Best Speed Fit EDF Scheduling for Performance Asymmetric Multiprocessors

Abstract

In order to improve the performance of a real-time system, asymmetric multiprocessors have been proposed. The benefits of improved system performance and reduced power consumption from such architectures cannot be fully exploited unless suitable task scheduling and task allocation approaches are implemented at the operating system level. Unfortunately, most of the previous research on scheduling algorithms for performance asymmetric multiprocessors is focused on task priority assignment. They simply assign the highest priority task to the fastest processor. In this paper, we propose BSF-EDF (best speed fit for earliest deadline first) for performance asymmetric multiprocessor scheduling. This approach chooses a suitable processor rather than the fastest one, when allocating tasks. With this proposed BSF-EDF scheduling, we also derive an effective schedulability test.

Heterogeneity-Aware Scheduling on SoCs for Autonomous Vehicles

“…We evaluate HetSched for synthetic and real-world traces derived from AV benchmark suites Mini-ERA [5], ADSuite [35] and MAVBench [17]. We also incrementally evaluate our proposed features and compare with state-ofthe-art schedulers, namely CPATH [21], ADS [57] and 2lvl-EDF [56] in terms of QoM metrics (Section II-D) and overall PE utilization.…”

“…In contrast, our work targets to meet deadlines in safety-critical multi-DAG scenarios. 2lvl-EDF schedules tasks with the earliest deadline on the earliest finish time PE, as described in [56]. However, it neither considers safety constraints of tasks nor the variation in the timing profile of tasks on the heterogeneous SoC with respect to deadlines.…”

“…In this regard, Xie et al [57] propose two dynamic schedulers; fairness-based dynamic scheduler (FDS MIMF) and an adaptive dynamic scheduler (ADS MIMF), to optimize the makespan of the DAGs and to achieve low deadline miss ratio (DMR) by considering safety and criticality of the system values for high-criticality DAGs, respectively. Wu and Ryu [56], present the best speed fit EDF scheduler that prioritizes tasks according to the earliest deadline and assigns the task to the best possible PE while considering the execution profile of the task, similar to the 2lvl-EDF implementation in this work. However, none of these work consider that meeting real-time deadlines does not translate to safe completion of mission at the least mission time, i.e.…”

“…4) We show significant reductions in mission time, SoC energy and area using HetSched in a design space exploration (DSE) loop for SoC design. We show that HetSched reduces the accelerator resource requirement of an efficient SoC to safely complete AV missions by 35% (on average), compared to state-of-the-art schedulers [56], [57].…”

WAEAS: An optimization scheme of EAS scheduler for wearable applications

“…For latency-sensitive tasks, multicore scheduling strategy [8] focuses on how to place such tasks on the appropriate CPU core while meeting the performance demand of tasks. Wu and Ryu [9] introduced Best Speed Fit for Earliest Deadline First (BSF-EDF), which is a task assignment method in a heterogeneous processor system based on Earliest Deadline First (EDF) [10] . When assigning tasks, BSF-EDF selects the lowest speed idle CPU whose performance capacity is no less than the utilization of the task to perform the task with the highest priority, instead of selecting the CPU with the highest speed like the traditional scheduling algorithm, which reserves high-speed processors for the following tasks, improves the overall performance, and reduces the power consumption of the system.…”