Task-I/O Co-scheduling for Pfair Real-Time Scheduler in Embedded Multi-core Systems

Abstract

Real-time embedded systems often require the ability of robust controls because interactions between the embedded systems and their physical environment that dynamically changes. Multi-core chips are regarded as ideal candidate hardware components for such environments, since each of them carries two or more cores on a single die, and has potential for providing execution parallelism as well as better performance at low cost. Parallelism, on the other hand, necessitates complex analysis of computation problems, such as task scheduling, while improving the realization of embedded controls. Pfair is an optimal scheduling algorithm that can fully utilize all cores in the system, but it incurs an excessive scheduling overhead which, in turn, diminishes its practicality in embedded systems. To mitigate this problem, hybrid partitioned-global Pfair (HPGP) scheduler was proposed, which significantly reduces the number of task migrations and global scheduling points by performing global scheduling only when absolutely necessary, while still achieving full processor utilization. This paper further extends the HPGP scheduler to support the robustness to interactions with the physical environment. Our evaluation results have shown that the extended HPGP can successfully limits the increase in response time caused by the hardware interrupts for physical interactions under a wide range of system utilization conditions., thus making it suitable for embedded real-time systems.

Early-Release Boundary Fair Scheduler with Task-interrupt Co-scheduling

“…Consider that arrival of interrupts can be processed by any core in a multi-core system, we can utilizing an idle or lightly-loaded core to process the interrupts in parallel, while not interfering with application tasks execution [2]. Based on this idea, we extend the ERBF scheduler with a taskinterrupt co-scheduling approach, instead of using a fixed core or fixed priority for interrupt processing.…”

“…2(b). In first time slice [0, 5), the mandatory time units allocated to 2 T and 3 T is 1, and they receives 2 optional time units respectively, such that the total amount of 2 T and 3 T executed is equal to i C , this means that they have completed the task ahead in current cycle [0, 15). Moreover, under BF, 6 T execution completed at time instant 30, and under ERBF, it was completed at time 10.…”

“…Such a control system uses a large number of sensors and performs interaction with the physical environment, interrupt scheduling, is as important as task scheduling to meet the system-level control performance [2] .…”

“…For a multi-core system, in order to disable interrupts and to replace interrupt handler, we can implement a polling mechanism as sporadic tasks [2] . When time-critical tasks are executing in a heavily-loaded system, the interrupts are usually disabled so as to avoid the possible deadline misses of time-critical tasks.…”

“…However, by making scheduling decision at every time unit, Pfair algorithms can incur quite high scheduling overhead, especially for systems with small time quantum. Sangsoo Park proposed the HPGP (hybrid partitioned-global Pfair) scheduler [2] , which based on the Pfair scheduler, it introduces a decision algorithm which calls for global scheduling only if it is absolutely necessary, and performs partitioned scheduling that tasks are scheduled within the prior-assigned core in most of the run-time, thus, reduces the scheduling overheads.…”