MARCO: A High-performance Task M apping a nd R outing Co -optimization Framework for Point-to-Point NoC-based Heterogeneous Computing Systems

Heterogeneous computing systems (HCSs), which consist of various processing elements (PEs) that vary in their processing ability, are usually facilitated by the network-on-chip (NoC) to interconnect its components. The emerging point-to-point NoCs which support single-cycle-multi-hop transmission, reduce or eliminate the latency dependence on distance, addressing the scalability concern raised by high latency for long-distance transmission and enlarging the design space of the routing algorithm to search the non-shortest paths. For such point-to-point NoC-based HCSs, resource management strategies which are managed by compilers, scheduler, or controllers, e.g., mapping and routing, are complicated for the following reasons: (i) Due to the heterogeneity, mapping and routing need to optimize computation and communication concurrently (for homogeneous computing systems, only communication). (ii) Conducting mapping and routing consecutively cannot minimize the schedule length in most cases since the PEs with high processing ability may locate in the crowded area and suffer from high resource contention overhead. (iii) Since changing the mapping selection of one task will reconstruct the whole routing design space, the exploration of mapping and routing design space is challenging. Therefore, in this work, we propose MARCO, the m apping a nd r outing co -optimization framework, to decrease the schedule length of applications on point-to-point NoC-based HCSs. Specifically, we revise the tabu search to explore the design space and evaluate the quality of mapping and routing. The advanced reinforcement learning (RL)algorithm, i.e., advantage actor-critic, is adopted to efficiently compute paths. We perform extensive experiments on various real applications, which demonstrates that the MARCO achieves a remarkable performance improvement in terms of schedule length (+44.94% ∼ +50.18%) when compared with the state-of-the-art mapping and routing co-optimization algorithm for homogeneous computing systems. We also compare MARCO with different combinations of state-of-the-art mapping and routing approaches.

Improved Associated Task Scheduling in Different Heterogeneous systems: تحسين جدولة المهام المترابطة في الأنظمة مختلفة التجانس

Scheduling of associated tasks in heterogeneous systems is very important to reduce the total completion time as heterogeneity and coherence introduce more complexity. Several studies have focused on studying task scheduling in homogeneous systems and some studies have been limited to scheduling in heterogeneous systems. In this paper, a new algorithm was created and a simulation of the proposed algorithm was designed and tested using a random schema generator. This algorithm showed better results than its predecessors in terms of schedule length ratio relative to heterogeneity factor and Communication to Computation Ratio factor.

Scheduling Multiprocessor Tasks with Equal Processing Times as a Mixed Graph Coloring Problem

This article extends the scheduling problem with dedicated processors, unit-time tasks, and minimizing maximal lateness for integer due dates to the scheduling problem, where along with precedence constraints given on the set of the multiprocessor tasks, a subset of tasks must be processed simultaneously. Contrary to a classical shop-scheduling problem, several processors must fulfill a multiprocessor task. Furthermore, two types of the precedence constraints may be given on the task set . We prove that the extended scheduling problem with integer release times of the jobs to minimize schedule length may be solved as an optimal mixed graph coloring problem that consists of the assignment of a minimal number of colors (positive integers) to the vertices of the mixed graph such that, if two vertices and are joined by the edge , their colors have to be different. Further, if two vertices and are joined by the arc , the color of vertex has to be no greater than the color of vertex . We prove two theorems, which imply that most analytical results proved so far for optimal colorings of the mixed graphs , have analogous results, which are valid for the extended scheduling problems to minimize the schedule length or maximal lateness, and vice versa.

Operating Room (Re)Scheduling with Bed Management via ASP

The Operating Room Scheduling (ORS) problem is the task of assigning patients to operating rooms (ORs), taking into account different specialties, lengths, and priority scores of each planned surgery, OR session durations, and the availability of beds for the entire length of stay (LOS) both in the Intensive Care Unit (ICU) and in the wards. A proper solution to the ORS problem is of primary importance for the healthcare service quality and the satisfaction of patients in hospital environments. In this paper we first present a solution to the problem based on Answer Set Programming (ASP). The solution is tested on benchmarks with realistic sizes and parameters, on three scenarios for the target length on 5-day scheduling, common in small–medium-sized hospitals, and results show that ASP is a suitable solving methodology for the ORS problem in such setting. Then, we also performed a scalability analysis on the schedule length up to 15 days, which still shows the suitability of our solution also on longer plan horizons. Moreover, we also present an ASP solution for the rescheduling problem, that is, when the offline schedule cannot be completed for some reason. Finally, we introduce a web framework for managing ORS problems via ASP that allows a user to insert the main parameters of the problem, solve a specific instance, and show results graphically in real time.

A Self-Organized Scheduling Algorithm for Embedded Real-Time Systems

In this paper, the authors present a self-organized approach for scheduling tasks on processors in embedded real-time systems. For such a mapping, two conflicting objectives have to be optimized: the reliability and the schedule length. Since the scheduling problem is NP-hard, a heuristic algorithm is used to produce schedules achieving different trade-offs between the two objectives. Moreover, a self-organization strategy based on dynamic crowding distance is adopted. This allows a better exploration of the objective space as well as an enhanced solution diversity. The proposed algorithm, reliability schedule length trad-offs algorithm (RSTA), is tested and compared with the popular SPEA2 algorithm where experimental results are promising on both quality and diversity of solutions.

Task-Level Aware Scheduling of Energy-Constrained Applications on Heterogeneous Multi-Core System

Minimizing the schedule length of parallel applications, which run on a heterogeneous multi-core system and are subject to energy consumption constraints, has recently attracted much attention. The key point of this problem is the strategy to pre-allocate the energy consumption of unscheduled tasks. Previous articles used the minimum value, average value or a power consumption weight value as the pre-allocation energy consumption of tasks. However, they all ignored the different levels of tasks. The tasks in different task levels have different impact on the overall schedule length when they are allocated the same energy consumption. Considering the task levels, we designed a novel task energy consumption pre-allocation strategy that is conducive to minimizing the scheduling time and developed a novel task schedule algorithm based on it. After getting the preliminary scheduling results, we also proposed a task execution frequency re-adjustment mechanism that can re-adjust the execution frequency of tasks, to further reduce the overall schedule length. We carried out a considerable number of experiments with practical parallel application models. The results of the experiments show that our method can reach better performance compared with the existing algorithms.

A New Intra-Cluster Scheduling Scheme for Real-Time Flows in Wireless Sensor Networks

Real-time flows using time division multiple access (TDMA) scheduling in cluster-based wireless sensor networks try to schedule more flows per time frame to minimize the schedule length to meet the deadline. The problem with the previously used cluster-based scheduling algorithm is that intra-cluster scheduling does not consider that the clusters may have internal or outgoing flows. Thus, intra-cluster scheduling algorithms do not utilize their empty time-slots and thus increase schedule length. In this paper, we propose a new intra-cluster scheduling algorithm by considering that clusters may have having internal or outgoing flows. Thus, intra-cluster scheduling algorithms do not differentiate the intra-cluster time slots and utilize their empty time slots. The objective is to schedule more flows per time frame, to reduce schedule length and improve the acceptance rate of flows. Simulation results show that the acceptance rate of the proposed scheme has a higher performance than the previous scheme.

Efficient Solution for Load Balancing in Fog Computing Utilizing Artificial Bee Colony

Fog computing is a set of mobile cloudlets which can fulfil the demand of the user who is already considered a mobile job in this architecture. The main aim of Fog computing is to provide the user with an optimal solution which is quick and cost-efficient. This article focuses on a load balancing mechanism for cloudlets along with keeping the cost-effectiveness as an optimal selection parameter. This article utilizes the Artificial Bee Colony (ABC) in order to prioritize the user demand using a fitness function. This work evaluates quality of service (QoS) parameters such as schedule length runtime (SLR), schedule length vm ratio (SLVMR), energy consumed (EC) and energy consumption ratio (ECR) and shows the effectiveness of proposed work.

An Optimal Solution in cloud computing for Load Balancing using Genetic Algorithm

Cloud computing is a research trend which bring various cloud services to the users. Cloud environment face various challenges and issues to provide efficient services. In this paper, a novel Genetic Algorithm based load balancing algorithm has been implemented to balance the load in the network. The literature review has been studied to understand the research gap. More specifically, load balancing technique authenticate the network by enabling Virtual Machines (VM). The proposed technique has been further evaluated using the Schedule Length Runtime (SLR) and Energy consumption (EC) parameters. Overall, the effective results has been obtained such as 46% improvement in consuming the energy and 12 % accuracy for the SLR measurement. In addition, results has been compared with the conventional approaches to validate the outcomes.