Scheduling Task-parallel Applications in Dynamically Asymmetric Environments

Chen, Jing; Soomro, Pirah Noor; Abduljabbar, Mustafa; Manivannan, Madhavan; Pericàs, Miquel

doi:10.1145/3409390.3409408

Cited by 1 publication

(2 citation statements)

References 21 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…The online performance modeling module enables ERASE to predict the execution time of a task when mapped on an execution place. In ERASE, we adopt a history-based model since it is a simple yet effective method for estimating performance and detecting performance heterogeneity and system interference for fine-grained tasking applications [11,13,24,37]. The online performance module is implemented as a look-up table for each kernel.…”

Section: Online Performance Modelingmentioning

confidence: 99%

“…These schedulers do not rely on DVFS but use other approaches for improving performance and energy efficiency. To compare the efficacy of the schedulers at different static frequency settings, we perform the evaluation using the four MIN and MAX frequency permutations on the two clusters, as shown in the x-axis labels of Figures 6,7,8,and 11. For example, MAX & MIN means that the Denver cluster is set to maximum frequency and the A57 cluster is set to minimum frequency.…”

Section: Evaluation On Asymmetric Platformmentioning

confidence: 99%

See 1 more Smart Citation

ERASE: Energy Efficient Task Mapping and Resource Management for Work Stealing Runtimes

Chen

Manivannan

Abduljabbar

et al. 2022

ACM Trans. Archit. Code Optim.

Self Cite

View full text Add to dashboard Cite

Parallel applications often rely on work stealing schedulers in combination with fine-grained tasking to achieve high performance and scalability. However, reducing the total energy consumption in the context of work stealing runtimes is still challenging, particularly when using asymmetric architectures with different types of CPU cores. A common approach for energy savings involves dynamic voltage and frequency scaling (DVFS) wherein throttling is carried out based on factors like task parallelism, stealing relations, and task criticality. This article makes the following observations: (i) leveraging DVFS on a per-task basis is impractical when using fine-grained tasking and in environments with cluster/chip-level DVFS; (ii) task moldability, wherein a single task can execute on multiple threads/cores via work-sharing, can help to reduce energy consumption; and (iii) mismatch between tasks and assigned resources (i.e., core type and number of cores) can detrimentally impact energy consumption. In this article, we propose EneRgy Aware SchedulEr (ERASE), an intra-application task scheduler on top of work stealing runtimes that aims to reduce the total energy consumption of parallel applications. It achieves energy savings by guiding scheduling decisions based on per-task energy consumption predictions of different resource configurations. In addition, ERASE is capable of adapting to both given static frequency settings and externally controlled DVFS. Overall, ERASE achieves up to 31% energy savings and improves performance by 44% on average, compared to the state-of-the-art DVFS-based schedulers.

show abstract

Section: Online Performance Modelingmentioning

confidence: 99%

Section: Evaluation On Asymmetric Platformmentioning

confidence: 99%