@inproceedings{87f343acd69c469abc6830b54fe5c45c,
title = "Elasticity in a Task-based Dataflow Runtime Through Inter-node GPU Work Stealing",
abstract = "Most contemporary HPC programming models assume an inelastic runtime in which the resources allocated to an application remain fixed throughout its execution. Conversely, elastic runtimes can expand and shrink resources based on availability and/or dynamic application requirements. In this paper, we implement elasticity for PaRSEC, a task-based dataflow runtime, using inter-node GPU work stealing. In addition to supporting elasticity, we demonstrate that inter-node GPU work stealing can enhance the performance of imbalanced applications by up to 45%.",
keywords = "Dataflow, Distributed Work Stealing, Elastic computing, Malleable computing, PaRSEC, Task-based programming",
author = "Joseph John and Josh Milthorpe",
note = "Publisher Copyright: {\textcopyright} 2024 IEEE.; 24th IEEE/ACM International Symposium on Cluster, Cloud and Internet Computing, CCGrid 2024 ; Conference date: 06-05-2024 Through 09-05-2024",
year = "2024",
month = oct,
day = "8",
doi = "10.1109/CCGrid59990.2024.00020",
language = "English",
isbn = "979-8-3503-9567-9",
series = "Proceedings - IEEE International Symposium on Cluster, Cloud and Internet Computing CCGrid",
publisher = "Institute of Electrical and Electronics Engineers Inc.",
number = "24",
pages = "97--105",
booktitle = "Proceedings of the 2024 IEEE 24th International Symposium on Cluster, Cloud and Internet Computing, CCGrid 2024",
address = "United States",
}