Grant: Collaborative Research and Development Grants
University of Alberta
$40,000.00
Runtime Binary Re-Compilation for Enterprise-Scale Computing
EF 2015-16
Résumé
In the near future a larger set of metrics to monitor program behaviour will be much more easily accessible are expected to become available in some computer architectures. This development leads to the possibility of using such metrics at runtime to change the code being executed to improve performance. Some of the developments that will be necessary to realize this vision include (1) a control mechanism to re-focus the hardware runtime instrumentation framework; (2) a cost-effective data representation and reduction framework that will have to resolve trade-offs between the amount of information gathered and the computational cost of the gathering and analysis; and (3) a set of new code transformations based on path specialization.
On the other hand, the hardware of computer systems rapidly evolves with new features offered with each processor generation. Also, new and more aggressive compilation strategies are developed by compiler researchers and developers. A long-standing problem is that many of the new hardware features can only be accessed if the compiler is aware of them at the time that it generates executable binaries. Moreover, none of the improved compilation strategies can be deployed unless the application is compiled with the newer compiler. However, many important software applications run the same binaries for many years --- or more likely are never re-compiled --- and therefore they do not access the newer features that the hardware makes available and do not benefit from more advanced compilation strategies. The goal of this proposal is to develop a run-time binary recompilation framework to solve this problem by allowing the newly generated binaries to access the most recent hardware features and compiler techniques.
Programme
Collaborative Research and Development Grants
Bénéficiaire
Emplacement
Alberta, Canada
Type de bénéficiaire
Academia
Type
Grant
Source
NSERC's Awards Data
Tel que publié
Application id
469056-2014
Name nom
Amaral, Jose
Department département
Computing Science
Organization id
9
Institution établissement
University of Alberta
Province en
Alberta
Province fr
Alberta
Country en
CANADA
Country fr
CANADA
Fiscal year exercice financier
2015
Competition year année de concours
2014
Award amount
40000
Program id
CRDPJ
Program name en
Collaborative Research and Development Grants
Program name fr
Subventions de recherche et développement coopérative
Group en
Research partnerships programs
Group fr
Programmes de partenariat
Committee code
96
Committee name en
RPP Internal Decision Cttee
Committee name fr
Cté de décision interne - PPR
Area of application code
801
Area of application group en
Information and communication services
Area of application group fr
Services d'information et de télécommunications
Area of application en
Computer software
Area of application fr
Logiciel
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
4
Application title
Runtime Binary Re-Compilation for Enterprise-Scale Computing
Keyword
NSERC
Application summary
In the near future a larger set of metrics to monitor program behaviour will be much more easily accessible are expected to become available in some computer architectures. This development leads to the possibility of using such metrics at runtime to change the code being executed to improve performance. Some of the developments that will be necessary to realize this vision include (1) a control mechanism to re-focus the hardware runtime instrumentation framework; (2) a cost-effective data representation and reduction framework that will have to resolve trade-offs between the amount of information gathered and the computational cost of the gathering and analysis; and (3) a set of new code transformations based on path specialization.
On the other hand, the hardware of computer systems rapidly evolves with new features offered with each processor generation. Also, new and more aggressive compilation strategies are developed by compiler researchers and developers. A long-standing problem is that many of the new hardware features can only be accessed if the compiler is aware of them at the time that it generates executable binaries. Moreover, none of the improved compilation strategies can be deployed unless the application is compiled with the newer compiler. However, many important software applications run the same binaries for many years --- or more likely are never re-compiled --- and therefore they do not access the newer features that the hardware makes available and do not benefit from more advanced compilation strategies. The goal of this proposal is to develop a run-time binary recompilation framework to solve this problem by allowing the newly generated binaries to access the most recent hardware features and compiler techniques.