P. Mohapatra, C.Yu and C.R.Das, "A Lazy Scheduling Scheme for Hypercube Comput- ers," Journal of Parallel and Distributed Computing, Vol.27 No.1 pp.26-37, May 1995.  Home/Search   Document Details and Download   Summary   Related Articles   Check

....can only be applied effectively when there are multiple jobs of each size, which implies large queueing delays. Thus the opposite approach may be better: whenever a job of a given size is already running, queue any additional requests for subcubes of the same size instead of running them together [410]. Then when the running job terminates, its subcube is reused. Multiple queues for jobs with different requirements can also be used to match jobs with complementary requirements that can run side by side. For example, this is done in PASM, where the different queues represent different potential ....

P. Mohapatra, C. Yu, and C. R. Das, "A lazy scheduling scheme for hypercube computers". J. Parallel & Distributed Comput. 27(1), pp. 26--37, May 1995.

....can only be applied effectively when there are multiple jobs of each size, which implies large queueing delays. Thus the opposite approach may be better: whenever a job of a given size is already running, queue any additional requests for subcubes of the same size instead of running them together [250]. Then when the running job terminates, its subcube is reused. A similar approach is used in NQS, a network queueing system designed to support submission of batch jobs to shared parallel machines. The different queues specify different limits on the number of PEs and the execution time [196] The ....

P. Mohapatra, C. Yu, and C. R. Das, "A lazy scheduling scheme for hypercube computers". J. Parallel & Distributed Comput. 27(1), pp. 26--37, May 1995.

....sharing or multitasking runtime environment gains more attention recently as parallel systems are used as general purpose computing servers where sequential as well as highly parallel programs execute concurrently. Considerable number of allocation algorithms have been proposed for hypercubes [1] [6] and meshes [7] 11] An allocation algo rithm dynamically partitions the interconnection topology and the corresponding processors and assigns subsets of processors to the requesting jobs. The problem of the processor allo cation assumes importance so as to enable higher system utilization and ....

P. Mohapatra, C.Yu and C.R.Das, "A Lazy Scheduling Scheme for Hypercube Comput- ers," Journal of Parallel and Distributed Computing, Vol.27 No.1 pp.26-37, May 1995.

....topology required for task execution. The performance gain observed from efficient allocation schemes is limited mainly by the first come firstserve (FCFS) discipline. A job that cannot be allocated to the system will block all the following jobs from entering the system. The scheduling schemes [9, 11, 12] arrange the order of execution so that jobs could be executed out of the order in which they arrived and thus reduce the blocking effect. However, these scheduling schemes do not provide fairness to various tasks. Some jobs may have to wait for a very long time before they get executed while a ....

....systems combining a priority technique with a reservation scheme is proposed. The reservation scheme allows a job to reserve processors for its execution and let other jobs bypass it to enter the system if they can. The priority technique is used to reduce the chance of fragmentation. The lazy [11] scheduling scheme proposed for hypercubes temporarily delays the allocation of a job if any other job of the same dimension is running. The delayed job is then executed on the existing subcube rather then acquiring a new subcube. The fragmentation of the system and the blocking problem with the ....

P. Mohapatra, C. Yu, and C. R. Das, "A Lazy Scheduling Scheme for Hypercube Computers," Journal of Parallel and Distributed Computing, 27, pp. 26-37, May, 1995.

....time complexity. In spite of its low subcube recognition ability, it has been shown that the buddy strategy performs well in hypercube systems when compared with the perfect recognition allocation schemes [9] The scheduling schemes proposed for non real time hypercubes include scan [9] and lazy [10]. Krishnamurti and Narahari have proposed a preemptive scheduling scheme for non real time, partitionable parallel architectures [8] Hong and Leung have proposed an on line scheduling algorithm for real time tasks with one common deadline [5] Sahni has shown that when the common deadline is more ....

P. Mohapatra, C. Yu, and C. R. Das, "A Lazy Scheduling Scheme for Hypercube Computers, " Journal of Parallel and Distributed Computing, 27, pp. 26-37, May 1995.

....comm so that the turnaround time of a job in Equation 1 can benefit from shorter queuing delay without being penalized for the communication latency. Experimental study indicates that the queuing delay can be greatly reduced resulting in a high DeltaT queue if the fragmentation can be reduced [8, 9, 10, 11]. Design of faster switching devices and wormhole routing techniques have made the message passing latency insensitive to the communication distance making DeltaT comm negligible. Non contiguous allocation is hence a very attractive alternative for processor allocation. Non contiguous allocations ....

P. Mohapatra, C. Yu and C. R. Das, "A Lazy Scheduling Scheme for Hypercube Computers, " Journal of Parallel and Distributed Computing, 27, pp. 26-37, May, 1995.

Online articles have much greater impact   More about CiteSeer.IST   Add search form to your site   Submit documents   Feedback  

CiteSeer.IST - Copyright Penn State and NEC