Supercomputing Frontiers 6th Asian Conference, SCFA 2020, Singapore, February 24-27, 2020, Proceedings

This open access book constitutes the refereed proceedings of the 6th Asian Supercomputing Conference, SCFA 2020, which was planned to be held in February 2020, but unfortunately, the physical conference was cancelled due to the COVID-19 pandemic. The 8 full papers presented in this book were carefu...

Full description

Saved in:
Bibliographic Details
Main Author: Panda, Dhabaleswar K
Format: eBook
Language:English
Published: Cham Springer Nature 2020
Springer International Publishing AG
Springer
Edition:1
Series:Lecture Notes in Computer Science
Subjects:
Computer programming, programs, data
ISBN:9783030488420, 303048842X, 9783030488413, 3030488411
Online Access:Get full text
Tags: Add Tag
No Tags, Be the first to tag this record!
Table of Contents:
  • 3 Investigation of Gusset Effect in Between Wave Paddles -- 4 Digital Twin of Marine Vessel for Remote and Autonomous Navigation -- 5 Concluding Remarks -- References -- Correcting Job Walltime in a Resource-Constrained Environment -- 1 Introduction -- 2 Productivity in Resource-Constrained HPC Systems -- 2.1 Job Delays -- 2.2 Walltime Accuracy Effect on Productivity -- 3 Understanding Real HPC Workloads -- 3.1 Walltime Charactertics -- 3.2 Job Diversity and Walltime Scheduling -- 4 Walltime-Based Scheduling -- 4.1 Walltime Prediction and Correction -- 4.2 User-Based Prediction -- 4.3 User and Job-Based Prediction -- 4.4 Correction -- 5 Experimental Setup -- 5.1 Workload Preparation -- 5.2 Scheduling Simulator -- 5.3 Performance Metric -- 6 Simulation Results Analysis -- 6.1 Wait Time Performance -- 6.2 Scheduling Slowdown -- 6.3 Results Synthesis -- 7 Conclusion -- References -- Author Index
  • 4 DAOS Data Model and I/O Interface -- 4.1 DAOS Data Model -- 4.2 DAOS POSIX Support -- 5 Performance -- 6 Conclusion -- References -- Cloud Platform Optimization for HPC -- Abstract -- 1 Introduction -- 2 Existing Gaps -- 3 Methods -- 3.1 Eliminate "Jitter" -- 3.2 Resource Virtualization -- 3.3 Software Ecosystem -- 4 Results -- 5 Future Work -- References -- Applications and Scheduling -- swGBDT: Efficient Gradient Boosted Decision Tree on Sunway Many-Core Processor -- 1 Introduction -- 2 Background -- 2.1 Sunway Many-Core Processor -- 2.2 Gradient Boosted Decision Tree -- 2.3 Challenges for Efficient GBDT Implementation on Sunway Processor -- 3 Methodology -- 3.1 Design Overview -- 3.2 Data Prefetching on CPE -- 3.3 Sort and Merge -- 4 Implementation -- 4.1 Processing Logic of Gain Computation -- 4.2 Processing Logic of Sorting -- 4.3 Synchronization Among CPEs -- 5 Evaluation -- 5.1 Experiment Setup -- 5.2 Performance Analysis -- 5.3 Roofline Model -- 5.4 Scalability -- 6 Related Work -- 6.1 Acceleration for Gradient Boosted Decision Tree -- 6.2 Machine Learning on Sunway Architecture -- 7 Conclusion and Future Work -- References -- Numerical Simulations of Serrated Propellers to Reduce Noise -- Abstract -- 1 Introduction -- 2 Numerical Setup -- 2.1 Aerodynamic Solver -- 2.2 Force Calculations -- 2.3 Solver Validation -- 2.4 Acoustic Solver -- 2.5 Simulation Setup and Grid Convergence Study -- 3 Experimental Setup -- 4 Methodology -- 4.1 Initial Baseline Propeller Selection and Serrated, Cut-off Propeller Design -- 5 Results and Discussions -- 5.1 Force Comparison -- 5.2 Flow Visualizations -- 5.3 Acoustic Analysis -- 6 Conclusions and Recommendations -- References -- High-Performance Computing in Maritime and Offshore Applications -- 1 Introduction -- 2 Digital Twin of a Large-Scale Wave Basin
  • Intro -- Preface -- Organization -- Contents -- File Systems, Storage and Communication -- A BeeGFS-Based Caching File System for Data-Intensive Parallel Computing -- Abstract -- 1 Introduction -- 2 Background and Related Work -- 2.1 Burst Buffer Overview -- 2.2 Uniform Storage Systems for HPC Storage Hierarchy -- 2.3 Parallel File System Overview -- 3 Design -- 3.1 Uniform Namespace -- 3.2 Caching Model -- 3.3 Data Consistency -- 3.4 Data Movement -- 4 Implementation -- 4.1 Data Distribution -- 4.2 Meta Servers -- 4.3 Storage Servers -- 4.4 Cache Eviction -- 5 Performance Evaluation -- 5.1 Meta-Data Performance -- 5.2 Data Performance -- 6 Conclusions -- Acknowledgements -- References -- Multiple HPC Environments-Aware Container Image Configuration Workflow for Large-Scale All-to-All Protein-Protein Docking Calculations -- 1 Introduction -- 2 Background -- 2.1 Containers for HPC Environment -- 2.2 Problems of Container Image Configuration Workflow -- 3 HPC Container Maker (HPCCM) Framework -- 4 Container Deployment Workflow for MEGADOCK Application Using HPC Container Maker -- 4.1 MEGADOCK: A High Performance All-to-All Protein-Protein Docking Application -- 4.2 HPC Container Workflow for MEGADOCK with HPCCM -- 4.3 Example of User Workflow -- 5 Evaluation Experiments -- 5.1 Experiment 1. Container Deployment for Target HPC Environment -- 5.2 Experiment 2. Performance Evaluation with Large-Scale Computing Nodes and over a Million Protein-Protein Pairs -- 6 Discussion -- 7 Conclusion -- References -- DAOS: A Scale-Out High Performance Storage Stack for Storage Class Memory -- Abstract -- 1 Introduction -- 2 Constraints of Using Traditional Parallel Filesystems -- 3 DAOS, a Storage Stack Built for SCM and NVMe Storage -- 3.1 DAOS System Architecture -- 3.2 DAOS I/O Service -- 3.3 Data Protection and Data Recovery