J. D. Padhye and L. W. Dowdy. Preemptive versus non-preemptive processor allocation policies for message passing parallel computers: An empirical comparison. In Proceedings of the 2nd Workshop on Job Scheduling Strategies for Parallel Processing, April 1996.  Home/Search   Document Details and Download   Summary   Related Articles   Check

....exhibits lower mean response times, as well as a smaller variance in the execution times, than that observed for dynamic equi partitioning and folding, with equi partitioning consistently providing better response times than folding. The latter result is in contrast to those of Padhye and Dowdy [29] which show that folding generally outperforms equi partitioning in a distributed memory environment under a workload based on scientific matrix computation programs. The differences between the results of these two studies are primarily due to the differences in the respective workloads, where ....

....workload based on scientific matrix computation programs. The differences between the results of these two studies are primarily due to the differences in the respective workloads, where the workload used in our experiments consists of applications with larger execution times than those studied in [29]. This further highlights the fundamental tradeoff between the two above dynamic partitioning approaches for reducing the repartitioning overheads in distributed memory environments. We are continuing to examine these and related scheduling issues in distributed memory parallel systems. 18 ....

J. D. Padhye and L. W. Dowdy. Preemptive versus non-preemptive processor allocation policies for message passing parallel computers: An empirical comparison. In Proceedings of the 2nd Workshop on Job Scheduling Strategies for Parallel Processing, April 1996.

....is one that changes of its own accord. However, note that evolving and malleable should usually come together, because one job s evolution will cause others to have to reconfigure. Analyzing the benefits of dynamic partitioning and malleable jobs has been the subject of much recent research [36,7,28,39,34,23]. This research typically compares the cost of reconfiguration with the resulting improvement in overall performance. But such comparisons do not give a full picture. In many cases, changing the number of processors allocated to a job requires complex interactions between the operating system and ....

....that has a balanced workload over a particular partition size is likely to remain balanced after a folding operation. Many speedup curves resemble step functions, with poor speedup values for non powers of two number of processors. However, there is some debate over the benefits of folding [28,39]. 4.2 Preemption and Time Slicing Dynamic partitioning, discussed above, requires certain processors to be preempted and re allocated in order to accommodate load changes. Another type of preemption is that used in order to time slice multiple applications, as is commonly done on uniprocessors. ....

J. D. Padhye and L. W. Dowdy, "Preemptive versus non-preemptive processor allocation policies for message passing parallel computers: an empirical comparison ". In Job Scheduling Strategies for Parallel Processing II, D. G. Feitelson and L. Rudolph (eds.), Springer-Verlag, 1996. Lecture Notes in Computer Science.

....for numerical integration. Matrix Multiplication performs matrix multiplication in parallel. Fast Fourier Transform is a parallel version of fast fourier transform. Parallel Make is a parallel version of the DQS make program. Conjugent Gradient performs a conjugate gradient algorithm in parallel [33]. AtEarth and Adaptive Quadrature are bag of tasks applications. Conjugate Gradient, on the other hand, is a symmetric application. Sim is a synthetic application that consists of small parallel applications with irregular finishing times for its constituent processes. Workloads We studied a ....

J. D. Padhye and L. W. Dowdy. Preemptive versus non-preemptive processor allocation policies for message passing parallel computers: An empirical comparison. In Job Scheduling Strategies for Parallel Processing, D. G. Feitelson and L. Rudolph (eds.), pages 224--243. Springer-Verlag, 1996. Lecture Notes in Computer Science Vol. 1162.

....smoothing exhibits lower mean response times, as well as a smaller variance in the execution times, than that observed for dynamic equi partitioning and folding, with equi partitioning consistently providing better response times than folding. This is in contrast to the results of Padhye and Dowdy [30] which show that folding generally outperforms equi partitioning in a distributed memory environment under a workload based on scientific matrix computation programs. The di#erences between the results of these two studies are primarily due to the di#erences in the respective workloads, where the ....

....a workload based on scientific matrix computation programs. The di#erences between the results of these two studies are primarily due to the di#erences in the respective workloads, where the workload used in our experiments consists of applications with larger execution times than those studied in [30]. This further highlights the fundamental tradeo# between these two dynamic partitioning approaches for reducing the repartitioning overheads in distributed memory environments. We are continuing to examine these and related scheduling issues in distributed memory parallel systems. ....

J. D. Padhye and L. W. Dowdy. Preemptive versus non-preemptive processor allocation policies for message passing parallel computers: An empirical comparison. In Proceedings of the 2nd Workshop on Job Scheduling Strategies for Parallel Processing, April 1996.

....section briefly describes the three scheduling policies studied in this paper, the two workload programs used for comparisons, and the architecture of the Intel Paragon. Due to space constraints, this section provides only a brief overview of the actual implementation. More details may be found in [P96]. 2.1 The Intel Paragon The Intel Paragon supercomputer consists of several nodes connected in a mesh configuration. The computer used for this study has 66 nodes connected in a 11x6 matrix. Several nodes are dedicated to special tasks such as disk and network control. Each node consists of two ....

J. Padhye, "Preemptive versus non-preemptive processor allocation policies: an empirical comparison", Technical Report, Department of Computer Science, Vanderbilt University, 1996.

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