IWOMP 2023 Program

The following OpenMP events are being hosted at the University of Bristol.

  • EuroMPI – 11-13 Sept.  See eurompi23.github.io for program details
  • IWOMP – 12-15 Sept.  See program below
  • MPI Forum Meeting – 14-15 Sept.  See www.mpi-forum.org for details

IWOMP 2023 Proceedings

Free access for conference attendees will be granted for 4 weeks until October 6, 2023. Attendees should use the password provided at the conference to access the IWOMP 2023 Proceedings.

Tuesday 12 September

08:30 Registration and Refreshments
09:00 Tutorial: Hands-On Using OpenMP Offload (Part 1)
Tom Deakin and Wei-Chen (Tom) Lin, University of Bristol.
View Presentation Slides (Parts 1-4)
10:30 Morning Break & Networking
11:00 Tutorial: Hands-On Using OpenMP Offload (Part 2)
12:30 Lunch & Networking
13:30 Tutorial: Hands-On Using OpenMP Offload (Part 3)
15:00 Afternoon Break & Networking
13:30 Tutorial: Hands-On Using OpenMP Offload (Part 4)
17:00 Close

Wednesday 13 September

08:30 Registration and Refreshments
09:00 Tutorial: Advanced OpenMP (Part 1)
View Presentation Slides (Parts 1 and 2)
10:30 Morning Break & Networking
11:00 Tutorial: Advanced OpenMP (Part 2)
12:30 Lunch & Networking
13:30 Welcome and Introductions
14:00: Keynote Presentation: Joint EuroMPI/IWOMP
Professor David May. Emeritus Professor, Department of Computer Science at the University of Bristol and a Fellow of The Royal Society. Wikipedia biography
15:00 Afternoon Break & Networking
Papers I: OpenMP Offload Experiences
Session Chair: Simon McIntosh-Smith
15:30 OpenMP Target Offload Utilizing GPU Shared Memory
Mathias Gammelmark, Anton Rydahl and Sven Karlsson [#07]
16:00 Improving a Multigrid Poisson Solver with Peer-to-Peer Communication and Task Dependencies
Mathias Gammelmark, Anton Rydahl and Sven Karlsson. [#11]
16:30 The Kokkos OpenMPTarget Backend: Implementation and Lessons Learned
Stephen Olivier, Rahulkumar Gayatri, Christian Trott, Johannes Doerfert, Jan Ciesko and Damien Lebrun-Grandie. [#13]
17:00 Close
19:00 Conference Dinner
Joint dinner for attendees of IWOMP, EuroMPI and MPI Forum Attendees. See Travel Information for location.

Thursday 14 September

8:30 Refreshments
09:00 Welcome and Introduction
09:15 Invited Talk: El Capitan: The First NNSA Exascale System
Bronis R. de Supinski, Lawrence Livermore National Laboratory (LLNL)

View Abstract

Livermore Computing (LC), Lawrence Livermore National Laboratory’s (LLNL’s) supercomputing center, and HPE are deploying the first US exascale system focused on national security. This talk will provide an overview of the preparations for LC’s first exascale system, as well as details of its system architecture. Throughout, the talk will explore considerations for energy efficiency in large-scale systems.

View Biography

As Chief Technology Officer (CTO) for Livermore Computing (LC) at Lawrence Livermore National Laboratory (LLNL), Bronis R. de Supinski formulates LLNL’s large-scale computing strategy and oversees its implementation. He frequently interacts with supercomputing leaders and oversees many collaborations with industry and academia. He has served as the Chair of the OpenMP Language Committee since 2009. In addition to his work with LLNL, Bronis is also a Professor of Exascale Computing at Queen’s University of Belfast. He is a Fellow of the ACM and the IEEE.

10:30 Morning Break & Networking
Papers II: OpenMP Infrastructure and Evaluation
Session Chair: Chunhua Liao
11:00 Improving Simulations of Task-Based Applications on Complex NUMA Architectures
Samuel Thibault, Idriss Daoudi, Thierry Gautier and Swann Perarnau. [#08]
11:30 Experimental Characterization of OpenMP Offloading Memory Operations and Unified Shared Memory Support
Wael Elwasif. [#19]
12:00 OpenMP Reverse Offloading Using Shared Memory Remote Procedure Calls
Jonathan Chesterfield and Joseph Huber. [#12]
12:30 Lunch & Networking
Papers III: Tasking Extensions
Session Chair: Stephen Olivier
13:30 How to Efficiently Parallelize Irregular DOACROSS Loops Using Fine Granularity and OpenMP Tasks: The SPEC mcf Case
Juan Salamanca and Alexandro Baldassin. [#16]
14:00 Introducing Moldable Task in OpenMP
Pierre-Etienne Polet, Ramy Fantar and Thierry Gautier . [#14]
14:30 Suspending OpenMP Tasks on Asynchronous Events: Extending the Taskwait Construct
Maël Martin, Romain Pereira, Adrien Roussel, Patrick Carribault and Thierry Gautier. [#03]
15:00 Afternoon Break & Networking
Implementer Updates
Session Chair: Tom Lin
15:30 GCC – Andrew Stubbs, Siemens
15:55 clang – Johannes Doerfert, LLVN
16:20 AMD Compiler for CPU and GPU – Greg Rodgers, AMD
16:45 Closing Remarks
17:00 Close

Friday 15 September

08:30 Refreshments
09:00 Welcome and Introduction
09:15 Invited Talk: Having it all: Can software be portable, performant and productive?
Dr Chris Maynard, Met Office.

View Abstract

Weather and climate models simulate complex, multi-scale physics. They can deliver substantial value for society utilising their predictive power. However, It can take a decade or more to develop a new model. Moreover, computer architectures and the programming models used to develop the software have evolved on a much shorter timescale and become more complex. How can large science applications achieve the three P’s? In this presentation the Domain Specific Language (DSL) approach being adopted by the Met Office in developing its next generation model, LFRic, will be discussed. How OpenMP is used to on modern CPUs and how it being employed to exploit the computational power of GPUs described. To conclude, some of the implications of the use of DSLs for parallelism, for general purpose languages and programming model specifications such as OpenMP will be considered.

View Biography

Dr Chris Maynard joined the Met Office in 2012 as a senior scientific software engineer. He was promoted to expert in 2013 and has led the software development of the Met Office next generation model, LFRic. Prior to 2012, at the Edinburgh Parallel Computing Centre (EPCC) he developed and optimised scientific software for supercomputers in diverse scientific fields. He obtained his PhD in 1998 from the University of Edinburgh in theoretical particle physics. His research interests include performance optimisation, parallel programming models, languages and algorithms and from 2018 until 2023 he was jointly Associate Professor in Computer Science at the University of Reading. Since August 2022 Dr Maynard has been a Met Office fellow and leads the performance optimisation of the next generation modelling system, Momentum.

10:30 Morning Break & Networking
Papers IV: Beyond Explicit GPU Support
Session Chair: Tom Scogland
11:00 Multipurpose Cacheing to accelerate OpenMP Target Regions on FPGAs
Julian Brandner, Florian Mayer and Michael Philippsen. [#06]
11:30 Generalizing Hierarchical Parallelism
Michael Kruse. [#05]
12:00 Exploring the Limits of Generic Code Execution on GPUs via Direct (OpenMP) Offload
Johannes Doerfert, Shilei Tian and Barbara Chapman. [#01]
12:30 Lunch & Networking
Papers V: OpenMP and AI
Session Chair: Johannes Doerfert
13:30 OpenMP Advisor: A Compiler Tool for Heterogeneous Architectures
Rolando Enriquez, Alok Mishra, Abid M. Malik, Meifeng Lin and Barbara Chapman. [#02]
14:00 Towards Effective Language Model Application in High-Performance Computing
Chunhua Liao, Le Chen, Pei-Hung Lin, Tristan Vanderbruggen, Murali Emani and Bronis R. de Supinski. [#18]
14:30 Advising OpenMP Parallelization via a Graph-Based Approach with Transformers
Tal Kadosh, Gal Oren, Nadav Schneider, Niranjan Hasabnis, Timothy Mattson and Yuval Pinter. [#04]
15:00 Close