Synergically Rebalancing Parallel Execution via DCT and Turbo Boosting
Sandro M. Marques, Thiarles S. Medeiros, Fábio Diniz Rossi, Marcelo Caggiani Luizelli, Antonio Carlos Schneider Beck, Arthur Francisco Lorenzon
Abstract
The increasing use of cloud and HPC systems put more pressure on the efficient utilization of hardware resources to keep costs low. Many dynamic concurrency throttling (DCT) techniques have successfully used to tune the number of executing threads to better balance a parallel application according to its available scalability. Similarly, boosting frequency strategies have been used to speed up the sequential parts’ execution. Given that, we propose Poseidon, the first transparent and automatic approach that cooperatively exploits both techniques to rebalance OpenMP applications without any preprocessing, with no code transformation, recompilation, or OS modification.
Ask about this paper
Ask your agent about it.
Lune has read the top-tier papers around this one, so every answer names the papers it rests on.
Related papers
- Dynamic Detection of Inefficient Data Mapping Patterns in Heterogeneous OpenMP ApplicationsLuke Marzen, Junhyung Shim, Ali JannesariPPoPP 2026
- Compiling Loop-Based Nested Parallelism for Irregular WorkloadsYian Su, Mike Rainey, Nick Wanninger, Nadharm Dhiantravan et al.ASPLOS 2024 · 5 citations
- Fewer Cores, More Hertz: Leveraging High-Frequency Cores in the OS Scheduler for Improved Application PerformanceRedha Gouicem, Damien Carver, Jean-Pierre Lozi, Julien Sopena et al.USENIX ATC 2020 · 12 citations
- Efficiently Supporting Dynamic Task Parallelism on Heterogeneous Cache-Coherent SystemsMoyang Wang, Tuan Ta, Lin Cheng, Christopher BattenISCA 2020 · 11 citations
- Dopia: online parallelism management for integrated CPU/GPU architecturesYounghyun Cho, Jiyeon Park, Florian Negele, Changyeon Jo et al.PPoPP 2022 · 9 citations
