CPU
3.2 General Improvements
4 Conclusion
References
Multi-node GPU-Enabled Pseudo-spectral Solver for Turbulence Problems
1 Introduction
2 Multi-node Multi-GPU Fast Fourier Transform
3 Structure of TARANG, a Pseudo-spectral Solver
4 Solving 2D MHD Turbulence Using TARANG
5 Conclusions
References
Multicriteria Optimization of Chemical Reactions Using Interval Analysis. Parallel Scheme for Implementing the Computational Process
1 Introduction
2 Formulation of the Problem of Multicriteria Optimization of Chemical Reactions Based on the Interval Kinetic Model
3 Interval Kinetic Model of Olefin Hydroalumination Reaction in the Presence of Triisobutylaluminum
4 Multicriteria Optimization of the Conditions of the Olefin Hydroalumination Reaction in the Presence of Triisobutylaluminum
5 Parallel Scheme for Implementing the Computational Process
6 Conclusions
References
Parallel Numerical Implementation 3D Wave Hydrodynamics and SWAN Models
1 Introduction
2 3D Wave Hydrodynamics Mathematical Model
3 Wave Effects in SWAN
4 SWAN Parallel Mode
5 3D Hydrodynamics Model Parallel Implementation
6 Conclusions
References
Parallel Numerical Simulation of Sonic Log for 3D Heterogeneous TTI Media
1 Introduction
2 Mathematical Formulation
2.1 Finite-Difference Approximation
2.2 Approximation on the Axis r=0
3 Numerical Experiments
3.1 Presentation of the Well Bore Model
3.2 Approximation at R=0
4 Brief Description of the Code
4.1 Hardware Used
4.2 Scalability
5 Conclusion
References
Quantum-Chemical Study of Some Tris(pyrrolo)benzenes and Tris(pyrrolo)-1,3,5-triazines
1 Introduction
2 Calculation Method
3 Results and Discussion
3.1 Enthalpy of Formation
3.2 IR Absorption Spectra
4 Computational Details
5 Conclusions
References
Reduced Precision Computations in the SL-AV Global Atmosphere Model
1 Introduction
2 Global Atmosphere Model SL-AV
3 Porting Parameterizations to Single Precision
3.1 Preliminary Analysis
3.2 Code Modifications
3.3 Experiments and Results
4 Memory Usage Optimization
4.1 Controlling Vector Length
4.2 Localization of Temporary Arrays in the Parameterizations Block
5 Conclusions
References
Scalability of the INM RAS Earth System Model
1 Introduction
2 Materials and Methods
2.1 INM RAS Earth System Model
2.2 Profiler for INMCM
2.3 Perftools
2.4 HPC Systems
3 Results
3.1 Optimal Configurations for INMCM6M
3.2 INMCM Scalability on Different HPC Systems
3.3 Parallel Efficiency of INMCM Modules
4 Discussion and Conclusion
References
Simulation of Free-Surface Fluid Dynamics: Parallelization for GPUs
1 Introduction
2 Basic Equations and Methods
2.1 SPH Method
2.2 Boundary Conditions
3 Parallelization on GPUs
3.1 Parallel Sorting Algorithm
4 Numerical Model Verification
4.1 Dispersion of Surface Gravity Waves in Linear Approximation
4.2 Placement of Immobile SPH Particles on Solid Surface
5 Discussion and Conclusion
References
Software Package USPARS for Large Sparse Systems of Linear Equations
1 Motivation
2 Theoretical Background
3 Extended Functionality
4 Numerical Results
5 Plans
6 Conclusions
References
Supercomputer Search for Coagulation Factor XIIa Inhibitors in the Chinese National Compound Library
1 Introduction
2 Materials and Methods
2.1 Chinese National Compound Library
2.2 Factor XIIa Model
2.3 SOL Program
2.4 Quantum-Chemical Post Processing
3 Results
4 Discussions
5 Conclusions
References
Supercomputer Technologies for Ultrasound Nondestructive Imaging of Low-Contrast Defects in Solids
1 Introduction
2 Formulation of the Problem of Ultrasound Tomography of Flat Objects and Its Solution Methods
3 Numerical Simulations
4 Comparison of the Capabilities of CPU and GPU Platforms for Solving Inverse Diagnostic Problems in NDT
5 Conclusion
References
The Effect of Data Structuring on the Parallel Efficiency of the HydroBox3D Relativistic Code
1 Introduction
2 Mathematical Model and Numerical Method
3 CPU Efficiency Evaluation
4 Conclusions
References
The Efficiency Optimization Study of a Geophysical Code on Manycore Computing Architectures
1 Introduction
2 The Intelligent Support System for Solving Compute-Intensive Problems of Mathematical Physics
3 Statement of the Problem and the Numerical Method
4 Considered Manycore Systems
4.1 Optimizations for Intel and IBM Processors
4.2 Optimizations for NVIDIA Accelerators
4.3 The Performance Comparison on Considerate Manycore Architectures
4.4 General Recommendations for Optimization and Parallelization
5 Conclusion
References
Using Parallel Technologies to Calculate Fluid Dynamic Processes in a Fractured-Porous Reservoir Taking into Account Non-isothermality
1 Introduction
2 Mathematical Model
3 Numerical Algorithm
4 Matrix Sweep
5 Calculation Results
6 Conclusion
References
Author Index