A system framework and API for run-time adaptable parallel software

Authors:
Mikyung Kang;Dong-In Kang;Seungwon Lee;Jaedon Lee
Affiliations:
University of Southern California, Arlington, VA;University of Southern California, Arlington, VA;Samsung Advanced Institute of Tech. Samsung Electronics CO., LTD., Nongseo-dong, Giheung-gu, Yongin-si Gyeonggi-do, Korea;Samsung Advanced Institute of Tech. Samsung Electronics CO., LTD., Nongseo-dong, Giheung-gu, Yongin-si Gyeonggi-do, Korea
Venue:
Proceedings of the 2013 Research in Adaptive and Convergent Systems
Year:
2013

Citing 12
Cited 0

The SPLASH-2 programs: characterization and methodological considerations

ISCA '95 Proceedings of the 22nd annual international symposium on Computer architecture
Co-array Fortran for parallel programming

ACM SIGPLAN Fortran Forum
The Vision of Autonomic Computing

Computer
A Malleable-Job System for Timeshared Parallel Machines

CCGRID '02 Proceedings of the 2nd IEEE/ACM International Symposium on Cluster Computing and the Grid
X10: an object-oriented approach to non-uniform cluster computing

OOPSLA '05 Proceedings of the 20th annual ACM SIGPLAN conference on Object-oriented programming, systems, languages, and applications
Advances, Applications and Performance of the Global Arrays Shared Memory Programming Toolkit

International Journal of High Performance Computing Applications
An Approach To Data Distributions in Chapel

International Journal of High Performance Computing Applications
Scioto: A Framework for Global-View Task Parallelism

ICPP '08 Proceedings of the 2008 37th International Conference on Parallel Processing
Self-adaptive software: Landscape and research challenges

ACM Transactions on Autonomous and Adaptive Systems (TAAS)
PetaBricks

XRDS: Crossroads, The ACM Magazine for Students - The Changing Face of Programming
EasyPDP: An Efficient Parallel Dynamic Programming Runtime System for Computational Biology

IEEE Transactions on Parallel and Distributed Systems
High-performance parallel accelerator for flexible and efficient run-time monitoring

DSN '12 Proceedings of the 2012 42nd Annual IEEE/IFIP International Conference on Dependable Systems and Networks (DSN)

Quantified Score

Hi-index	0.00

Visualization

Abstract

A system framework supporting run-time dynamic adaptation of parallel applications is proposed. The framework consists of a run-time manager and an API. The run-time manager profiles applications and guides adaption of applications at run-time for optimization and adaptation to the changes of environment such as the number of cores to use. With the proposed API, a range of different parallelization can be provided inside of an application, and the run-time manager chooses an appropriate parallelization at run-time. With the aid of the proposed run-time manager, a program can adapt/morph itself to dynamic changes in the system resources and/or performance requirements at run-time. The proposed adaptive run-time manager monitors and guides the applications to make the system adaptable at run-time.