VisSched: An Auction-Based Scheduler for Vision Workloads on Heterogeneous Processors

• Published in: IEEE transactions on computer-aided design of integrated circuits and systems Vol. 39; no. 11; pp. 4252 - 4265
• Main Authors: Moolchandani, Diksha, Kumar, Anshul, Martinez, Jose F., Sarangi, Smruti R.
• Format: Journal Article
• Language: English
• Published: New York IEEE 01.11.2020; The Institute of Electrical and Electronics Engineers, Inc. (IEEE)
• Subjects: Adaptation models; Algorithms; Asymmetric multicores; auction theory; Benchmark testing; Computer vision; Edge computing; hardware scheduler; Linear algebra; Message systems; Nash equilibrium; Optimization; Processor scheduling; Schedules; Scheduling; Task analysis; Workload; Workloads
• ISSN: 0278-0070; 1937-4151
• DOI: 10.1109/TCAD.2020.3013076

With the growth of edge computing, application-specific workloads based on computer vision are steadily migrating to edge cloudlets. Scheduling has been identified to be a major problem in these cloudlets. In this article, we propose a generic architectural solution, VisSched , that leverages the fact that most vision workloads share similar code kernels (such as library code for linear algebra), and as a result, they tend to exhibit similar phase behavior. This allows us to create an auction theory-based scheduling mechanism, where we give each thread a replenishable virtual wallet, and threads are scheduled based on the amounts that they bid for executing on a free core. We show that in 20%-40% of the cases, our scheduling algorithm is theoretically optimal, and in the remaining cases, it reaches a global optimum obtained using Monte Carlo simulations 90%-95% of the time. Our results for the MEVBench vision workloads show a 17% higher performance and a 14% lower <inline-formula> <tex-math notation="LaTeX">ED^{2} </tex-math></inline-formula> as compared to the nearest competing algorithm in the literature.