Program Overview
2019 Exascale Computing Project Annual Meeting
The goal of the 2019 ECP Annual Meeting is to highlight technical accomplishments that are being enabled by interactions and collaborations within the ECP community, including the ECP focus areas, DOE HPC facilities, industry and vendors. In addition to the Keynote and Plenary Sessions, the core of the program will include Breakouts, Tutorials and Posters Sessions.
Program details and final daily agenda are available by clicking here.  |
 |
 |
 |
=  |
| Monday Agenda (PDF) | Tuesday Agenda (PDF) | Wednesday Agenda (PDF) | Thursday Agenda (PDF) |
Side Meetings (Mon-Fri) - Individual or Group Meetings
Breakout rooms will be available from Monday, January 14 to Friday, February 18 for use by teams to hold side meetings. Side Meeting discussions can be used for individual project teams to hold organizational or technical interactions, or for multiple teams to work together on integration activities, such as shared fate milestones. The number of rooms and their size will vary depending on the day.Plenaries
● Continuous Integration – A Status Update on Newly Developed HPC Features, and Goals for Deployment● Exascale Deep Learning
● Update from the Facilities
Breakout Sessions
The 2019 ECP Annual Meeting offers a diverse set of 45-minute, 60-minute and 90-minute breakout sessions that will take place Monday, Tuesday, Wednesday, and Thursday of the Annual Meeting. Participants are encouraged to attend breakouts to learn more about these important topics and areas in the ECP community. Presenters will be discussing accomplishments, requirements sharing, and opportunities for interactions. Details including dates and times of the sessions will follow. See list of Breakout Sessions below:● AMReX and AMReX-based Applications
● Architectural Challenges of Genomic Analysis
● CEED: High-Order Methods and Applications
● Checkpointing from Application Perspective: Use Cases, Software Integration, Challenges
● Containers in HPC, and Beyond
● Contributing to Upstream Software: Why and How
● Cray Performance Tools Deep Dive (Cray - NDA Session)
● Demonstrating new hardware capabilities with ECP applications (IBM - NDA Session)
● ECP Continuous Integration Framework - getting started
● ECP Hardware Evaluation – Assessing the Characteristics of Exascale Class Machines
● Fostering a culture of passion and productivity in ECP teams
● From Good to Great: Getting Started with PSIP (Productivity Sustainability Improvement Planning)
● Implications of Exascale Hardware Trends on Application Development
● Jupyter at High Performance Computing Centers
● Kokkos Applications: Lessons Learned
● Kokkos EcoSystem 3.0 Roadmap
● Lessons Learned from National Security Applications
● Lossy data reduction/compression for simulations and experiments
● Making Agile Work for Your R&D Software Project
● Management of Complex, Heterogeneous & Hierarchical Memory Resources
● Modeling the Coupling of Climate, Urban Microclimate, and the Built Environment
● MPI Breakout
● Online Data Analysis and Reduction Software
● OpenMP BoF at ECP
● PathForward Vendor Presentation - AMD (NDA Session) + Repeat
● PathForward Vendor Presentation - Cray (NDA Session) + Repeat
● PathForward Vendor Presentation - HPE (NDA Session) + Repeat
● PathForward Vendor Presentation - IBM (NDA Session) + Repeat
● PathForward Vendor Presentation - Intel (NDA Session) + Repeat
● PathForward Vendor Presentation - NVIDIA (NDA Session) + Repeat
● Performance Portability at the Intersection of the Exascale Computing Project and the DOE Centers of Excellence
● Perlmutter - a 2020 pre-exascale GPU-accelerated system for NERSC.
● Portable, High-Performance C++ Programming Models for Heterogeneous Architectures
● Preparing Applications for the Argonne Aurora 2021 System
● Preparing Applications for NERSC-9 (Perlmutter) as an Exascale Waypoint
● Resource Management and Scheduling Challenges at Exascale
● Software Assurance and HPC
● Software Sustainability for ECP and Beyond
● Spack Roundtable
● Supporting Complex Workflows from Experimental and Observational Facilities
● System-level architecture simulation for exascale: challenges and opportunities
● The LLVM Compiler Infrastructure in ECP -- Usage, Updates, Challenges, and Improvements
● Toward Common and Interoperable Software Infrastructures
● Towards enabling applications to exploit SmartNIC/FPGA with integrated networking and acceleration
● Using the Flang/LLVM compiler for ECP Applications and Tools Development
● What's New in Proxy Applications
Tutorials
The ECP agenda includes an exciting series of 60-minute, 90-minute, 180-minute, and 360-minute Tutorials offered Monday through Thursday of the annual meeting. When planning your schedule for the ECP Annual Meeting, be sure to allow time for you and your teammates to attend some tutorials. See list of Tutorials below:● A Tutorial Introduction to RAJA
● Advanced Programming on Hybrid Computers in PaRSEC Using Symbolic Representation
● An Introduction to BEE Containerized Application Workflows
● An introduction to the xSDK, a community of diverse numerical HPC software packages
● Application-driven Fault-Tolerance for High Performance Distributed Computing
● Atlassian Confluence Fundamentals (Beginners to Intermediate)
● Atlassian Jira Fundamentals (Beginners to Intermediate)
● Better Scientific Software
● Boosting power efficiency of HPC applications with GEOPM
● CANDLE Tutorial
● Container Computing for HPC and Scientific Workloads
● Contemporary Peer Code Review in Scientific Software Development (Part 1) + (Part 2)
● Deep Learning at Scale
● ECP 101
● Exascale I/O Technologies
● Flux: Using Next-Generation Resource Management and Scheduling Infrastructure for Exascale Workflows
● Graph and Combinatorial Methods for Enabling Scientific Applications
● In Situ Visualization and Analysis with Ascent (Part 1) + (Part 2)
● Introduction to the Capabilities and Use of the SUNDIALS Suite of Nonlinear and Differential/Algebraic Equation Solvers
● Introduction to the Global Arrays Toolkit
● Kokkos Short Tutorial
● Modern CMake Tutorial
● NVBit: A Binary Instrumentation Tool for CUDA (NVIDIA - NDA Session)
● OpenMP 4.5 and 5.0
● PDEs, optimization, and eigenproblems with PETSc/TAO and SLEPc
● Performance Tuning with the Roofline Model on GPUs and CPUs
● SLATE and MAGMA: linear algebra for HPC
● Spack Tutorial
● Task-Based Parallel Programming in Legion
● TAU Performance System
● Umpire: Managing Heterogeneous Memory Resources
● UnifyCR - A file system for burst buffers
Posters
Application Development
Exascale-Ready Multi-Grid Algorithms for Lattice QCD on GPUs
Lattice QCD 2.2.1.01 ADSE03
Accelerating Large Sparse Matrix Inversion on Multi-Node GPU Systems with Local Preconditioning and Tensor Cores
Grid QCD code: Performance Portability
NWChemEx - Moving Computational Chemistry to the Exascale
Molecular Dynamics at the exascale with EXAALT
IDEAS-EXAALT Collaboration: Adopting Continuous Integration for Long-Timescale Materials Simulation
Exascale for Additive Manufacturing (ExaAM)
QMCPACK: Algorithm, Performance, and Software Improvements
ExaWind: Towards predictive wind plant simulations
Pele Suite of Codes for Transforming Combustion Science and Technology with Exascale Simulations
MFiX-Exa: Designing Next Generation Chemical Looping Reactors
WarpX: Exascale Modeling of Plasma Particle Accelerators
ExaSky: Computing the Sky at Extreme Scales
Towards Exascale Simulations of Regional Earthquake Hazard and Risk
Exa-SGD: Data Intensive Scalable Optimization for Power System Transmission Expansion
CANDLE: Exascale Deep Learning Enable Precision Medicine for Cancer
ExaBiome: Exascale Solutions to Microbiome Analysis
ExaFEL First Results
The MARBL Multi-physics Code
Coupled Urban Modeling: The Big Data Challenge in Exascale Computing
Multiscale Coupled Urban Systems Framework
CODAR Technologies
AMReX Co-Design Center for Exascale Computing
Center for Efficient Exascale Discretizations
ExaGraph: Graph and Combinatorial Methods for Enabling Exascale Applications
GAMESS ECP: Enabling GAMESS for Extreme Computing in Chemistry and Materials
E3SM-MMF Cloud Resolving Earth System Model
High Energy Physics Detector Simulation (Geant)
Proxy Apps Mysteries Revealed
Performance Analysis and Optimization Challenges in Scaling a Transportation Simulation
ExaLearn: Codesign Center for Exascale Machine Learning Technologies
CoPA: Co-design center for Particle Applications
Ristra: next-generation multi-physics
Towards core-edge coupled simulations for fusion plasmas: kinetic electrons, time telescoping and software framework
Exascale Models of Stellar Explosions: Quintessential Multi-Physics Simulation
Exascale Subsurface Simulator of Coupled Flow, Transport, Reactions, and Mechanics
ExaSMR: Coupled Monte Carlo Neutronics and Fluid Flow Simulations of Small Modular Reactors
SPEC: Scalable Predictive methods for Excitations and Correlated phenomena
Facilities
Interactive Supercomputing through Jupyter at NERSC
Oak Ridge Leadership Computing Facility (OLCF)
Argonne Leadership Computing Facility: Accelerating Discovery
ASTRA
Hardware & Integration
Node Level Simulation
Deciphering Exascale Hardware Trends for the Broader ECP Community
Potential of Quantum Accelerators
Scaling the exascale memory wall ... No trace left behind
Hardware Evaluation Interconnect Working Group
Analytical Modeling
Perlmutter: A Pre-Exascale Systems for the Office of Science
Application Integration at the Facilities
NERSC: HPC for the DOE Office of Science
Improving Tomographic Reconstruction Parallelism in TomoPy1
Optimizing Python data processing for the DESI experiment on the NERSC Cori supercomputer: A case study
Performance Portability: A case study with OpenMP 4.5
Optimization of ATLAS simulations on Cori
Use of HPC resources at NERSC by the LZ dark matter experiment
ECP/E6 Software Deployment – Facilities, Teams, Containers
A Healthy Software Ecosystem with Continuous Integration
Argonne Training Program on Extreme-Scale Computing 2019
Using Software Productivity and Sustainability Improvement Planning (PSIP) to Improve Application Development Productivity
User Stories: Communicating about Software Requirements
Better Scientific Software (BSSw): So our code will see the future
IDEAS-ECP: Advancing Software Productivity for Exascale Applications
Holistic Measurement Driven System Assessment (HMDSA)
Industry
Exascale Computing Enables the Future of Mobility
Application of HPC to Enhance Product Development at General Motors
Computing Challenges in Kinetic Simulation of FRC Stability and Transport
Design of Next Gen Industrial Products
Software Technology
VTK-m: Updating HPC Visualization Software for Exascale-Era Processors
Open MPI for Exascale (OMPI-X)
Legion and Multiphysics, LANL Advanced Technology Development and Mitigation – Programming Models and Runtimes
RajaChecker: A Customized Loop Pattern Recognition Tool
Kokkos Programming Model Update
Kokkos Ecosystem 3.0
xGA: New Features of the Global Arrays Toolkit
Integrated Software Components for Managing Computation and Memory Interplay at Exascale
Improving MPICH On-boarding Training Process using ECP IDEAS PSIPs
Exascale MPI (MPICH)
The Legion Programming System
Capturing Parallelism in LLVM
PaRSEC: A Runtime for Distributed Tasking at Exascale
PaRSEC: DSLs and Interfaces for ECP
Performance Portable Finite Volume Magnetohydrodynamics on Structured Grids
SOLLVE: OpenMP Programming Environment for ECP
SOLLVE: Validation and Verification of OpenMP 4.5+ Features
BOLT: Lightning-Fast OpenMP
Pagoda: Lightweight Communications and Global Address Space Support for Exascale Applications - GASNet-EX
Pagoda: Lightweight Communications and Global Address Space Support for Exascale Applications - UPC++
SICM: Simplified Interface to Complex Memory
Metall (meta allocator for persistent memory)
Ubiquitous Performance Analysis
Sandia ATDM DevOps and Performance Analysis
EXA-PAPI: The Exascale Performance Application Programming Interface
Y-Tune: Autotuning for Cross-Architecture Transformation and Code Generation
Extending HPCToolkit to Measure and Analyze Code Performance on GPU-accelerated Platforms
Recent Developments in HPCToolkit
Understanding Congestion in High Performance Interconnection Networks Using Sampling
Dynamic Data Race Detection for OpenMP Programs
STTO12-PROTEAS: Programming Toolchain for Emerging Architectures and System
Clacc: Translating OpenACC to OpenMP in Clang
The TAU Performance System
LLVM in ECP
Extreme-scale Scientific Software Development Kit (xSDK) for ECP
hypre - High Performance Linear Solvers and Preconditioners
LLNL ATDM Mathematical Libraries
Enabling Mission Simulations Via Next Generation Trilinos Solver Libraries
Hierarchical Parallelism for Performant Embedded Automatic Differentiation on GPUs
Enabling Time Integrators for Exascale Through SUNDIALS
PETSc/TAO Enhancements for Exascale
libEnsemble: A library for the concurrent evaluation of ensembles of computations
FFTX, a new framework for building high-performance FFT-based applications for exascale machines
STRUMPACK -- STRUctured Matrix PACKage
SuperLU Sparse Direct Solver
ForTrilinos: Object-oriented Fortran interfaces to Trilinos packages
SLATE: Software for Linear Algebra Targeting Exascale
FFT-ECP
Iterative Solvers and Preconditioners for extreme Parallelism
ALExa: Accelerated Libraries for Exascale
Cinema
The ATDM Workflow Project at LLNL
The Cognitive Simulation Project at Lawrence Livermore National Laboratory
SNL ATDM: I/O AND DATA MANAGEMENT
SNL ATDM: In-situ Compression with ParaView/TuckerMPI
2.3.4.04 SNL Data and Visualization: ML Projects as Sandia
VELOC: Very Low Overhead Checkpointing
UnifyCR: A File System for Burst Buffers
ExaHDF5: Delivering Efficient Parallel I/O on Exascale Computing Systems
The ADIOS framework for Scientific Data on Exascale Systems
DataLib: Data Libraries and Services Enabling Exascale Science
ZFP: A Compressed Array Primitive
ALPINE project: Infrastructure and Algorithms for ECP Science
ECP ST Software Development Kits (SDKs)
Extreme-Scale Scientific Software Stack (E4S)
Profiling Application Resilience with Decaf-FSEFI
BEE Workflows for HPC
Spack: a package manager for exascale
SNL ATDM Software Ecosystem
Argo: Improved Operating System and Resource Management in the Exascale Era
Argo: Resource Partitioning and Power Management in the Exascale Era
Argo: Management of Deep Memory Hierarchy in the Exascale Era
A Production-quality Fortran Front-end for LLVM
Runtime Techniques for Boosting Application Performance and Power Efficiency in the ECP Software Stack
Sonar: LLNL HPC Facility Performance Monitoring Infrastructure
Livermore Computing: A Legacy of Outstanding HPC Solutions
EZ: Exascale Lossy Compression for Scientific Data