2020 International Workshop on Performance, Portability & Productivity in HPC

Program

The P3HPC program will be live-streamed on the virtual SC20 platform, on Friday 13 November 2020. All times below are in Eastern.

10:00 - 10:10
Welcome and Introduction
Rob Neely, Lawrence Livermore National Laboratory


Keynote

10:10 - 11:10
ECP: Lessons Learned in Porting Complex Applications to Accelerator-based Systems
Andrew Siegel, Argonne National Laboratory


Session 1

11:10 - 11:40
Tracking Performance Portability on the Yellow Brick Road to Exascale
Tom Deakin, University of Bristol


11:40 - 12:10
Performance Portability of Molecular Docking Miniapp on Leadership Computing Platforms
Mathialakan Thavappiragasam, Oak Ridge National Laboratory


Break

12:10 - 12:40


Session 2

12:40 - 12:55
P3HPC Forum Recap
Scott Parker, Argonne National Laboratory


12:55 - 13:30
OpenMP: State of the Standard
Tom Scogland, Lawrence Livermore National Laboratory


13:30 - 14:00
Interpreting and Visualizing Performance Portability Metrics
Jason Sewall, Intel Corporation


Break

14:00 - 14:30


Session 3

14:30 - 15:10
Findings from the ECP Performance Portability Panel Series
Anshu Dubey, Argonne National Laboratory


15:10 - 15:40
Mini-Panel with ECP Representatives
Anshu Dubey, Katherine Riley, Jack Deslippe, David Richards, Brian VanEssen, Bronson Messer, Shreyas Anathan


15:40 - 16:55
Workflows are the New Applications: Challenges in Performance, Portability, and Productivity
C.J. Newburn, Daisy Hollman, Todd Gamblin, Harinarayan Krishnan, Tal Ben-Nun


Break

16:55 - 17:25


Session 4

17:25 - 17:55
Cross-Platform Performance Portability of DNN Models using SYCL
Mehdi Goli, Codeplay Software Ltd


17:55 - 18:25
Evaluating the Performance Portability of Contemporary SYCL Implementations
Beau Johnston, Oak Ridge National Laboratory


18:25 - 18:30
Wrapup
John Pennycook, Intel Corporation