ATune: Compiler-Driven Adaptive Execution
ATune Objectives:
The goal of this project is to study
dynamically adapting compiler-translated MPI applications in a distributed
system.
In previous work, we have created a system
for dynamic adaptation and tuning of computational applications (PEAK).
The current study includes, deploying a
translator that converts OpenMP parallel applications directly into MPI message
passing programs,
and investigating new methods to enable
dynamic, adaptive optimization and tuning in diverse architectures.
To facilitate matching different
application needs with diverse available resources, our Internet-sharing system
(iShare) is exploited.
ATune Block diagram:
Related Projects:
PEAK – Program Evolution by
Adaptive Compilation
iShare – Internet-sharing
middleware and collaboration technology
Publications:
Seyong Lee and Rudolf
Eigenmann. Adaptive Runtime Tuning of Parallel Sparse
Matrix-Vector Multiplication on Distributed Memory Systems. 22nd ACM
International Conference on Supercomputing (ICS), June 2008. (pdf)
Seyong Lee and Rudolf Eigenmann.
Adaptive Tuning in a Dynamically Changing Resource Environment. Workshop on National Science Foundation Next
Generation Software Program (NSFNGS) held in conjunction with the IEEE
International Parallel & Distributed Processing Symposium (IPDPS), April
2008. (pdf)
Seyong Lee, Xiaojuan Ren, and Rudolf Eigenmann. Efficient Content Search in
iShare, a P2P based Internet-Sharing System. The 2nd Workshop on Large-scale, volatile Desktop Grids
(PCGrid) held in conjunction with the IEEE International Parallel &
Distributed Processing Symposium (IPDPS), April 2008. (pdf)
Xiaojuan Ren, Seyong Lee, Rudolf
Eigenmann, Saurabh Bagchi. Prediction of Resource
Availability in Fine-Grained Cycle Sharing Systems and Empirical Evaluation. Journal
of Grid Computing, Volume 5, Number 2, pp. 173-195, June 2007.(pdf)
Xiaojuan Ren, Seyong Lee, Rudolf
Eigenmann, Saurabh Bagchi. Resource Availablilty Prediction
in Fine-Grained Cycle Sharing Systems. The 15th IEEE
International Symposium on High Performance Distributed Computing (Nominated
for Best Paper Award), June 2006. (pdf)
Xiaojuan Ren, Seyong Lee, Saurabh Bagchi, Rudolf Eigenmann. Resource Fault Prediction in
Fine-Grained Cycle Sharing. DSN-2005: The International Conference on
Dependable Systems and Networks, Fast Abstracts, June 2005
Funding:
This work is supported in part by the National Science Foundation. Any opinions,
findings, and conclusions or recommendations expressed in this material are
those of the authors and do not necessary reflect the views of the National
Science Foundation.