Research
Overview
My research is in the general area of parallel and distributed computing. Current research topics center around power and performance in high-performance computing and include:
- Building power-aware run-time systems for future exascale machines,
- Developing models for power and performance for high-performance computing applications
- Developing power-aware scheduling techniques for future high-performance computing systems
- Developing models and techniques for inter-job optimization in high-performance computing systems
Current Projects
-
Exascale Computing Project (ECP). This project is investigating power constraints in task-based parallel programs.
[ECP Project web page] -
Zero-S Environment. This project is investigating the interaction of hardware features and HPC run-time systems and applications.
[Zero-S Project web page] -
Inter-Job Optimization in HPC Systems. This project is developing models and run-time techniques aimed at exascale systems.
[Inter-Job Optimization Project web page] -
Conductor. This project is developing models and run-time techniques aimed
at exascale systems.
[Conductor Project web page] -
Cloudster. This project is investigating techniques for cost-effective execution
of high-performance computing applications on the Cloud.
[Cloudster Project web page] -
Hobbes.
This project (joint with Sandia, UNM, and others) is developing OS support
for extreme-scale systems.
[Hobbes Project web page] -
Power at Scale.
This project (joint with LLNL) is developing models and schedulers to optimize
power at all levels for massively parallel machines.
[Power At Scale Project web page]