P. Kacsuk. Wavefront scheduling in logflow. In Proc. of the Euromicro Workshop on Parallel and Distributed Processing, pages 503--510. IEEE Computer Society Press, 1994. Home/Search Document Not in Database Summary Related Articles Check |
This paper is cited in the following contexts:
Side-effects in PloSys OR-parallel Prolog on distributed.. - Morel Briat Chassin (1996) (2 citations) (Correct)
....will be continued, in particular to test the cut and side effects implementation and their relation with the scheduling policies. In the longer term, the prototype will be entirely distributed. The distributed scheduling algorithm will use some of the techniques developed in the LOGFLOW system [21, 14] for which there already exists an animation tool [20] that can be used to tune load balancing strategies. Another foreseen extension is to combine constraint propagation techniques, already available in the inference engine wamcc [11] with ORparallelism. Such a combination was already proven to ....
P. Kacsuk. Wavefront scheduling in logflow. In Proc. of the Euromicro Workshop on Parallel and Distributed Processing, pages 503--510. IEEE Computer Society Press, 1994.
Graphical Environments to Support Parallel Program Development - Peter Kacsuk Research (1997) (1 citation) Self-citation (Kacsuk) (Correct)
....the LOGFLOW system is able to collect information about the state of the processors. This information are written into a file at the end of the execution. This method is called post mortem visualisation and was required for analysing the efficiency of the Wavefront Scheduler of the LOGFLOW system [6]. In the load balancing visualizer there are columns representing the following information: position: identify processors . height: shows the quantity of tokens (jobs) on the processor . colour: determines the type of token (job) on the processor . pattern: gives information about which ....
P. Kacsuk, Wavefront scheduling in LOGFLOW, Proceedings of Euromicro Workshop on Parallel and Distributed Processing, Malaga, 1994. pp. 503-510.
Granularity Control In The Logflow Parallel Prolog System - Kacsuk Self-citation (Kacsuk) (Correct)
....to give exact predictions on them. Logic programs typically belong to this class of irregular problems. Accordingly, handling of granularity and load balancing was one of the main issue in the design of the LOGFLOW system. To solve the load balancing problem the so called Wavefront Scheduler [7] has been proposed and applied on multitransputer systems with success. In the current paper, the granularity control mechanism of LOGFLOW is described. Since the LOGFLOW execution mechanism combines a very fine grain data driven scheme and the well known coarse grain WAM based scheme, the LOGFLOW ....
Kacsuk, P. (1994) Wavefront Scheduling in LOGFLOW, in Proc. of the 2nd EUROMICRO Workshop on Parallel and Distributed Processing, Malaga, 503-510.
Tools for Mapping, Load Balancing and Monitoring in the.. - Kacsuk, Németh, Puskás Self-citation (Kacsuk) (Correct)
....related to Prolog execution. The main features of these versions of LOGFLOW are compared in Section 2. Load balancing is crucial to achieve high performance in both types of distributed memory architectures. Two scheduling schemes have been developed in the LOGFLOW project: the Wavefront Scheduler [10] is employed for the multitransputer implementation, and the Distributed Token Scheduler [20] is introduced for the workstation cluster implementations. These scheduling schemes are described and compared in Section 3. The transputer version of LOGFLOW has been ported to a 34 processor Supernode ....
....Wavefront scheduling for LOGFLOW The so called Wavefront Scheduler has been implemented in LOGFLOW. The first implementation of the scheduler dealt with the horizontal distribution of work only, i.e. newly generated tokens were sent to the neighbouring processing element with the lowest workload [10]. The workload of neighbouring processing elements are approximated and measured by the length of token queues However, the scheduler did not change the type of call, i.e. the parallel call remained parallel and the sequential one remained sequential. Results from performance monitoring showed ....
[Article contains additional citation context not shown here]
P. Kacsuk, Wavefront Scheduling in LOGFLOW, in 2 nd EUROMICRO Workshop on Parallel and Distributed Processing, Malaga, 1994. 503-510.
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