Hobbes: An Operating System for Extreme-Scale Systems
The goal of this three-year project is to deliver an operating system for future extreme-scale parallel computing platforms that will address the major technical challenges of energy efficiency, managing massive parallelism and deep memory hierarchies, and providing resilience in the presence of increasing failures. Full details on Hobbes can be found here. Hobbes is a collaboration between 4 national labs and 8 universities. The UA part of the project is develop models for power as well as ones that integrate power and resilience.
People
- Faculty:
- David Lowenthal
- Graduate Students:
- Peter Bailey
Publications
Peter Bailey, David K. Lowenthal, Vignesh Ravi, Bronis de Supinski,
Barry Rountree, and Martin Schulz.
Adaptive Configuration Selection for Power-Constrained
Heterogeneous Systems
43rd IEEE International Conference on Parallel Processing
(ICPP), September 2014.
Paper: PDF
This material is based upon work supported by Sandia National Laboratory and DOE under subcontract number 1399016. Any opinions, findings, and conclusions or recommendations expressed in this material are those of the author(s) and do not necessarily reflect the views of Sandia National Laboratory or DOE.