國立高雄大學圖資館 |

Language: English

Back

Parallel Scheduling with Batching.

Georgetown University.

Parallel Scheduling with Batching.

Record Type:	Electronic resources : Monograph/item
Title/Author:	Parallel Scheduling with Batching.
Author:	Sheridan, Brendan.
Published:	Ann Arbor : ProQuest Dissertations & Theses, 2018
Description:	141 p.
Notes:	Source: Dissertation Abstracts International, Volume: 79-08(E), Section: B.
Notes:	Adviser: Jeremy Fineman.
Contained By:	Dissertation Abstracts International79-08B(E).
Subject:	Computer science.
Online resource:	http://pqdd.sinica.edu.tw/twdaoapp/servlet/advanced?query=10748372
ISBN:	9780355807905

Parallel Scheduling with Batching.
Sheridan, Brendan.

Parallel Scheduling with Batching. - Ann Arbor : ProQuest Dissertations & Theses, 2018 - 141 p.

Source: Dissertation Abstracts International, Volume: 79-08(E), Section: B.

Thesis (Ph.D.)--Georgetown University, 2018.

This work provides and analyses three provably good parallel scheduling algorithms. Each algorithm utilizes batching by delaying work until it can be executed as part of a larger batch. Batching is often necessary for good performance in problems that have a high start-up cost to do any work or reduced cost for a large group of work. These problems are often challenging because the delay of work conflicts with standard scheduling constraints such as deadlines as well as useful scheduling objectives such as makespan and flow.

ISBN: 9780355807905Subjects--Topical Terms:

199325
Computer science.

Parallel Scheduling with Batching.
LDR:02258nmm a2200301 4500 001 547545
005 20190513114555.5
008 190715s2018 ||||||||||||||||| ||eng d
020 $a 9780355807905
035 $a (MiAaPQ)AAI10748372
035 $a (MiAaPQ)georgetown:13869
035 $a AAI10748372
040 $a MiAaPQ $c MiAaPQ
100 1 $a Sheridan, Brendan. $3 826844
245 1 0 $a Parallel Scheduling with Batching.
260 1 $a Ann Arbor : $b ProQuest Dissertations & Theses, $c 2018
300 $a 141 p.
500 $a Source: Dissertation Abstracts International, Volume: 79-08(E), Section: B.
500 $a Adviser: Jeremy Fineman.
502 $a Thesis (Ph.D.)--Georgetown University, 2018.
520 $a This work provides and analyses three provably good parallel scheduling algorithms. Each algorithm utilizes batching by delaying work until it can be executed as part of a larger batch. Batching is often necessary for good performance in problems that have a high start-up cost to do any work or reduced cost for a large group of work. These problems are often challenging because the delay of work conflicts with standard scheduling constraints such as deadlines as well as useful scheduling objectives such as makespan and flow.
520 $a Specifically, this dissertation focuses on two scheduling problems: dynamically multithreaded computations with implicit batching and Integrated Stockpile Evaluation (ISE). The former is an online scheduling problem with precedence constraints where the delay of data structure operations can be used to reduce total work and increase parallelism with batched operations. ISE is a traditional offline scheduling problem where n jobs, each with an arbitrary release time, must be scheduled non-preemptively on m machines. It has the extra constraint that machines may only be used if they have been recently calibrated and batching can be used to reduce the necessary number of calibrations. In both cases, competitive scheduling algorithms must appropriately balance the benefit of batching against the cost of delaying work.
590 $a School code: 0076.
650 4 $a Computer science. $3 199325
690 $a 0984
710 2 $a Georgetown University. $b Computer Science. $3 826845
773 0 $t Dissertation Abstracts International $g 79-08B(E).
790 $a 0076
791 $a Ph.D.
792 $a 2018
793 $a English
856 4 0 $u http://pqdd.sinica.edu.tw/twdaoapp/servlet/advanced?query=10748372