On Task Relocation in Two-Dimensional Meshes

In parallel computer systems with a number of processors, external fragmentation is caused by continuous allocation and deallocation of processors to tasks which require exclusive use of several contiguous processors. With this condition, the system may not be able to find contiguous processors to be allocated to an incoming task even with a sufficient number of free processors. Relocation is an approach for alleviating this problem by reassigning the running tasks to other processors. In this paper, we examine two relocation schemes—full relocation and partial relocation scheme—for two-dimensional meshes. The full relocation scheme is desirable when the system is highly fragmented, while the partial relocation scheme is used for minimizing the number of relocated tasks. For the relocation process, we formally define and use two basic submesh movement operations—shifting and rotating. Comprehensive computer simulation reveals that the proposed schemes are beneficial when the relocation overhead is not high, which is machine dependent.     

Task migration in all-port wormhole-routed 2D mesh multicomputers

In a mesh multicomputer, submeshes are allocated to perform jobs according to processor allocation schemes, with each task assigned to occupy processors of one submesh with an appropriate size. To assign regions for incoming tasks, task compaction is needed to produce a large contiguous free region. The overhead of task compaction relies mainly on designing an efficient task migration scheme. This paper investigates task migration schemes in 2D wormhole-routed mesh multicomputers with an all-port communication model. Two constraints are given between two submeshes for task migration. Two task migration schemes that follow one of the constraints in 2D mesh multicomputers are then developed. In addition, the proposed schemes are proven to be deadlock-free and congestion-free. Finally, performance analysis is adopted to compare the proposed task migration schemes.     

Using task migration to improve non-contiguous processor allocation in NoC-based CMPs

In this paper, a processor allocation mechanism for NoC-based chip multiprocessors is presented. Processor allocation is a well-known problem in parallel computer systems and aims to allocate the processing nodes of a multiprocessor to different tasks of an input application at run time. The proposed mechanism targets optimizing the on-chip communication power/latency and relies on two procedures: processor allocation and task migration. Allocation is done by a fast heuristic algorithm to allocate the free processors to the tasks of an incoming application when a new application begins execution. The task-migration algorithm is activated when some application completes execution and frees up the allocated resources. Task migration uses the recently deallocated processors and tries to rearrange the current tasks in order to find a better mapping for them. The proposed method can also capture the dynamic traffic pattern of the network and perform task migration based on the current communication demands of the tasks. Consequently, task migration adapts the task mapping to the current network status. We adopt a non-contiguous processor allocation strategy in which the tasks of the input application are allowed to be mapped onto disjoint regions (groups of processors) of the network. We then use virtual point-to-point circuits, a state-of-the-art fast on-chip connection designed for network-on-chips, to virtually connect the disjoint regions and make the communication latency/power closer to the values offered by contiguous allocation schemes. The experimental results show considerable improvement over existing allocation mechanisms.     

Task allocation techniques for distributed computing systems: a review

An efficient strategy to allocate tasks to processors in distributed computer systems is required for proper utilization of available computational power. Various task allocation models developed so far have the basic assumption that an application program is suitably divided into a number of subtasks or modules and the execution costs of modules and intermodule communication costs are also known. The subdivision of a given program into modules is assumed to be carried out by a software mechanism which forms a separate discipline of research work.The task allocation models developed earlier did not consider various constraints like load balancing, precedence relations, timing constraints, etc. Most of the recent works in the formulation of task allocation strategies take into account a number of constraints. This paper presents a brief outline of the various available models and algorithms for task allocation with suggestions for possible future directions in the field.     

Power and performance management for parallel computations in clouds and data centers

We address scheduling independent and precedence constrained parallel tasks on multiple homogeneous processors in a data center with dynamically variable voltage and speed as combinatorial optimization problems. We consider the problem of minimizing schedule length with energy consumption constraint and the problem of minimizing energy consumption with schedule length constraint on multiple processors. Our approach is to use level-by-level scheduling algorithms to deal with precedence constraints. We use a simple system partitioning and processor allocation scheme, which always schedules as many parallel tasks as possible for simultaneous execution. We use two heuristic algorithms for scheduling independent parallel tasks in the same level, i.e., SIMPLE and GREEDY. We adopt a two-level energy/time/power allocation scheme, namely, optimal energy/time allocation among levels of tasks and equal power supply to tasks in the same level. Our approach results in significant performance improvement compared with previous algorithms in scheduling independent and precedence constrained parallel tasks.     

A comparative study of job allocation and migration in the pancake network

Pancake networks are an attractive class of Cayley graphs functioning as a viable interconnection scheme for large multi-processor systems. The hierarchy of the pancake graph allows the assignment of its special subgraphs, which have the same topological features as the original graph, to a sequence of incoming jobs. We investigate the hierarchical structure of the pancake network and derive a job allocation scheme for assigning processors to incoming jobs. An algorithm is presented for job migration. Finally, we compare the assignment scheme to those derived previously for the star network and address the shortcomings of the pancake network.     

Allocating precise submeshes in mesh connected systems

We propose a new processor allocation strategy that applies to any mesh system and recognizes submeshes of arbitrary sizes at any locations in a mesh system. The proposed strategy allocates a submesh of exactly the size requested by an incoming task, completely avoiding internal fragmentation. Because of its efficient allocation, this strategy exhibits better performance than an earlier allocation strategy based on the buddy principle. An efficient implementation of this strategy is presented. Extensive simulation runs are carried out to collect experimental cost and performance measures of interest under different allocation schemes     

Multi-heuristic dynamic task allocation using genetic algorithms in a heterogeneous distributed system

We present a multi-heuristic evolutionary task allocation algorithm to dynamically map tasks to processors in a heterogeneous distributed system. It utilizes a genetic algorithm, combined with eight common heuristics, in an effort to minimize the total execution time. It operates on batches of unmapped tasks and can preemptively remap tasks to processors. The algorithm has been implemented on a Java distributed system and evaluated with a set of six problems from the areas of bioinformatics, biomedical engineering, computer science and cryptography. Experiments using up to 150 heterogeneous processors show that the algorithm achieves better efficiency than other state-of-the-art heuristic algorithms.     

New Products

EAI's new hybrid computer system, called Hyshare, is aimed at the multi-user, multi-task application demands of larger- scale simulation and scientific computation laboratories. Consisting of an EAI 3200 digital computer and up to six EAI analog processors, the analog/digital and digital/ analog communications interface employs on-line, dynamic resource allocation techniques which allow analog processors to be assigned to separate tasks or linked together to meet specific application requirements.     

同构型分布式计算机系统的启发式任务分配算法

本文讨论一种启发式任务分配方法，称之为改进的ｌｉｓｔ分配方法，它适用于分配一组具有先后关系和通信延迟的任务集到同构型分布式计算机系统上。文中描述了此分配方法的原理和算法，给出相应的仿真流程图，并对具有不同拓扑结构，任务运行时间和通信时间满足多种概率分布的任务集进行了分析和仿真。结果表明，当处理器个数小于任务集的并行度，任务粒度大于５时，任务分配效率大于８０％。     

Analysis of processor allocation in multiprogrammed,distributed-memory parallel processing systems

A main objective of scheduling independent jobs composed of multiple sequential tasks in shared-memory and distributed-memory multiprocessor computer systems is the assignment of these tasks to processors in a manner that ensures efficient operation of the system. Achieving this objective requires the analysis of a fundamental tradeoff between maximizing parallel execution, suggesting that the tasks of a job be spread across all system processors, and minimizing synchronization and communication overheads, suggesting that the job's tasks be executed on a single processor. The authors consider a class of scheduling policies that represent the essential aspects of this processor allocation tradeoff, and model the system as a distributed fork-join queueing system. They derive an approximation for the expected job response time, which includes the important effects of various parallel processing overheads (such as task synchronization and communication) induced by the processor allocation policy     

混合粒子群算法的异构多核处理器间任务调度

针对异构多核处理器间的任务调度问题,为了更好地发挥异构多核处理器间的平台优势,提出一种基于将有关联的且不在同一处理器上的任务进行复制的思想,从而使每个异构多核的处理器能独立执行任务,来减少不同处理器之间的通信开销,并且通过混合粒子群算法(HPSO)来调度异构多核处理器中的任务,避免由于当任意一个异构多核处理器由于任务分配过多而导致计算机不能及时且准确地得出结果.最后实验证明,对比传统的启发式分配方案和常见的遗传算法(GA),基于任务复制思想分配方案和混合粒子群算法(HPSO)具有更好的求解能力,并且可以提供执行时间更少的调度分配方案,具有较好的应用价值.     

Task migration in n-dimensional wormhole-routed mesh multicomputers

In a mesh multicomputer, performing jobs needs to schedule submeshes according to some processor allocation scheme. In order to assign the incoming jobs to a free submesh, a task compaction scheme is needed to generate a larger contiguous free region. The overhead of compaction depends on the efficiency of the task migration scheme. In this paper, two simple task migration schemes are first proposed in n-dimensional mesh multicomputers with supporting dimension-ordered wormhole routing in one-port communication model. Then, a hybrid scheme which combines advantages of the two schemes is discussed. Finally, we evaluate the performance of all of these proposed approaches.     

Effective reformulations for task allocation in distributed systems with a large number of communicating tasks

In any distributed processing environment, decisions need to be made concerning the assignment of computational task modules to various processors. Many versions of the task allocation problem have appeared in the literature. Intertask communication makes the assignment decision difficult; capacity limitations at the processors increase the difficulty. This problem is naturally formulated as a nonlinear integer program, but can be linearized to take advantage of commercial integer programming solvers. While traditional approaches to linearizing the problem perform well when only a few tasks communicate, they have considerable difficulty solving problems involving a large number of intercommunicating tasks. This paper introduces new mixed integer formulations for three variations of the task allocation problem. Results from extensive computational tests conducted over real and generated data indicate that the reformulations are particularly efficient when a large number of tasks communicate, solving reasonablylarge problems faster than other exact approaches available.     

Processor scheduling and allocation for 3D torus multicomputersystems

Multicomputer systems achieve high performance by utilizing a number of computing nodes. Recently, by achieving significant reductions in communication delay, the three-dimensional (3D) torus has emerged as a new candidate interconnection topology for message-passing multicomputer systems. In this paper, we propose an efficient processor allocation scheme-scan search scheme-for the 3D torus based on a first-fit approach. The scan search scheme minimizes the average allocation time for an incoming task by effectively manipulating the 3D information on a torus as 2D information using a data structure called the CST (Coverage Status Table). Comprehensive computer simulation reveals that the allocation time of the scan search scheme is always smaller than that of the earlier scheme based on a best-fit approach. The difference gets larger as the input load increases, and it is as much a factor of 3 for high load. To investigate the performance of the proposed scheme in different scheduling environments, we also consider a non-FCFS scheduling policy along with the typical FCFS policy. The allocation time complexity of the scan search scheme is O(LW²H²). This is significantly smaller than that of the existing scheme which is O(L⁴W⁴H⁴). Here, L, W, and H represent the length, width, and height of 3D torus, respectively     

How useful is old information?

We consider the problem of load balancing in dynamic distributed systems in cases where new incoming tasks can make use of old information. For example, consider a multiprocessor system where incoming tasks with exponentially distributed service requirements arrive as a Poisson process, the tasks must choose a processor for service, and a task knows when making this choice the processor queue lengths from T seconds ago. What is a good strategy for choosing a processor in order for tasks to minimize their expected time in the system? Such models can also be used to describe settings where there is a transfer delay between the time a task enters a system and the time it reaches a processor for service. Our models are based on considering the behavior of limiting systems where the number of processors goes to infinity. The limiting systems can be shown to accurately describe the behavior of sufficiently large systems and simulations demonstrate that they are reasonably accurate even for systems with a small number of processors. Our studies of specific models demonstrate the importance of using randomness to break symmetry in these systems and yield important rules of thumb for system design. The most significant result is that only small amounts of queue length information can be extremely useful in these settings; for example, having incoming tasks choose the least loaded of two randomly chosen processors is extremely effective over a large range of possible system parameters. In contrast, using global information can actually degrade performance unless used carefully; for example, unlike most settings where the load information is current, having tasks go to the apparently least loaded server can significantly hurt performance     

Cluster-based multicore real-time mixed-criticality scheduling

Cluster-based scheduling is recently gaining importance to be applied to mixed-criticality real-time systems on multicore processors platform. In this approach, the cores are grouped into clusters, and tasks that are partitioned among different clusters are scheduled by global scheduler in each cluster. This research work introduces a new cluster-based task allocation scheme for the mixed-criticality real-time task sets on multicore processors. For task allocation, smaller clusters sizes (sub-clusters) are used for mixed-criticality tasks in low criticality mode, while relatively larger cluster sizes are used for high criticality tasks in high criticality mode. In this research paper, the mixed-criticality task set is allocated to clusters using worst-fit heuristic. The tasks from each cluster are also allocated to its sub-clusters, using the same worst-fit heuristic. A fixed-priority response time analysis approach based on Audsley’s approach is used for the schedulability analysis of tasks in each cluster and sub-cluster. If the high criticality job is not completed after its worst case execution time in low mode, then the system is switched to high criticality mode. After mode switch, all the low criticalities tasks are discarded and only high criticality tasks are further executed in high criticality mode. Simulation results indicate that the percentage of schedulable task sets significantly increases under cluster scheduling as compared to partitioned and global mixed-criticality scheduling schemes.     

优化能耗的可变电压禁忌任务调度算法

能耗是影响异构式并行和分布式系统性能的一个重要因素,动态电压缩放(DVS)技术通过将处理器降低到不同频率来达到有效地节约能耗的目标。通常DVS技术包含任务调度及空闲时间片分配两阶段。当前绝大部分研究均针对时间片分配阶段,而在此考虑的是任务分配与空闲时间片间的关系。为了降低异构分布式系统的能耗,提出了一个利用禁忌(Tabu)策略进行调度的DVS算法。此算法首先调度用有向无环图(DAG)表示的任务集到处理器上,再应用禁忌策略来改进它,通过禁止任务再调度到特定处理器,从而增加时间片,分配阶段可用的空闲时间片达到进一步减少能耗的目标。仿真结果表明,本算法能有效地减少计算机系统的能耗。     

Task migration in three-dimensional meshes

As a result of the emerging use of mesh-based multicomputers (and recently mesh-based multiprocessor systems-on-chip), issues related to processor management have attracted much attention. In a mesh-based multiprocessor, after repeated submesh allocations and de-allocations, the system network may be fragmented, i.e. there might be unallocated nodes in the network. As a result, in a system with contiguous processor allocation, no new tasks can start running due to the lack of enough free adjacent processors to form a suitable submesh. Although there might be enough free processors available, they remain idle until the allocator can find a set of adjacent free nodes forming a submesh to be used for the new task. This can lead to low system performance. Task migration was introduced as a solution to this problem through migration of tasks running on some submeshes to other free areas in order to reduce fragmentation by chaining the newly freed areas and disengaging nodes to form larger submeshes. In this paper, we propose a novel structured and formulated way to code task migration, which is helpful for congestion detection in different steps of task migration algorithms. Moreover, considering the fact that the 3D mesh-based multicomputers are now very popular, a new task migration algorithm in 3D meshes is proposed. We also address the special case of the 2D migration in a 3D mesh multicomputer.     

分布式容错系统的任务分配算法

文章提出了分布式容错系统的任务分配算法,该算法考虑了系统任务的周期性、冗余性特点,以处理机负载平衡为目标,通过三步静态分配实现了任务在处理机中的冗余分布,在系统执行过程中的处理机故障,通过冗余任务动态唤醒实现系统重构。