Real-Time Workload Allocation on a Uni-Processor

IEEE Transactions on Parallel and Distributed Systems, 2008

Multicore processors deliver a higher throughput at lower power consumption than unicore processors. In the near future, they will thus be widely used in mobile real-time systems. There have been many research on energy-efficient scheduling of real-time tasks using DVS. These approaches must be modified for multicore processors, however, since normally all the cores in a chip must run at the same performance level. Thus, blindly adopting existing DVS algorithms that do not consider the restriction will result in a waste of energy. This article suggests Dynamic Repartitioning algorithm based on existing partitioning approaches of multiprocessor systems. The algorithm dynamically balances the task loads of multiple cores to optimize power consumption during execution. We also suggest Dynamic Core Scaling algorithm, which adjusts the number of active cores to reduce leakage power consumption under low load conditions. Simulation results show that Dynamic Repartitioning can produce energy savings of about 8 percent even with the best energy-efficient partitioning algorithm. The results also show that Dynamic Core Scaling can reduce energy consumption by about 26 percent under low load conditions.

—The usage of heterogeneous multicore platforms is appealing for applications, e.g. hard real-time systems, due to the potential reduced energy consumption offered by such platforms. However, the power wall is still a barrier to improving the processor design process due to the power consumption of components. Hard real-time systems are part of life critical environments and reducing the energy consumption on such systems is an onerous and complex process. This paper assesses the problem of finding optimum allocations and frequency assignments of hard real-time tasks among heterogeneous processors targeting low power consumption but taking into account timing constraints. We also propose models based on a well-established formulation in the operational research literature of the Multilevel Generalized Assignment Problem (MGAP). We tackle the problem from the perspective of different integer programming mathematical formulations and their interplay on the search for optimal solutions for RM and EDF. Computational experiments show that providing upper bounds determined by a meta-heuristic based on genetic algorithm reduces the time to finding optimal solution from hours to milliseconds, enabling us to still pursue optimum in larger instances.

To my beloved husband, Abd Hadi, my loving childrens, Aizat and Ainaa, and my loving and supportive parents, Hj Chuprat and Hajjah Rabahya. iv ACKNOWLEDGEMENT First of all, I thank ALLAH (SWT), the Lord Almighty, for giving me the health, strength, ability to complete this work and for blessing me with supportive supervisors, family and friends.

2003

When executing different real-time applications on a single processor system, one problem is how to compose these applications and guarantee at the same time that their timing requirements are not violated. A possible way of composing applications is through the resource reservation approach. Each application is handled by a dedicated server that is assigned a fraction of the processor. Using this approach, the system can be seen as a two-level hierarchical scheduler. A considerable amount of work has been recently addressed to the analysis of this kind of hierarchical systems. However, a question is still unanswered: given a set of real-time tasks to be handled by a server, how to assign the server parameters so that the task set is feasible? In this paper, we answer to the previous question for the case of fixed priority local scheduler by presenting a methodology for computing the class of server parameters that make the task set feasible.

2000

This paper introduces a maximal allowable workload task allocation problem (MAW) to address a class of distributed real-time systems that have unpredictable execution times due to varying workload. It is known that worst case execution time analysis is not well suited for these systems. This problem seeks to maximize the upper bound of permissible workload in an allocation so that

IEEE transactions on systems, man, and cybernetics, 2020

Proceedings 21st IEEE Real-Time Systems Symposium

In this paper we propose a novel scheduling framework for a real-time environment that experiences dynamic changes. This framework is capable of adjusting the system workload in incremental steps under overloaded conditions such that the most critical tasks in the system are always scheduled and the total value of the system is m i m i z e d. Each task has an assigned criticality value and consists of two parts, a mandatory part and an optional part. A timely answer is available after the mandatory part completes execution and its value may be improved by executing the entire optional part. Optional parts can be discarded in overloaded conditions. The process of selecting optional parts to discard while maximizing the value of the system requires the exploration of a potentially large number of combinations. Since this process is too time consuming to be computed on-line, an approximate algorithm is executed incrementally whenever the processor would otherwise be idle, progressively rejning the quality of the solution. This criteria allows the scheduler to handle overloads with low cost while maximizing the use of the available resources and without jeopardizing the temporal constraints of the most critical tasks in the system. Simulation results show that few stages of the algorithm need to be executed for achieving a performance with near-optimal results.

2008

High-performance microprocessors, e.g., multithreaded and multicore processors, are being implemented in embedded real-time systems because of the increasing computational requirements. These complex microprocessors have two major drawbacks when they are used for real-time purposes. First, their complexity difficults the calculation of the WCET (Worst Case Execution Time). Second, power consumption requirements are much larger, which is a major concern in these systems.

In real-time systems, it is required to complete all work on a timely basis. There are mainly two types of real time systems: hard real-time systems (HRT) and soft-real time (SRT) systems. In hard real-time systems, a missed deadline is considered a system failure; in soft real-time systems some deadlines may be missed. The aim of real-time scheduling analysis is to ensure a sequence of jobs meets their deadlines. Many real-time systems allow jobs to interrupt, or preempt, one another. In multiprocessor systems a preemption may result in a job migrating from one processor to another. Both preemptions and migrations cause scheduling overheads. In this dissertation, we present two approaches for reducing scheduling overheads. One approach reduces the number of preemptions and migrations by adjusting job priorities. Another approach incorporates genetic algorithms to classify HRT task sets, and uses heuristics to reduce the number of preemptions and migrations. Another type of overhead that this dissertation addresses is energy consumption. This dissertation presents an algorithm to use Dynamic Voltage and Frequency Scaling (DVFS) processors for conserving energy. The proposed algorithm drastically reduces the power consumption of the systems by slowing down Contents ACKNOWLEDGMENTS iv LIST OF FIGURES vi

IEEE Embedded Systems Letters, 2019

The shift from homogeneous multi-core to heterogeneous multi-core introduces challenges in scheduling the tasks to the appropriate cores maintaining the time deadline. This paper studies the existing scheduling schemes in heterogeneous multi-core system and finds an approach to enhance the homogeneous system model to heterogeneous scheduling architecture. The proposed model increases the overall system utilization by accommodating almost all the tasks (low power task and high power task) into appropriate cores (big high power and small low power). It further enhances the system performance by allocating rejected jobs from small cores into the big cores through dispatcher.

Computing Research Repository, 2011

An optimal solution to the problem of scheduling real-time tasks on a set of identical processors is derived. The described approach is based on solving an equivalent uniprocessor real-time scheduling problem. Although there are other scheduling algorithms that achieve optimality, they usually impose prohibitive preemption costs. Unlike these algorithms, it is observed through simulation that the proposed approach produces no more than three preemptions points per job.

Proceeding. 10th EUROMICRO Workshop on Real-Time Systems (Cat. No.98EX168), 1998

This paper introduces improvements in partitioning schemes for multiprocessor real-time systems which allow higher processor utilization and enhanced schedulability by using exact feasibility tests to evaluate the schedulability limit of a processor. The paper analyzes how to combine these tests with existing bin-packing algorithms for processor allocation and provides new variants which are exhaustively evaluated using two assumptions: variable and fixed number of processors. The problem of evaluating this algorithms is complex, since metrics are usually based on comparing the performance of a given algorithm to an optimal one, but determining optimals is often NP-hard on multiprocessors. This problem has been overcome by defining lower or upper bounds on the performance of an optimal algorithm and then defining metrics with respect this bounds. The evaluation has shown that the algorithms exhibit extremely good behaviors and they can be considered very close to the maximum achievable utilization. It is also shown that dynamic priority policies produces significantly better results than fixed priorities policies when task sets require high processor utilizations.

1998

In a parallelizable task model, a task can be parallelized and the component tasks can be executed concurrently on multiple processors. We use this parallelism in tasks to meet their deadlines and also obtain better processor utilisation compared to non-parallelized tasks. Non-preemptive parallelizable task scheduling combines the advantages of higher schedulability and lower scheduling overhead o ered by the preemptive and non-preemptive task scheduling models, respectively. We propose a new approach to maximize the bene ts from task parallelization. It involves checking the schedulability of periodic tasks (if necessary, by parallelizing them) o-line and run-time scheduling of the schedulable periodic tasks together with dynamically arriving aperiodic tasks. To avoid the run-time anomaly that may occur when the actual computation time of a task is less than its worst case computation time, we propose e cient run-time mechanisms. We have carried out extensive simulation to study the e ectiveness of the proposed approach by comparing the schedulability o ered by it with that of dynamic scheduling using Earliest Deadline First (EDF), and by comparing its storage e ciency with that of the static table-driven approach. We found that the schedulability o ered by parallelizable task scheduling is always higher than that of the EDF algorithm for a wide variety of task parameters and the storage overhead incurred by it is less than 3.6% of the static table-driven approach even under heavy task loads.

Real-Time Technology and Applications - Proceedings, 2002

In this paper we propose a novel scheduling framework for a dynamic real-time environment with energy constraints. This framework dynamically adjusts the CPU voltage/frequency so that no task in the system misses its deadline and the total energy savings of the system are maximized. In this paper we consider only realistic, discrete-Ievel speeds. Each task in the system consumes a certain amount of energy, which depends on a speed chosen for execution. The process of selecting speeds for execution while maximizing the energy savings of the system requires the exploration of a large number of combinations, which is too time consuming to be computed on-line. Thus, we propose an integrated heuristic methodology which executes an optimization procedure in a low computation time. This scheme allows the scheduler to handle power-aware realtime tasks with low cost while maximizing the use of the available resources and without jeopardizing the temporal constraints of the system. Simulation results show that our heuristic methodology is able to generate power-aware scheduling solutions with near-optimal performance.

Most currently existing optimal real-time multiprocessor scheduling algorithms follow the fairness rule, in which all tasks are forced to make progress in their executions proportional to their utilization, to ensure the optimality of the algorithm. However, obeying the fairness rule results in large number of task preemptions and migrations and these highly affect the practicability of the algorithm. In this paper, we present an efficient real-time multiprocessor scheduling algorithm in which the fairness rule is completely relaxed and a semi-greedy algorithm is introduced. In the simulation, the proposed algorithm showed promising results in terms of number of task preemptions and migrations that are very few compared to the current state of the art real-time multiprocessor scheduling algorithms. Although the algorithm can sometimes miss a very few deadlines, we assume that these deadline misses can be tolerated in view of the great reduction of task preemptions and migrations.