A Survey on Task Scheduling of CPU-GPU Heterogeneous Cluster

This paper reviews task scheduling frameworks,methods,and evaluation metrics of central processing unit-graphics processing unit(CPU-GPU)heterogeneous clusters.Task scheduling of CPU-GPU heterogeneous clusters can be carried out on the system level,nodelevel,and device level.Most task-scheduling technologies are heuristic based on the experts’experience,while some technologies are based on statistic methods using machine learning,deep learning,or reinforcement learning.Many metrics have been adopted to evaluate and compare different task scheduling technologies that try to optimize different goals of task scheduling.Although statistic task scheduling has reached fewer research achievements than heuristic task scheduling,the statistic task scheduling still has significant research potential.

阴/非离子型表面活性剂对CMP后SiO_(2)颗粒的去除效果

为了更有效地去除铜晶圆化学机械抛光(CMP)后清洗残留的SiO_(2)颗粒,选择了2种阴离子型表面活性剂(SLS、TD⁃40)和2种非离子型表面活性剂(AEO⁃5、JFC⁃6),通过接触角、表面张力、电化学、分子动力学模拟实验探究了4种表面活性剂在铜表面的润湿性、吸附构型及吸附稳定性。通过优化表面活性剂质量浓度,选择达到吸附稳定时的质量浓度配置4种表面活性剂来清洗铜晶圆,利用扫描电子显微镜观测铜表面形貌,对比它们的清洗效果。随后选择TD⁃40和JFC⁃6进行复配,研究复配后表面活性剂对硅溶胶颗粒的去除效果。实验结果表明,使用体积比为2∶1的TD⁃40与JFC⁃6进行复配得到的CMP清洗液对SiO_(2)颗粒的去除效果比单一表面活性剂的更好。

DR-IS:Dynamic Response Incremental Scheduling in Time-Sensitive Network

Time-Sensitive Network(TSN)with deterministic transmission capability is increasingly used in many emerging fields.It mainly guarantees the Quality of Service(QoS)of applications with strict requirements on time and security.One of the core features of TSN is traffic scheduling with bounded low delay in the network.However,traffic scheduling schemes in TSN are usually synthesized offline and lack dynamism.To implement incremental scheduling of newly arrived traffic in TSN,we propose a Dynamic Response Incremental Scheduling(DR-IS)method for time-sensitive traffic and deploy it on a software-defined time-sensitive network architecture.Under the premise of meeting the traffic scheduling requirements,we adopt two modes,traffic shift and traffic exchange,to dynamically adjust the time slot injection position of the traffic in the original scheme,and determine the sending offset time of the new timesensitive traffic to minimize the global traffic transmission jitter.The evaluation results show that DRIS method can effectively control the large increase of traffic transmission jitter in incremental scheduling without affecting the transmission delay,thus realizing the dynamic incremental scheduling of time-sensitive traffic in TSN.

Energy-Saving Distributed Flexible Job Shop Scheduling Optimization with Dual Resource Constraints Based on Integrated Q-Learning Multi-Objective Grey Wolf Optimizer

The distributed flexible job shop scheduling problem(DFJSP)has attracted great attention with the growth of the global manufacturing industry.General DFJSP research only considers machine constraints and ignores worker constraints.As one critical factor of production,effective utilization of worker resources can increase productivity.Meanwhile,energy consumption is a growing concern due to the increasingly serious environmental issues.Therefore,the distributed flexible job shop scheduling problem with dual resource constraints(DFJSP-DRC)for minimizing makespan and total energy consumption is studied in this paper.To solve the problem,we present a multi-objective mathematical model for DFJSP-DRC and propose a Q-learning-based multi-objective grey wolf optimizer(Q-MOGWO).In Q-MOGWO,high-quality initial solutions are generated by a hybrid initialization strategy,and an improved active decoding strategy is designed to obtain the scheduling schemes.To further enhance the local search capability and expand the solution space,two wolf predation strategies and three critical factory neighborhood structures based on Q-learning are proposed.These strategies and structures enable Q-MOGWO to explore the solution space more efficiently and thus find better Pareto solutions.The effectiveness of Q-MOGWO in addressing DFJSP-DRC is verified through comparison with four algorithms using 45 instances.The results reveal that Q-MOGWO outperforms comparison algorithms in terms of solution quality.

Research on Flexible Job Shop Scheduling Based on Improved Two-Layer Optimization Algorithm

To improve the productivity,the resource utilization and reduce the production cost of flexible job shops,this paper designs an improved two-layer optimization algorithm for the dual-resource scheduling optimization problem of flexible job shop considering workpiece batching.Firstly,a mathematical model is established to minimize the maximum completion time.Secondly,an improved two-layer optimization algorithm is designed:the outer layer algorithm uses an improved PSO(Particle Swarm Optimization)to solve the workpiece batching problem,and the inner layer algorithm uses an improved GA(Genetic Algorithm)to solve the dual-resource scheduling problem.Then,a rescheduling method is designed to solve the task disturbance problem,represented by machine failures,occurring in the workshop production process.Finally,the superiority and effectiveness of the improved two-layer optimization algorithm are verified by two typical cases.The case results show that the improved two-layer optimization algorithm increases the average productivity by 7.44% compared to the ordinary two-layer optimization algorithm.By setting the different numbers of AGVs(Automated Guided Vehicles)and analyzing the impact on the production cycle of the whole order,this paper uses two indicators,the maximum completion time decreasing rate and the average AGV load time,to obtain the optimal number of AGVs,which saves the cost of production while ensuring the production efficiency.This research combines the solved problem with the real production process,which improves the productivity and reduces the production cost of the flexible job shop,and provides new ideas for the subsequent research.

半导体材料CMP过程中磨料的研究进展

对磨料在半导体材料化学机械抛光(CMP)中的应用和研究进展进行了简单阐述,从各代半导体材料制成半导体器件的加工要求介绍了磨料在半导体材料CMP中的重要性,从CMP过程中磨料与半导体材料的相互作用介绍了磨料在半导体材料CMP中的环保性,从磨料的改性和制备介绍了磨料在半导体材料CMP中应用的限制性,重点从半导体材料的去除速率和表面质量介绍了磨料对半导体材料抛光性能的影响,并对国内外研究中单一磨料、混合磨料和复合磨料对半导体材料抛光性能的影响进行了评述,总结了近年来磨料在半导体材料CMP中的研究进展。最后,对磨料在半导体材料CMP中存在的共性问题进行了总结,并对该领域所面临的挑战及发展方向进行了展望。

C^(*)-代数上的CMP逆

给出了有单位元的C^(*)-代数中CMP逆的刻画,讨论了CMP逆与Moore-Penrose逆,Drazin逆和(b,c)-逆等广义逆之间的关系。

复合磨料的制备及其对层间介质CMP性能的影响

以SiO_(2)为内核、CeO_(2)为外壳制备出了核壳结构复合磨料,用以提升集成电路层间介质的去除速率及表面一致性。采用扫描电子显微镜(SEM)观察复合磨料的表面形貌,利用X射线衍射仪(XRD)、傅里叶变换红外光谱仪(FTIR)和X射线光电子能谱仪(XPS)分析复合磨料的表面物相结构及化学键组成。研究结果表明,所制备的复合磨料呈现出“荔枝”形,平均粒径为70~90 nm,CeO_(2)粒子主要以Si—O—Ce键与SiO_(2)内核结合。将所制备的复合磨料配置成抛光液进行层间介质化学机械抛光(CMP)实验。实验结果表明,Zeta电位随着pH值的降低而升高,当pH值约为6.8时达到复合磨料的等电点。当pH值为3时,层间介质去除速率达到最大,为481.6 nm/min。此外,研究发现去除速率还与摩擦力和温度有关,CMP后的SiO_(2)晶圆均方根表面粗糙度为0.287 nm。

MCWOA Scheduler:Modified Chimp-Whale Optimization Algorithm for Task Scheduling in Cloud Computing

Cloud computing provides a diverse and adaptable resource pool over the internet,allowing users to tap into various resources as needed.It has been seen as a robust solution to relevant challenges.A significant delay can hamper the performance of IoT-enabled cloud platforms.However,efficient task scheduling can lower the cloud infrastructure’s energy consumption,thus maximizing the service provider’s revenue by decreasing user job processing times.The proposed Modified Chimp-Whale Optimization Algorithm called Modified Chimp-Whale Optimization Algorithm(MCWOA),combines elements of the Chimp Optimization Algorithm(COA)and the Whale Optimization Algorithm(WOA).To enhance MCWOA’s identification precision,the Sobol sequence is used in the population initialization phase,ensuring an even distribution of the population across the solution space.Moreover,the traditional MCWOA’s local search capabilities are augmented by incorporating the whale optimization algorithm’s bubble-net hunting and random search mechanisms into MCWOA’s position-updating process.This study demonstrates the effectiveness of the proposed approach using a two-story rigid frame and a simply supported beam model.Simulated outcomes reveal that the new method outperforms the original MCWOA,especially in multi-damage detection scenarios.MCWOA excels in avoiding false positives and enhancing computational speed,making it an optimal choice for structural damage detection.The efficiency of the proposed MCWOA is assessed against metrics such as energy usage,computational expense,task duration,and delay.The simulated data indicates that the new MCWOA outpaces other methods across all metrics.The study also references the Whale Optimization Algorithm(WOA),Chimp Algorithm(CA),Ant Lion Optimizer(ALO),Genetic Algorithm(GA)and Grey Wolf Optimizer(GWO).

pH调节剂在CMP工艺中的应用研究进展

pH调节剂在化学机械抛光(CMP)工艺中有重要应用,可以调节抛光液的pH值以确保抛光过程的化学反应在理想的pH值下进行,同时保持抛光化学环境的稳定等。对无机酸、有机酸、无机碱和有机碱四大类pH调节剂在合金、金属和金属化合物等材料的CMP中的应用及其作用机理进行综述。无机酸pH调节剂的主要作用机理是快速腐蚀材料表面,但其主要缺点是会将多余的金属离子引入抛光液中污染金属表面。有机酸pH调节剂的主要作用机理是螯合金属离子形成大分子络合物,但其主要缺点是稳定性差,难以保存。无机碱pH调节剂的主要作用机理是在基底表面生成一层软化层,使其在机械作用下更容易被去除,但其主要缺点是仍会引入金属离子污染材料表面。有机碱pH调节剂的主要作用机理是加速钝化膜的形成,但其主要缺点是制备困难、成本高。最后对pH调节剂在CMP中的应用前景进行了展望。

CMP抛光垫表面及材料特性对抛光效果影响的研究进展

对化学机械抛光(CMP)工艺中的抛光垫特性、劣化以及修整进行了简单阐述,重点先从抛光垫表面特性(抛光垫表面微形貌、微孔及抛光垫的结构、表面沟槽纹理的形状、微凸体的分布)和材质特性(硬度、弹性模量和化学性能)入手,对近年来国内外的实验研究与理论模拟分析两方面进行了概括,总结了目前各个特性参数对抛光垫性能以及对CMP过程影响的进展,此外,从机械磨损和化学腐蚀两方面对抛光垫的劣化机理进行简要分析。随后,为进一步探究抛光垫修整对抛光性能影响,对抛光垫的两种修整方式和修整参数对修整的效果进行了归纳,介绍了几种新型自修整材料。最后,指出了抛光垫特性和修整在发展现状中存在的问题,未来抛光垫的发展趋势将逐渐走向创新化、智能化、理论化以及应用集成化。

基于流固耦合的碳化硅衬底CMP过程温度场仿真分析

在碳化硅衬底化学机械抛光过程中,抛光界面温度是影响抛光效率的关键因素之一,掌握抛光界面温度分布情况,有助于更深入理解CMP机理并为工艺优化提供理论指导。为此,对碳化硅衬底的CMP过程中温度场分布情况进行了探究,分析了不同抛光工艺参数和抛光液组分对抛光界面温度的影响。利用有限元分析软件ANSYS的流固耦合模块,综合考虑抛光垫与抛光液对SiC衬底的磨削作用,得到抛光过程中SiC衬底表面温度分布。仿真结果表明,SiC衬底径向温度从中心到边缘逐渐增大,边缘处上升趋势逐渐减小甚至出现温度小幅下降,最大温差接近0.4℃(约为4%)。通过单因素实验探究不同影响因素与温度之间的关系,得出结论:随着抛光转速和抛光压力的增大,SiC表面平均温度上升,均近似成线性关系,并且边缘点与中心点温度变化相差越来越大;同时,衬底界面温度随着抛光液磨料浓度的增加而上升,但变化相对较小。

CMP抛光液中SiO_(2)磨料分散稳定性的研究进展

对SiO_(2)磨料在化学机械抛光(CMP)抛光液中的应用以及影响抛光液分散稳定性的因素进行了阐述,重点从SiO_(2)磨料分散稳定性的角度介绍了SiO_(2)磨料质量分数和粒径、抛光液pH值、表面活性剂种类和表面改性等对抛光液稳定性的影响,通过分析Zeta电位绝对值的范围、凝胶时间的长短、粒径随时间的变化和接触角等,总结了小粒径(35 nm左右)SiO_(2)磨料在抛光液中的分散机理,同时探讨了弱碱性环境对磨料Zeta电位的影响,阳离子、阴离子和非离子表面活性剂对磨料表面的作用机理和复配使用的效果,以及表面疏水化或亲水化改性对磨料分散稳定性的影响。最后对该领域未来的发展方向进行了展望。

E1310P和FMEE复配对铜膜CMP性能的影响

针对化学机械抛光(CMP)中传统唑类缓蚀剂毒性强、成本高,在表面形成坚硬钝化膜难以去除等问题,在甘氨酸-双氧水体系下,使用阴离子表面活性剂聚氧乙烯醚磷酸酯(E1310P)和非离子表面活性剂脂肪酸甲酯乙氧基化物(FMEE)复配替代传统唑类缓蚀剂。通过去除速率、表面粗糙度和表面形貌等的实验结果研究了E1310P和FMEE协同作用对CMP过程中表面质量的影响。通过表面张力、电化学性能、X射线光电子能谱(XPS)和密度泛函理论(DFT)揭示了E1310P和FMEE的协同吸附行为及其机理。结果表明,E1310P可以吸附在Cu的表面,降低Cu的去除速率;FMEE的加入能有效屏蔽E1310P离子头基间的电性排斥作用,使更多的E1310P吸附在Cu的表面,在协同作用下形成了更致密的抑制膜,使得CMP抛光性能得以提升。

A Layered Energy-Efficient Multi-Node Scheduling Mechanism for Large-Scale WSN

In recent years, target tracking has been considered one of the most important applications of wireless sensornetwork (WSN). Optimizing target tracking performance and prolonging network lifetime are two equally criticalobjectives in this scenario. The existing mechanisms still have weaknesses in balancing the two demands. Theproposed heuristic multi-node collaborative scheduling mechanism (HMNCS) comprises cluster head (CH)election, pre-selection, and task set selectionmechanisms, where the latter two kinds of selections forma two-layerselection mechanism. The CH election innovatively introduces the movement trend of the target and establishesa scoring mechanism to determine the optimal CH, which can delay the CH rotation and thus reduce energyconsumption. The pre-selection mechanism adaptively filters out suitable nodes as the candidate task set to applyfor tracking tasks, which can reduce the application consumption and the overhead of the following task setselection. Finally, the task node selection is mathematically transformed into an optimization problem and thegenetic algorithm is adopted to form a final task set in the task set selection mechanism. Simulation results showthat HMNCS outperforms other compared mechanisms in the tracking accuracy and the network lifetime.

A Novel Scheduling Framework for Multi-Programming Quantum Computing in Cloud Environment

As cloud quantum computing gains broader acceptance,a growing quantity of researchers are directing their focus towards this domain.Nevertheless,the rapid surge in demand for cloud-based quantum computing resources has led to a scarcity,which in turn hampers users from achieving optimal satisfaction.Therefore,cloud quantum computing service providers require a unified analysis and scheduling framework for their quantumresources and user jobs to meet the ever-growing usage demands.This paper introduces a new multi-programming scheduling framework for quantum computing in a cloud environment.The framework addresses the issue of limited quantum computing resources in cloud environments and ensures a satisfactory user experience.It introduces three innovative designs:1)Our framework automatically allocates tasks to different quantum backends while ensuring fairness among users by considering both the cloud-based quantum resources and the user-submitted tasks.2)Multi-programming mechanism is employed across different quantum backends to enhance the overall throughput of the quantum cloud.In comparison to conventional task schedulers,our proposed framework achieves a throughput improvement of more than two-fold in the quantum cloud.3)The framework can balance fidelity and user waiting time by adaptively adjusting scheduling parameters.

Dynamic Offloading and Scheduling Strategy for Telematics Tasks Based on Latency Minimization

In current research on task offloading and resource scheduling in vehicular networks,vehicles are commonly assumed to maintain constant speed or relatively stationary states,and the impact of speed variations on task offloading is often overlooked.It is frequently assumed that vehicles can be accurately modeled during actual motion processes.However,in vehicular dynamic environments,both the tasks generated by the vehicles and the vehicles’surroundings are constantly changing,making it difficult to achieve real-time modeling for actual dynamic vehicular network scenarios.Taking into account the actual dynamic vehicular scenarios,this paper considers the real-time non-uniform movement of vehicles and proposes a vehicular task dynamic offloading and scheduling algorithm for single-task multi-vehicle vehicular network scenarios,attempting to solve the dynamic decision-making problem in task offloading process.The optimization objective is to minimize the average task completion time,which is formulated as a multi-constrained non-linear programming problem.Due to the mobility of vehicles,a constraint model is applied in the decision-making process to dynamically determine whether the communication range is sufficient for task offloading and transmission.Finally,the proposed vehicular task dynamic offloading and scheduling algorithm based on muti-agent deep deterministic policy gradient(MADDPG)is applied to solve the optimal solution of the optimization problem.Simulation results show that the algorithm proposed in this paper is able to achieve lower latency task computation offloading.Meanwhile,the average task completion time of the proposed algorithm in this paper can be improved by 7.6%compared to the performance of the MADDPG scheme and 51.1%compared to the performance of deep deterministic policy gradient(DDPG).

A Novel Predictive Model for Edge Computing Resource Scheduling Based on Deep Neural Network

Currently,applications accessing remote computing resources through cloud data centers is the main mode of operation,but this mode of operation greatly increases communication latency and reduces overall quality of service(QoS)and quality of experience(QoE).Edge computing technology extends cloud service functionality to the edge of the mobile network,closer to the task execution end,and can effectivelymitigate the communication latency problem.However,the massive and heterogeneous nature of servers in edge computing systems brings new challenges to task scheduling and resource management,and the booming development of artificial neural networks provides us withmore powerfulmethods to alleviate this limitation.Therefore,in this paper,we proposed a time series forecasting model incorporating Conv1D,LSTM and GRU for edge computing device resource scheduling,trained and tested the forecasting model using a small self-built dataset,and achieved competitive experimental results.

Performance Prediction Based Workload Scheduling in Co-Located Cluster

Cloud service providers generally co-locate online services and batch jobs onto the same computer cluster,where the resources can be pooled in order to maximize data center resource utilization.Due to resource competition between batch jobs and online services,co-location frequently impairs the performance of online services.This study presents a quality of service(QoS)prediction-based schedulingmodel(QPSM)for co-locatedworkloads.The performance prediction of QPSM consists of two parts:the prediction of an online service’s QoS anomaly based on XGBoost and the prediction of the completion time of an offline batch job based on randomforest.On-line service QoS anomaly prediction is used to evaluate the influence of batch jobmix on on-line service performance,and batch job completion time prediction is utilized to reduce the total waiting time of batch jobs.When the same number of batch jobs are scheduled in experiments using typical test sets such as CloudSuite,the scheduling time required by QPSM is reduced by about 6 h on average compared with the first-come,first-served strategy and by about 11 h compared with the random scheduling strategy.Compared with the non-co-located situation,QPSM can improve CPU resource utilization by 12.15% and memory resource utilization by 5.7% on average.Experiments show that the QPSM scheduling strategy proposed in this study can effectively guarantee the quality of online services and further improve cluster resource utilization.

Q-Learning-Assisted Meta-Heuristics for Scheduling Distributed Hybrid Flow Shop Problems

The flow shop scheduling problem is important for the manufacturing industry.Effective flow shop scheduling can bring great benefits to the industry.However,there are few types of research on Distributed Hybrid Flow Shop Problems(DHFSP)by learning assisted meta-heuristics.This work addresses a DHFSP with minimizing the maximum completion time(Makespan).First,a mathematical model is developed for the concerned DHFSP.Second,four Q-learning-assisted meta-heuristics,e.g.,genetic algorithm(GA),artificial bee colony algorithm(ABC),particle swarm optimization(PSO),and differential evolution(DE),are proposed.According to the nature of DHFSP,six local search operations are designed for finding high-quality solutions in local space.Instead of randomselection,Q-learning assists meta-heuristics in choosing the appropriate local search operations during iterations.Finally,based on 60 cases,comprehensive numerical experiments are conducted to assess the effectiveness of the proposed algorithms.The experimental results and discussions prove that using Q-learning to select appropriate local search operations is more effective than the random strategy.To verify the competitiveness of the Q-learning assistedmeta-heuristics,they are compared with the improved iterated greedy algorithm(IIG),which is also for solving DHFSP.The Friedman test is executed on the results by five algorithms.It is concluded that the performance of four Q-learning-assisted meta-heuristics are better than IIG,and the Q-learning-assisted PSO shows the best competitiveness.