Tools and Techniques for High Performance Computing: Selected Workshops, HUST, SE-HER and WIHPC, Held in Conjunction with SC 2019, Denver, CO, USA, ... in Computer and Information Science, 1190)

Preface
	Review Process
Organization
Contents
HUST - Annual Workshop on HPC User Support Tools
Buildtest: A Software Testing Framework with Module Operations for HPC Systems
	1 Introduction
	2 Buildtest
		2.1 Motivation
		2.2 Inception
		2.3 Framework
		2.4 Integration with Lmod Spider
		2.5 Setup
		2.6 Architecture Overview
		2.7 Building Test
		2.8 Building Job Scripts
	3 Module Operations
		3.1 Module Load Test
		3.2 Module Tree Difference for Parallel Software Stack in Heterogeneous Cluster
		3.3 Reporting Easybuild and Spack Modules
		3.4 User Collections and Buildtest Module Collections
		3.5 Module Permutation
		3.6 Report Unique Software and Modules
	4 Future Work
	5 Related Work
	6 Conclusion
	References
Using Malleable Task Scheduling to Accelerate Package Manager Installations
	1 Introduction
	2 Build Systems
		2.1 Build Systems and Spack Package Manager
	3 Parallel Builds and Task Scheduling
		3.1 Theory
		3.2 Prior Work
	4 Scheduling Algorithm
		4.1 Task Execution Time Heuristic
		4.2 CPR and MCPA Implementation
	5 Methodology
	6 Results and Discussion
		6.1 Schedule Creation Time
		6.2 Schedule Execution Time
		6.3 Scheduler Choice and DAG Size
		6.4 Improving Schedule Creation Times
	7 Future Work
		7.1 Task Cost Heuristic
		7.2 Package Fetching
		7.3 Hyperthreading
		7.4 Phase Tasks and Build System Tasks
	8 Conclusion
	References
Enabling Continuous Testing of HPC Systems Using ReFrame
	1 Introduction
	2 Framework Design
		2.1 The Regression Test Pipeline
		2.2 The Frontend
		2.3 Pluggable Backends
	3 Writing a Regression Test in ReFrame
	4 Configuring and Running ReFrame
		4.1 Performance Logging
		4.2 Running ReFrame
		4.3 Dealing with Test Failures
	5 Use Cases
		5.1 ReFrame at CSCS
		5.2 ReFrame at NERSC
		5.3 ReFrame at OSC
	6 Related Work
	7 Conclusions and Future Directions
	References
Tools for Monitoring CPU Usage and Affinity in Multicore Supercomputers
	1 Introduction
	2 Background
		2.1 Process and Thread Affinity
		2.2 Related Work
	3 Three Innovative Tools
		3.1 core_usage
		3.2 show_affinity
		3.3 amask
	4 Case Study
		4.1 Unexpected Slow VASP Runs
		4.2 MPI Library Evaluation on a New System
		4.3 Affinity Discovery
		4.4 General
	5 Best Practice
	6 Conclusion
	References
SE-HER - International Workshop on Software Engineering for HPC-Enabled Research
A Study of Hydrodynamics Based Community Codes in Astrophysics
	1 Introduction
	2 Background
	3 Methodology
	4 Analysis and Inferences
	5 Conclusions
	References
Lightweight Software Process Improvement Using Productivity  and Sustainability Improvement Planning (PSIP)
	1 Introduction
	2 Background
	3 Methodology
	4 Related Work
	5 Case Studies
		5.1 EXAALT
		5.2 Exascale MPI
	6 Discussion and Future Work
	7 Conclusion
	References
Role-Oriented Code Generation in an Engine for Solving Hyperbolic PDE Systems
	1 Introduction
	2 The ExaHyPE Engine
		2.1 A High-Order ADER-DG Solver with A-Posteriori Limiting
		2.2 Application-Specific Programming Interface
		2.3 Architecture-Aware Optimization of Kernels
	3 Code Generation in ExaHyPE
		3.1 Model-View-Controller Design
		3.2 Templates
		3.3 Architecture-Oblivious Templates and Architecture-Aware Optimization Macros
	4 Expanding the PDE: Navier-Stokes Equations
		4.1 Expanding the DSL
		4.2 Processing the New Specification File Option
		4.3 Expanding the Views
		4.4 Result Evaluation
	5 Improved Space-Time Predictor for Linear Applications
		5.1 Prototyping the New Algorithm
		5.2 Inclusion in the Kernel Generator
		5.3 Template Generalization and Optimization
		5.4 Performance Evaluation
	6 Vectorization of User Functions
		6.1 Optimized Transpose – from AoS to SoA (and Back)
		6.2 Abstracting the Call to the User Function Behind a Macro
		6.3 Expanding the callFlux Macro
		6.4 Performance Evaluation
	7 Conclusions
	References
FQL: An Extensible Feature Query Language and Toolkit on Searching Software Characteristics for HPC Applications
	1 Introduction
	2 Related Work
		2.1 Software Analysis Tools Without Using Domain-Specific Languages
		2.2 Software Analysis Tools Using Domain-Specific Languages
	3 The FQL Language and Toolkit
		3.1 Feature Query Language (FQL)
		3.2 Overall Workflow of the FQL Software Toolkit
		3.3 Implementation of the FQL Toolkit
	4 Exemplar Applications
	5 Conclusions
	References
WIHPC – Workshop on Interactive High-Performance Computing
Accelerating Experimental Science Using Jupyter and NERSC HPC
	1 Introduction
	2 NERSC Jupyter Infrastructure
	3 Scaling up Analyses on HPC
		3.1 NCEM Bragg Disk Detection with Py4DSTEM
		3.2 Transforming the Code for Parallel Execution
		3.3 Upgrading Disk Detection
		3.4 Integrating the Parallel Version Back into Py4DSTEM
		3.5 NCEM Impact
		3.6 Insights for Parallel Performance of Scientific Software
		3.7 Future Work
	4 Streamlining Exploration of Parameter Spaces
		4.1 Future Work
	5 Reproducible and Replicable Workflows
		5.1 Sharing and Cloning Jupyter Notebooks
		5.2 Future Work
	6 Conclusions
	References
Interactive Supercomputing for Experimental Data-Driven Workflows
	1 Introduction
	2 Background and Motivation
		2.1 Scientific IT Infrastructure Design Principles
		2.2 Dedicated vs. Shared Cluster Computing Resources
		2.3 Slurm Queues, Partitions and Fairshare
	3 Design and Scope of Data Driven Interactive Supercomputing
		3.1 Reservation Concepts and Use Cases
		3.2 Analysis of a Data-Driven Workflow
		3.3 Co-Designed Reservation Service for Data-Driven Workflow
		3.4 Controlled Privileges and Limitations
		3.5 Reservation Management Tool
	4 Implementation and Analysis
	5 Summary and Future Work
	References
Portals for Interactive Steering of HPC Workflows
	1 Introduction
	2 HPC Workflows
		2.1 DNA Monitoring
		2.2 Basic NGS QC
		2.3 RNASeq
		2.4 Weakly Supervised Deep Neural Network
	3 Conclusion and Future Work
	References
The Pangeo Ecosystem: Interactive Computing Tools for the Geosciences: Benchmarking on HPC
	1 Introduction
	2 Pangeo and the Pangeo Platform
	3 HPC Deployments of the Pangeo Platform
		3.1 Hal
		3.2 Cheyenne
	4 Benchmark of Pangeo on HPC
		4.1 Benchmark Method
		4.2 Results and Discussions
	5 Conclusion and Further
	References
Author Index