IWOMP 2022

University of Tennessee

IWOMP 2022 Program

IWOMP 2022 will be held in Chattanooga, TN, United States, at the University of Tennessee. Travel and Venue information.

Registration for onsite attendance has closed, but registration for online attendance is still open. Registration options include single-event passes for IWOMP and OpenMP tutorials as well as all-week passes that also include EuroMPI/USA. All passes are available for either on-site, in-person attendance, or online-only attendance.

IMPORTANT: Please make sure to register online. On-site registration during the workshop will not be possible.

The main IWOMP workshop will be held on September 27th-30th, 2022. All times EDT.

Proceedings: Access the Proceedings

Tuesday 27th September

Start Time Presentation
12:00pm Lunch – 90 mins
OpenMP Tutorial I
1:30pm Utilizing LLVM/OpenMP to the Fullest: Tips & Tricks for
Application Developers

Johannes Doerfert
3:00pm Break – 30 mins
3:30pm OpenMP Tutorial 1 continues
5:00pm End of Day

Wednesday 28th September

Start Time Presentation
OpenMP Tutorial II
9:00am Partitioned Communication
Presenter: tbd
10:00am BREAK – 30 mins
10:30am OpenMP Tutorial II continues
12:00pm Lunch – 90 mins
Opening Session
1:30pm Introduction and Welcome
Bronis R. de Supinski
Invited Talk I
2:00pm OpenMP and MPI at Exascale – Early Experiences on Frontier
Al Geist
3:00pm Break – 30 mins
Paper Session: OpenMP and Multiple Nodes (Chair: Vivek Kale)
3:30pm Enhancing MPI+OpenMP Task based Applications for Heterogenous Architectures with GPU support
Manuel Ferat, Romain Pereira, Adrien Roussel, Patrick Carribault, Luiz-Angelo Steffenel and Thierry Gautier
4:00pm Towards Efficient Remote OpenMP Offloading
Wenbin Lu, Baodi Shan, Eric Raut, Jie Meng, Mauricio Araya-Polo, Johannes Doerfert, Abid M. Malik and Barbara Chapman
4:30pm End of Day

Thursday 29th September

Start Time Presentation
Paper Session: Exploring New and Recent OpenMP Extensions (Chair: Deepak Eachempati)
9:00am Characterizing the Performance of Task Reductions in OpenMP 5.X Implementations
Jan Ciesko and Stephen Olivier
9:30am Using Off-the-Shelf Hardware Transactional Memory to Implement Speculative While in OpenMP
Juan Salamanca and Alexandro Baldassin
10:00am Break – 30 mins
Paper Session: Effective Use of Advanced Heterogenous Node Architectures
10:30am Towards Automatic OpenMP-Aware Utilization of Fast GPU Memory
Delaram Talaashrafi, Marc Moreno Maza and Johannes Doerfert
11:00am Feasibility Studies in Multi-GPU Target Offloading
Anton Rydahl, Mathias Gammelmark and Sven Karlsson
11:30am Reducing OpenMP to FPGA Round-trip Times with Predictive Modelling
Julian Brandner, Florian Mayer and Michael Philippsen
12:00pm Lunch
Vendor Session I
1:30pm Intel
Xinmin Tian
2:15pm AMD
Gregory Rodgers
3:00pm Break – 30 mins
Paper Session: OpenMP Tool Support (Chair: Sara Royuela)
3:30pm Improving Tool Support for Nested Parallel Regions with Introspection Consistency
Vladimir Indic and John Mellor-Crummey
4:00pm On the migration of OpenACC-based applications into OpenMP 5+
Harald Servat, Giacomo Rossi, Alejandro Duran and Ravi Narayanaswamy
4:30pm End of Day

Friday 30th September

Start Time Presentation
Invited Talk II (Chair: Harald Servat)
9:00am Critical real-time systems: A new horizon for OpenMP
Sara Royuela
10:00am Break – 30 mins
Vendor Session II
10:30am Hewlett Packard Enterprise
Deepak Eachempati
11:15am clang/LLVM community
Johannes Doerfert
12:00pm Lunch
Paper Session: OpenMP and Multiple Translation Units
1:30pm Just-in-Time Compilation and Link Time Optimization for OpenMP Target Offloading
Shilei Tian, Joseph Huber, John Tramm, Barbara Chapman and Johannes Doerfert
2:00pm Extending OpenMP to Support Automated Function Specialization Across Translation Units
Giorgis Georgakoudis, Thomas Scogland, Chunhua Liao and Bronis de Supinski
Closing session
2:30pm Best Paper Award and Concluding Remarks
Bronis R. de Supinski