P. Mohapatra, C. Yu, C.R. Das, and J. Kim, "A Lazy Scheduling Scheme for Improving Hypercube Performance," Proc. 1993.  Home/Search   Document Details and Download   Summary   Related Articles   Check

....time. There has been considerable prior research into each of the two topics of i) scheduling of parallel jobs [1,2,6,10,14,19,20,23,24,26] and ii) contiguous node allocation strategies [3,4,5,11,13,27] There have also been a few studies that have considered both these issues in combination [12,15,18]. However, only [15] addresses the impact of contiguous node allocation schemes in conjunction with a job scheduling policy that takes fairness into consideration by use of a FCFS (First Come First Served) scheduling policy. In [15] contiguous and non contiguous node allocation schemes for ....

....in [17] 16] presents a strategy that minimizes network contention due to both communication and I O traffic. All the above studies focus exclusively on the topic of contiguous node allocation schemes, but do not address the issue of job scheduling onto the parallel systems. The studies in [12,18] focus on job scheduling techniques for improving the performance of hypercube computers. In [12] the roles of processor allocation and job scheduling in improving the performance of hypercube computers are compared. A new scheduling algorithm is proposed for improving the average response time ....

[Article contains additional citation context not shown here]

P. Mohapatra, C. Yu, C.R. Das, and J. Kim, "A Lazy Scheduling Scheme for Improving Hypercube Performance," Proc. Int'l Conf. Parallel Processing, 1993.

....which arises from the contiguity constraint. Although no statistics have been compiled for mesh systems, we believe the same trend will be exhibited under the assumption of contiguity. Thus, improved performance requires exploration of other alternatives, including scheduling policies [3] [21] [24] and the approach we propose: non contiguous allocation. 3 Non contiguous Allocation Strategies 3.1 Random Allocation Strategy One of the most straightforward non contiguous allocation strategies is the Random allocation strategy, under which a request for k processors is satisfied with k ....

P. Mohapatra, C. Yu, C. R. Das, and J. Kim. A lazy scheduling scheme for improving hypercube performance. In International Conference on Parallel Processing, 1993.

....task is the time since the task enters the system until it exits. It is the sum of the delay time in the queue and the service time. From a task point of view, the average response time is more important than the overall system utilization. In fact these two metrics are not proportionally related [9]. Evaluation results We first evaluate the algorithms for hypercubes. Six test cases were studied. In each case two versions of the algorithms were experimented: one without reordering and the other with reordering. For the cases with reordering, the reordering success rate was also recorded. ....

P. Mohapatra, C. Yu, C.R. Das and J. Kim, "A lazy scheduling scheme for improving hypercube performance," Proc. of Int. Conf. on Parallel Processing, Aug. 1993

....in a round robin fashion All tasks in the same class are allocated before the scheduler moves on to the next class. One major problem with the eager scheduling policies are that they tend to favor those requesting a smaller number of processors and thus may increase the fragmentation in the system [3]. Lazy scheduling, on the other hand, delays the scheduling of some tasks so that more appropriate tasks may be accommodated later. In this way tasks with different requests for free processors will have equal opportunities to schedule. In the lazy scheduling policy LAZY [3] this is done by ....

....in the system [3] Lazy scheduling, on the other hand, delays the scheduling of some tasks so that more appropriate tasks may be accommodated later. In this way tasks with different requests for free processors will have equal opportunities to schedule. In the lazy scheduling policy LAZY [3], this is done by assigning, as best as one can, the subcube released from a completed task to a task requesting exactly the same number of processors. Unfortunately, lazy scheduling by its nature is conservative in scheduling. Tasks may be kept in the queue even if there are free processors ....

Mohapatra, P., Yu, C., Das, C.R., Kim, J.: A lazy scheduling scheme for improving hypercube performance. Proc. of Int. Conf. on Parallel Processing (1993) This article was processed using the L A T E X macro package with LLNCS style

....which arises from the contiguity constraint. Although no statistics have been compiled for mesh systems, we believe the same trend will be exhibited under the assumption of contiguity. Thus, improved performance requires exploration of other alternatives, including scheduling policies [2] [8] [11] and the approach we propose: non contiguous allocation. 3 Contention on Real Systems The potential increased message contention caused by non contiguous allocation could greatly reduce performance. As a first step in evaluating the feasibility of non contiguous allocation, we measured ....

Prasant Mohapatra, Chansu Yu, Chita R. Das, and Jong Kim. A lazy scheduling scheme for improving hypercube performance. In International Conference on Parallel Processing, pages 110--117, 1993.

....different job scheduling policies need to be considered. A good job scheduling policy should provide low waiting time for the input jobs, cause low overhead, and guarantee a certain fairness. In hypercube, a reservation idea has been proposed to reserve a subcube of processors for a scheduled job [12], which has been shown to improve the performance. For a torus or mesh system, we may adopt the similar idea. If a job request could not be allocated due to insufficient free processors, a submesh can be reserved for the job. However, some problems have to be considered. For example, how and where ....

P. Mohapatra and et al, "A lazy scheduling scheme for improving hypercube performance," International Conference on Parallel Processing, vol. 1, pp. 110--117, 1993.

No context found.

P. Mohapatra, C. Yu, C.R. Das, and J. Kim, "A Lazy Scheduling Scheme for Improving Hypercube Performance," Proc. 1993.

....this threshold value, it gets priority over incoming jobs as well as all jobs in the system queue. No waiting jobs are serviced until it is allocated. A predefined threshold value is useful in imposing deadline for job completion. A heuristic may be used to dynamically derive the value of ffit [3]. We use a predefined threshold value in this paper. SJF Scheduling Let J head denote the job at the head of the system queue and Deadline[J i ] denote the deadline of job J i . Job Arrival: 1. Let J i be the incoming job. 2. If current time Deadline[J head ] then Deadline[J i ] current ....

P. Mohapatra, C. Yu, C. R. Das and J. Kim, "A Lazy Scheduling Scheme for Improving Hypercube Performance," Proc. Int'l Conf. on Parallel Processing, Vol. I, pp. 110-117, Aug. 1993.

No context found.

P.Mohapatra, C.Yu, C.R.Das and J.Kim, "A Lazy Scheduling Scheme for Improving Hypercube Performance," Proc. Int. Conf. Parallel Processing, pp.I-110-117, 1993.

....policies strive to improve system performance in terms of low response time, high system utilization and high throughput. However, comparison of the aforementioned hypercube allocation policies shows that performance improvement through better subcube recognition ability is not significant [8][9]. The difference between a policy having the complete subcube recognition ability and the simple buddy scheme is minimal. Task miss ratio is relatively high and system utilization attains a maximum of 50 in spite of the most sophisticated allocation policy [9] In order to improve the system ....

....ability is not significant [8] 9] The difference between a policy having the complete subcube recognition ability and the simple buddy scheme is minimal. Task miss ratio is relatively high and system utilization attains a maximum of 50 in spite of the most sophisticated allocation policy [9]. In order to improve the system performance, it is more important to address the first type of overflow the fragmentation problem. Fragmentation of the free nodes often does not allow the formation of a subcube large enough to accommodate an incoming job. An allocator has to wait until some of ....

[Article contains additional citation context not shown here]

P.Mohapatra, C.Yu, C.R.Das and J.Kim, "A Lazy Scheduling Scheme for Improving Hypercube Performance," Proc. Int. Conf. Parallel Processing, pp.I-110-117, 1993.

....algorithm is proposed for assigning incoming tasks whose resource requirements are not cubic. The objective here is to minimize internal fragmentation. Comparison of all the hypercube allocation policies shows that performance improvement with an efficient allocation policy is not significant [18][19]. All these schemes provide almost equal performance and system utilization suffers due to FCFS scheduling. It is possible to improve system performance significantly by changing the scheduling scheme from FCFS to Scan [18] or Lazy [19] while using the simple buddy allocation algorithm. The same ....

....with an efficient allocation policy is not significant [18] 19] All these schemes provide almost equal performance and system utilization suffers due to FCFS scheduling. It is possible to improve system performance significantly by changing the scheduling scheme from FCFS to Scan [18] or Lazy [19] while using the simple buddy allocation algorithm. The same trend is also observed on a MIN machine with buddy allocation and Lazy scheduling. 3 We conduct a simulation study to analyze the performance of buddy strategy with the conventional matching pattern and buddy and free list schemes with ....

[Article contains additional citation context not shown here]

P.Mohapatra, C.Yu, C.R.Das and J.Kim, "A Lazy Scheduling Scheme for Improving Hypercube Performance," Proc. Int. Conf. Parallel Processing, pp.I-110-117, 1993.

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