Grant: Strategic Projects Supplemental Competition
Simon Fraser University
$71,800.00
Operating system scheduling for heterogeneous multicore systems
FY 2008-09
Summary
As space and energy consumption in computing systems become a crucial concern for systems ranging from small portable devices to data centers, computer architects search for processor technologies that accomplish more computation using less chip real estate and consuming less energy. An emerging technology that promises to deliver these benefits is heterogeneous multicore processors. It is well known that in order to
realize the benefits of HMC processors, running tasks must be dynamically assigned to cores by the operating system in a way that maximizes performance. Several scheduling algorithms addressing this objective were proposed, but despite offering good performance, they are not viable for use in commercial operating systems.
Schedulers in high-performance operating systems are designed, first and foremost, to achieve crucial qualitative objectives affecting system robustness, such as fair CPU sharing and consistent priority enforcement. Performance optimization is usually a secondary concern. In order to be commercially viable, new HMC schedulers must ensure that they improve performance without hurting these qualitative objectives.
This is challenging, because the core assignment strategy needed to maximize performance may not be the same as the core assignment strategy needed to enforce fairness. So the two parts of the system will interfere, hurting performance and fairness. Previous work has not evaluated the effect of this interference on system behaviour and did not consider how to address it. Until we find how to do this, however, we are unlikely to see the new performance-maximizing scheduling algorithms make it into commercial HMC systems. Our project addresses this problem. We will build and evaluate a HMC scheduling framework that balances among qualitative and quantitative performance objectives according to the needs of the user. The technology produced by our project will enable adoption of performance-maximizing algorithms in HMC systems, enabling users to benefit from energy and area saving properties of these systems.
Program
Strategic Projects Supplemental Competition
Recipient
Location
British Columbia, Canada
Recipient Type
Academia
Type
Grant
Source
NSERC's Awards Data
As published
Application id
356815-2007
Name nom
Fedorova, Alexandra
Department département
Computing Science
Organization id
5
Institution établissement
Simon Fraser University
Province en
British Columbia
Province fr
Colombie-Britannique
Country en
CANADA
Country fr
CANADA
Fiscal year exercice financier
2008
Competition year année de concours
2007
Award amount
71800
Program id
STPSC
Program nane en
Strategic Projects Supplemental Competition
Program name fr
Projets stratégiques compétition supplémentaire
Group en
Research partnerships programs
Group fr
Programmes de partenariat
Committee code
1064
Committee name en
Strategic Advanced Communications and Management of Information 1
Committee name fr
Stratégique -Technologies avancées de communication et de gestion de l'information 1
Area of application code
802
Area of application group en
Information and communication services
Area of application group fr
Services d'information et de télécommunications
Area of application en
Information systems and technology
Area of application fr
Systèmes d'information et technologie de l'information
Research subject code
2720
Research subject group en
Information technology
Research subject group fr
Technologie de l'information
Research subject en
Computer systems software
Research subject fr
Logiciel de systèmes informatiques
Partie
2
Nb partie
2
Application title
Operating system scheduling for heterogeneous multicore systems
Keyword
NA
Application summary
As space and energy consumption in computing systems become a crucial concern for systems ranging from small portable devices to data centers, computer architects search for processor technologies that accomplish more computation using less chip real estate and consuming less energy. An emerging technology that promises to deliver these benefits is heterogeneous multicore processors. It is well known that in order to
realize the benefits of HMC processors, running tasks must be dynamically assigned to cores by the operating system in a way that maximizes performance. Several scheduling algorithms addressing this objective were proposed, but despite offering good performance, they are not viable for use in commercial operating systems.
Schedulers in high-performance operating systems are designed, first and foremost, to achieve crucial qualitative objectives affecting system robustness, such as fair CPU sharing and consistent priority enforcement. Performance optimization is usually a secondary concern. In order to be commercially viable, new HMC schedulers must ensure that they improve performance without hurting these qualitative objectives.
This is challenging, because the core assignment strategy needed to maximize performance may not be the same as the core assignment strategy needed to enforce fairness. So the two parts of the system will interfere, hurting performance and fairness. Previous work has not evaluated the effect of this interference on system behaviour and did not consider how to address it. Until we find how to do this, however, we are unlikely to see the new performance-maximizing scheduling algorithms make it into commercial HMC systems. Our project addresses this problem. We will build and evaluate a HMC scheduling framework that balances among qualitative and quantitative performance objectives according to the needs of the user. The technology produced by our project will enable adoption of performance-maximizing algorithms in HMC systems, enabling users to benefit from energy and area saving properties of these systems.