刘炯鑫:
> Dear dickrp,
> My research topic is "Techniques of detecting and adapting to
> system fault with multiprocessors". After reading the paper "Reliability
> Multiprocessor system-on-chip synthesis", I think the run-time techniques
> to detect and adapt to processor failure is very critical in the
> calculating of system MTTF. Once a component fault happens, the
> rescheduling of
> tasks running on the failed component (processor) is necessary. However,
> the change of task assignments will in turn affect the MTTF of other
> processing components which is associated with thermal profile. In order to
> prolong the MTTF, at the start of system design, resources and floorplan
> should be considered carefully. In addition, a clear understanding of the
> fault lattice and a correct way to calculate system MTTF are helpful to
> find the optimal rescheduling strategy. So, how to base on these factors to
> come up with a wise technique is a difficult but interesting task.
> Best wishes

Ran Ren and Zhong Li are also working on this topic. You should talk with
them to see if you can work on differnt parts of the problem.

-Robert Dick-
Received on Sun Aug 26 13:17:09 2007