NERSCPowering Scientific Discovery Since 1974

NUG 2020

NUG 2020 focused on the upcoming Perlmutter system its capabilities, how NERSC is preparing its facility to support the system, resources for helping get your applications running on the Phase I's GPUs, plans for software and services, how and when users will get access, and system features to support AI/learning and data analysis of experimental/observational data.

Agenda - Monday, August 17, 2020

Pacific Time Eastern Time Topic Presenter(s) / Discussion Leader(s)
8:30 - 9:00 11:30 - 12:00 Welcome and set up and test Zoom connections  Alica Clum & Balint Joo NUGEX Chair and Co-Chair; Richard Gerber, Rebecca Hartman-Baker, NERSC
9:00 - 9:30 12:00 - 12:30 View from Washington Ben Brown, ASCR Facilities Director
Carol Hawk, ASCR, NERSC Program Manager
9:30 - 10:00 12:30 - 1:00 NERSC year in review, plans and Perlmutter overview Sudip Dosanjh, NERSC Director
10:00 - 10:20 1:00 - 1:20 Break/Lunch (20 minutes)  
10:20 - 10:50  1:20 - 1:50 The Perlmutter system, installation, and site preparation Tina Declerck, NERSC Systems Department Head; Doug Jacobsen, NERSC Systems Group Lead, Ben Maxwell, NERSC Building Infrastructure Group Lead
10:50 - 11:10 1:50 - 2:10 Perlmutter access, allocations and user environment Rebecca Hartman-Baker, NERSC User Engagement Lead
11:10 - 11:40 2:10 - 2:40 NESAP: Progress preparing applications for GPUs and lessons learned Jack Deslippe, NERSC Application Performance Lead
11:40 - 12:00 2:40 - 3:00 Break/Lunch (20 minutes)  
12:00 - 12:30 3:00 - 3:30 Superfacility and gateways for experimental and observational data Debbie Bard, NERSC Superfacility Lead; Cory Snavely, NERSC Infrastructure Services Lead
12:30 - 1:30 3:30 - 4:30 Breakout 1: NUGEX SIG: Experimental Science Users David Lawrence, NUGEX member, Jefferson Lab
Breakout 2: Application performance and portability across centers and architectures Brian Friesen, NERSC Application Performance Specialist
Breakout 3: Submitting a request for 2021 NERSC computer time and data storage; IRIS tricks and tips Clayton Bagwell, NERSC Account Support
Breakout 4: AI and Deep Learning on Perlmutter Wahid Bhimji, NERSC Big Data Architect
1:30 - 2:45 4:30 - 5:45 Contributed lightning talks  
2:45 - 3:00 5:45 - 6:00 Break  
3:00 - 3:30 6:00 - 6:30 NERSC operational metrics report Steve Leak, NERSC User Engagement
3:30 - 4:00 6:30 - 7:00 Open Discussion with NUGEX and NERSC  
4:00 7:00 Adjourn  

NUGEX SIG: Experimental Science Users Breakout Agenda

David Lawrence Introduction/Overview
David Lawrence GlueX at Jefferson Lab Data Processing w/ NERSC
Stephen Bailey Dark Energy Spectroscopic Instrument (DESI)
Michael Poat STAR
Jeff Porter ALICE
Chris O'Grady LCLS Realtime Analysis at NERSC
Bryce Foster JGI Workloads
David Lawrence Wrapup

Contributed Lightning Talks

Peter Ercius, LBL Analysis of 4D-STEM Data from a 480 Gbit/s Electron Detector
Matthew Ryder, ORNL Response Properties of Highly Porous Materials: Thermal, Pressure, and Electrical Response
Eric Stern, FNAL Refactoring the Synergia Accelerator Modeling Code with Kokkos to Achieve Perlmutter Readiness
Oisin Creaner, LBL Opticks on CoriGPU
Nikhel Gupta, U of Melbourne Mass Estimation of Galaxy Clusters with Deep Learning
Xueqiao Xu, LLNL Applying Microscopic Physics to the Macroscale through Deep Learning
Rose Yu, UCSD Towards Physics-informed Deep Learning for Turbulent Flow Prediction