Selected Publications
Balancing Tracking Granularity and Parallelism in Many-Task Systems: The Horizons Approach
Peter Thoman, Philip Salzmann
SN Computer Science, 2024
Automatic Discovery of Collective Communication Patterns in Parallelized Task Graphs
Fabian Knorr, Philip Salzmann, Peter Thoman, and Thomas Fahringer
International Journal of Parallel Programming, 2024
Command Horizons: Coalescing Data Dependencies while Maintaining Asynchronicity
Peter Thoman, Philip Salzmann
WAMTA 2023
An Asynchronous Dataflow-Driven Execution Model For Distributed Accelerator Computing
Philip Salzmann, Fabian Knorr, Peter Thoman, Philipp Gschwandtner, Biagio Cosenza, and Thomas Fahringer
CCGRID 2023
Declarative Data Flow in a Graph-Based Distributed Memory Runtime System
Fabian Knorr, Peter Thoman, Thomas Fahringer
International Journal of Parallel Programming, 2022
The Celerity High-level API: C++20 for Accelerator Clusters
Peter Thoman, Florian Tischler, Philip Salzmann, and Thomas Fahringer
HLPP 2021
Porting Real-World Applications to GPU Clusters: A Celerity and CRONOS Case Study
Philipp Gschwandtner, Ralf Kissmann, David Huber, Philip Salzmann, Fabian Knorr, Peter Thoman, and Thomas Fahringer
eScience 2021
Celerity: High-level C++ for Accelerator Clusters
Peter Thoman, Philip Salzmann, Biagio Cosenza, Thomas Fahringer
Euro-Par 2019
Selected Talks & Demos
A High-Level API for Dynamic Load Balancing in Large-Scale Parameter Sweeps
Philip Salzmann
HLPP 2024
SYCL and Celerity: High(-ish) level, vendor-independent C++ for GPU parallelism
Peter Thoman
Guest Lecture, Trento 2023
VideoAutomatic Discovery of Collective Communication Patterns in Parallelized Task Graphs
Fabian Knorr
HLPP 2023
SYCL Panel Discussion
Peter Thoman
IWOCL/SYCLcon 2021
Celerity — High-Level Distributed Accelerator C++ Programming
Philipp Gschwandtner
Talk at AHPC 2020
Celerity: High-productivity Programming for Accelerator Clusters
Peter Thoman
Talk at ScalPerf 2019
Introducing Celerity: High-level C++ for Accelerator Clusters
Philip Salzmann
Demo session at HPCS 2019
Acknowledgements
This project has received funding from the FFG (Austrian Research Promotion Agency) under grant agreement FO999903595.
This project has received funding from the European High-Performance Computing Joint Undertaking (JU) under grant agreement No 956137.
This research has been partially funded by the FWF (I 3388) and DFG (CO 1544/1-1, project number 360291326) as part of the CELERITY project.