ARCTIC: Approximate Real-Time Computing in a Cache-Conscious Multicore Environment

Saha, Sangeet; Chakraborty, Shounak; Agarwal, Sukarn; Själander, Magnus; McDonald-Maier, Klaus D.

Abstract

Improving result-accuracy in approximate computing (AC)-based time-critical systems, without violating power constraints of the underlying circuitry, is gradually becoming challenging with the rapid progress in technology scaling. The execution span of each AC real-time tasks can be split into a couple of parts: 1) the mandatory part, execution of which offers a result of acceptable quality, followed by 2) the optional part, which can be executed partially or completely to refine the initially obtained result in order to increase the result-accuracy, while respecting the time constraint. In this article, we introduce a novel hybrid offline–online scheduling strategy, ARCTIC, for AC real-time tasks. The goal of real-time scheduler of ARCTIC is to maximize the results-accuracy (QoS) of the task set with opportunistic shedding of the optional part, while respecting system-wide constraints. During execution, ARCTIC retains exclusive copy of the private cache blocks only in the local caches in a multicore system and no copies of these blocks are maintained at the other caches, and improves performance (i.e., reduces execution-time) by accumulating more live blocks on-chip. Combining offline scheduling with the online cache optimization improves both QoS and energy efficiency. While surpassing prior arts, our proposed strategy reduces the task rejection rate by up to 25%, whereas enhances QoS by 10%, with an average energy–delay–product gain of up to 9.1%, on an 8-core system.

Citation

Saha, S., Chakraborty, S., Agarwal, S., Själander, M., & McDonald-Maier, K. D. (2024). ARCTIC: Approximate Real-Time Computing in a Cache-Conscious Multicore Environment. IEEE Transactions on Computer-Aided Design of Integrated Circuits and Systems, 43(10), 2944-2957. https://doi.org/10.1109/tcad.2024.3384442

• Journal Article Type: Article
• Acceptance Date: Mar 24, 2024
• Publication Date: 2024-10
• Deposit Date: Jan 9, 2025
• Journal: IEEE Transactions on Computer-Aided Design of Integrated Circuits and Systems
• Print ISSN: 0278-0070
• Electronic ISSN: 1937-4151
• Publisher: Institute of Electrical and Electronics Engineers
• Peer Reviewed: Peer Reviewed
• Volume: 43
• Issue: 10
• Pages: 2944-2957
• DOI: https://doi.org/10.1109/tcad.2024.3384442
• Public URL: https://durham-repository.worktribe.com/output/3328972