IWOMP 2021 – Program
Due to the COVID-19 pandemic IWOMP 2021 was completely virtual. All the sessions were presented live. The start times for each session are shown in US Pacific Standard Time (PST) and European Central Time (CET). All speakers and delegates were required to Register to attend. Instructions for joining the online sessions were included in the booking confirmation. In the details below, Speakers are marked in Bold.
Best Paper Award
- A Case Study of LLVM-Based Analysis for Optimizing SIMD Code Generation
- Authors: Joseph Huber, Weile Wei, Giorgis Georgakoudis, Johannes Doerfert and Oscar Hernandez
Which was the highest ranked paper by the reviewers.
Proceedings
- Attendees can download the proceedings from: https://www.iwomp.org/proceedings/download/
- Password required. See your Eventbrite registration page for the password.
Monday 13 September | OpenMPCon
- See OpenMPCon for details
Tuesday 14 September
| Start Time | Presentation |
|---|---|
| Keynote:
Productivity, performance and portability: towards cross-domain DSL compiler architecture Prof. Paul Kelly, Imperial College London Domain-specific languages enable us to automate generation of high-performance code from a high-level abstraction. This talk will show, through a couple of example projects (Firedrake and Devito) that DSLs can deliver productivity, performance, and performance-portability. The key to success is compiler architecture – designing intermediate representations that making optimisations easy and analysis trivial. But the DSL software ecosystem is dysfunctional: DSL compilers (including ours) are typically standalone projects, reliant on support from a narrow developer base. Few, if any, components are shared between DSLs. The talk will conclude with a manifesto for fixing this – building on MLIR to establish community support for code generation tools that underpin multiple front-end DSLs. I will argue that this is in fact the only way we can tackle the complexity involved in achieving high performance for complex applications on diverse hardware. Biography: Paul Kelly leads the Software Performance Optimisation group at Imperial College London. His research focus is domain-specific program optimisation, leading to close engagement with colleagues in computational science, robotics and computer vision. This talk covers joint work with many such collaborators. |
|
| BREAK – 30 mins | |
| Papers Session I: Synchronization and Data | |
| View Slides | Improving Speculative taskloop in Hardware Transactional Memory
Juan Salamanca and Alexandro Baldassin |
| Vectorized Barrier and Reduction in LLVM OpenMP Runtime
Muhammad Nufail Farooqi and Miquel Pericàs |
|
| BREAK – 30 mins | |
| Papers Session II: Tasking Extensions I | |
| View Slides | Enhancing OpenMP Tasking Model: Performance and Portability
Chenle Yu, Sara Royuela and Eduardo Quiñones |
| View Slides | OpenMP Taskloop Dependences
Marcos Maroñas, Xavier Teruel and Vicenç Beltran |
| Day 1 Concludes | |
Wednesday 15 September
| Start Time | Presentation |
|---|---|
| Papers Session III: Applications | |
| Outcomes of OpenMP Hackathon: OpenMP Application Experiences with the Offloading Model
Swaroop Pophale, Barbara Chapman, Buu Pham, Charlene Yang, Christopher Daley, Colleen Bertoni, Dhruva Kulkarni, Dhruva Kulkarni, Dossay Oryspayev, Ed D’Azevedo, Helen He, Johannes Doerfert, Keren Zhou, Kiran Ravikumar, Mark Gordon, Mauro Del Ben, Meifeng Lin, Melisa Alkan, Michael Kruse, Oscar Hernandez, P. K. Yeung, Paul Lin, Peng Xu, Tosaporn Sattasathuchana, Vivek Kale, William Huhn and Dhruva Kulkarni |
|
| View Slides | An empirical investigation of OpenMP based implementation of Simplex Algorithm
Arkaprabha Banerjee, Pratvi Shah, Shivani Nandani, Shantanu Tyagi, Sidharth Kumar and Bhaskar Chaudhury |
| View Slides | Task inefficiency patterns for a wave equation solver
Holger Schulz, Gonzalo Brito Gadeschi, Oleksandr Rudyy and Tobias Weinzierl |
| BREAK – 30 mins | |
| Papers Session IV: Case Studies | |
| View Slides | Comparing OpenMP Implementations With Applications Across A64FX Platforms
Benjamin Michalowicz, Eric Raut, Yan Kang, Tony Curtis, Dossay Oryspayev and Barbara Chapman Save |
| View Slides | A Case Study of LLVM-Based Analysis for Optimizing SIMD Code Generation
Joseph Huber, Weile Wei, Giorgis Georgakoudis, Johannes Doerfert and Oscar Hernandez |
| Day 2 Concludes | |
Thursday 16 September
| Start Time | Presentation |
|---|---|
| Papers Session V: Heterogenous Computing & Memory | |
| View Slides | Experience Report: Writing A Portable GPU Runtime with OpenMP 5.1
Shilei Tian, Jon Chesterfield, Johannes Doerfert and Barbara Chapman |
| View Slides | FOTV: A generic device offloading framework for OpenMP
Jose Luis Vázquez and Pablo Sanchez |
| View Slides | Beyond Explicit Transfers: Shared and Managed Memory in OpenMP
Brandon Neth, Thomas Scogland, Alejando Duran, Michelle Mills Strout and Bronis R. de Supinski |
| BREAK – 30 mins | |
| Papers Session VI: Tasking Extensions II | |
| View Slides | Communication-Aware Task Scheduling Strategy in Hybrid MPI+OpenMP Applications
Romain Pereira, Adrien Roussel, Patrick Carribault and Thierry Gautier |
| View Slides | An OpenMP Free Agent threads implementation
Victor Lopez, Joel Criado, Raúl Peñacoba, Roger Ferrer, Xavier Teruel and Marta Garcia-Gasulla |
| IWOMP 2021 Concludes | |