Scheduling and Parallel Computing

Lead Research Organisation: University of Manchester
Department Name: Mathematics

Abstract

Consider a large computational task, for instance, solving a large system of linear equations. This task can be split into many smaller jobs which are then scheduled and queued at a large number of different heterogeneous computing resources and are executed in parallel. The processing requirements of the different resources (for instance CPUs and GPUs) are different, and they may well have different communication costs. In this project we will be interested in understanding the stochastic effects before designing (and implementing) novel distributed scheduling algorithms.

Publications

10 25 50

Studentship Projects

Project Reference Relationship Related To Start End Student Name
EP/N509565/1 01/10/2016 30/09/2021
1925936 Studentship EP/N509565/1 18/09/2017 31/03/2021 Thomas McSweeney
 
Title Static task scheduling simulator for accelerated platforms 
Description This is a set of Python modules and scripts that allow users to simulate the scheduling of arbitrary static task dependency graphs on user-defined accelerated computing platforms. 
Type Of Technology Software 
Year Produced 2019 
Open Source License? Yes  
Impact None so far. 
URL https://github.com/mcsweeney90/heterogeneous_optimistic_finish_time
 
Description Talk at Numerical Linear Algebra group meeting 
Form Of Engagement Activity A talk or presentation
Part Of Official Scheme? No
Geographic Reach Local
Primary Audience Postgraduate students
Results and Impact I gave a short talk at the University of Manchester Numerical Linear Algebra (NLA) group meeting. The content closely followed a conference paper which I have recently submitted and the purpose was to make the members of the group aware of my current research.
Year(s) Of Engagement Activity 2019
URL https://nla-group.org/meetings/