Using Imprecise Computing for Improved Real-Time Scheduling
MetadataShow full item record
Conventional hard real-time scheduling is often overly pessimistic due to the worst case execution time estimation. The pessimism can be mitigated by exploiting imprecise computing in applications where occasional small errors are acceptable. This leverage is investigated in a few previous works, which are restricted to preemptive cases. We study how to make use of imprecise computing in uniprocessor non-preemptive real-time scheduling, which is known to be more difficult than its preemptive counterpart. Several heuristic algorithms are developed for periodic tasks with independent or cumulative errors due to imprecision. Simulation results show that the proposed techniques can significantly improve task schedulability and achieve desired accuracy– schedulability tradeoff. The benefit of considering imprecise computing is further confirmed by a prototyping implementation in Linux system. Mixed-criticality system is a popular model for reducing pessimism in real-time scheduling while providing guarantee for critical tasks in presence of unexpected overrun. However, it is controversial due to some drawbacks. First, all low-criticality tasks are dropped in high-criticality mode, although they are still needed. Second, a single high-criticality job overrun leads to the pessimistic high-criticality mode for all high-criticality tasks and consequently resource utilization becomes inefficient. We attempt to tackle aforementioned two limitations of mixed-criticality system simultaneously in multiprocessor scheduling, while those two issues are mostly focused on uniprocessor scheduling in several recent works. We study how to achieve graceful degradation of low-criticality tasks by continuing their executions with imprecise computing or even precise computing if there is sufficient utilization slack. Schedulability conditions under this Variable-Precision Mixed-Criticality (VPMC) system model are investigated for partitioned scheduling and global fpEDF-VD scheduling. And a deferred switching protocol is introduced so that the chance of switching to high-criticality mode is significantly reduced. Moreover, we develop a precision optimization approach that maximizes precise computing of low-criticality tasks through 0-1 knapsack formulation. Experiments are performed through both software simulations and Linux proto- typing with consideration of overhead. Schedulability of the proposed methods is studied so that the Quality-of-Service for low-criticality tasks is improved with guarantee of satisfying all deadline constraints. The proposed precision optimization can largely reduce computing errors compared to constantly executing low-criticality tasks with imprecise computing in high-criticality mode.
Huang, Lin (2019). Using Imprecise Computing for Improved Real-Time Scheduling. Doctoral dissertation, Texas A&M University. Available electronically from