Showing papers by "Min-You Wu published in 2000"
01 May 2000
TL;DR: An algorithm is presented which improves the min-min algorithm by scheduling large tasks first, balances the load well and demonstrates even better performance in both makespan and running time.
Abstract: The min-min algorithm is a simple algorithm. It runs fast and delivers good performance. However, the min-min algorithm schedules small tasks first, resulting in some load imbalance. We present an algorithm which improves the min-min algorithm by scheduling large tasks first. The new algorithm, segmented min-min, balances the load well and demonstrates even better performance in both makespan and running time.
150 citations
TL;DR: The authors explain the test results they achieved in developing an experimental software tool called CASCH (Computer-Aided SCHeduling), which provides a unified environment for performing automatic parallelization and scheduling of applications without relying on simulations.
Abstract: The authors explain the test results they achieved in developing an experimental software tool called CASCH (Computer-Aided SCHeduling). This system provides a unified environment for performing automatic parallelization and scheduling of applications without relying on simulations.
34 citations
21 Aug 2000
TL;DR: A runtime parallel incremental DAG scheduling approach is described in this paper, which shows that it is superior to other approaches and can also execute dynamic DAGs.
Abstract: A runtime parallel incremental DAG scheduling approach is described in this paper. A DAG is expanded incrementally, scheduled, and executed on a parallel machine. A DAG scheduling algorithm is parallelized to scale to large systems. In this approach, a large DAG can be executed without consuming large amount of memory space. Inaccurate estimation of task execution time and communication time can be tolerated. This runtime approach can also execute dynamic DAGs. Implementation of this parallel incremental system demonstrates the feasibility of this approach. Preliminary results show that it is superior to other approaches.
15 citations
30 Jul 2000
TL;DR: Experimental results show that this novel VBR scheduling algorithm can achieve near 100% network utilization with a negligible block miss rate.
Abstract: Due to the burstiness, scheduling VBR streams and yet achieving high resources utilization on a clustered video server has been considered a difficult problem. We propose a simple but effective VBR scheduling algorithm for generating conflict-free network transmission schedules on clustered VoD systems. Experimental results show that this novel scheduling algorithm can achieve near 100% network utilization with a negligible block miss rate.
6 citations
30 Jul 2000
TL;DR: This work presents an algorithm used to build a schedule table to accommodate interactive operations, especially fast-forward operations, in a feasible parallel Video-on-Demand (VoD) server.
Abstract: To implement a feasible parallel Video-on-Demand (VoD) server the first task is to build a schedule table for each new incoming client request. The table is then updated with time and/or requests are rescheduled as any client changes its play mode. We present an algorithm used to build a schedule table to accommodate interactive operations, especially fast-forward operations. Any client's request can be classified to a particular group according to its playout speed requirement and then scheduled to the corresponding time slots of that group.
1 citations
01 May 2000
TL;DR: A runtime system is described here for dynamic DAG execution that has been parallelized to scale to large systems and preliminary results show that it is superior to other approaches.
Abstract: A runtime system is described here for dynamic DAG execution. A large D A G whihc represents an application program can be executed on a parallel system without consuming large amount of memory space. A DAG scheduling algorithm has been parallelized to scale to large systems. Inaccurate estimation of task execution time and comm unication time can be tolerated. Implementation of this parallel incremental system demonstrates the feasibility of this approach. Preliminary results show that it is superior to other approaches.