Holly Dail. A modular framework for adaptive scheduling in grid application development environments. Master's thesis, University of California at San Diego, March 2002. Available as UCSD Tech. Report CS2002-0698.  Home/Search   Document Details and Download   Summary   Related Articles   Check

....makes it feasible, in principle, to execute even communication intensive applications on distributed computation and storage resources [22] In heterogeneous environments, however, the discovery and configuration of suitable resources for applications remain challenging problems. Like others [1, 6, 11, 21, 23, 28], we postulate the existence of a resource selector service responsible for selecting Grid resources appropriate for a particular problem run based on that run s characteristics; organizing those resources into a virtual machine with an appropriate topology; and potentially also assisting with the ....

Dail, H. A Modular Framework for Adaptive Scheduling in Grid Application Development Environments Computer Science, University of California San Diego, 2002.http://grail.sdsc.edu

....for the application, but it assumes that communication is the major factor determining the execution time of the algorithm. Other greedy techniques using different orderings have been implemented within the GrADS resource selection process, for example, using CPU load to order the machines [8]. 2 Global Scheduling Strategies The scheduling problem can be viewed as an multivariate optimization problem, where the application is being assigned to a set of machines so as to optimize some metric (i.e. the overall execution time) Techniques exist for finding locally optimal solutions for ....

Holly Dail. A Modular Framework for Adaptive Scheduling in Grid Application Development Environments. Master's thesis, University of California, San Diego, 2002.

....iteration time on the fastest processor. The default search tolerance is 0.01, meaning that the search re nement ends when the relax factor changes by less than 0.01 between search steps. We now brie y describe our speci cation of this problem as a constrained optimization problem; see [11] for a thorough explanation and [13, 36] for previous work that applied a similar solution for the data mapping problem. The unknowns are the strip widths to be assigned to each processor: n 0 ; n p 1 . Since strip widths are constrained to integer values, the problem can be framed as an ....

.... predicted application performance (predTime) with actual application performance (actualT ime) We calculate the prediction error as: predError = 100 predTime actualT ime actualT ime : 9) We do not have space here to fully describe our experimental design; a full explanation is available in [11]. We tested model accuracy for both the Jacobi and the Game of Life applications on both a single site testbed and a three site testbed (see Section 4.1 for details) In total, we obtained 344 comparisons of actual and predicted times. A histogram of the prediction errors we measured in those ....

Dail, H. A modular framework for adaptive scheduling in grid application development environments. Master's thesis, University of California at San Diego, March 2002. Available as UCSD Tech. Report CS2002-0698.

....schedules are compared to find the schedule with the minimum predicted execution time; this schedule is returned as the bestSched. Given a base machine list of size p with s distinct sites, the upper bound on the number of CMGs that must be evaluated by our search procedure is 3p2 s ; see [8] for details of this bound development. An exhaustive search requires evaluation of 2 CMGs. As long as the number of sites is significantly smaller than the number of resources (universally true in production Computational Grids today) then our search procedure greatly reduces search space as ....

....We frame work allocation constraints as a constrained optimization problem and we use the freely available lp solve package [23] to solve it. Our mappers also re arrange machine ordering to limit wide area communication costs. For details on these designs the interested reader is referred to [8]. Testbeds Our experiments were performed on a subset of the GrADS testbed composed of workstations at the University of Tennessee, Knoxville (UTK) the University of Illinois, Urbana Champaign (UIUC) and the University of California, San Diego (UCSD) At UTK and at UCSD the resources we ....

DAIL, H. A modular framework for adaptive scheduling in grid application development environments. Master's thesis, University of California at San Diego, March 2002. Available as UCSD Tech. Report CS2002-0698.

No context found.

Holly Dail. A modular framework for adaptive scheduling in grid application development environments. Master's thesis, University of California at San Diego, March 2002. Available as UCSD Tech. Report CS2002-0698.

No context found.

Dail, H.J., A Modular Framework for Adaptive Scheduling in Grid Application Development Environments. Computer Science, University of California, California, San Diego, 2001.

No context found.

Dail, H. "A Modular Framework for Adaptive Scheduling in Grid Application Development Environments," Masters. Thesis, University of California at San Diego (2002), Available as UCSD Tech. Report CS2002-0698

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