… impact of soft errors on ve iterative methods, which are ubiquitous in scienti c applications. A
… We remark that rolling back an application because of an incorrect labeling of an iteration (…

… For our target applications, soft error detection can be efficiently performed by monitoring
the convergence variable only, since an abnormal change in the convergence variable can be …

… Further, we evaluate a variety of soft error detection and tolerance techniques, including …
the impact of soft errors on a model application that uses iterative methods. The model …

… applications, namely those involving the iterative … an error-detecting version of a parallel
algorithm for iteratively solving the Laplace equation over a rectangular grid. This error-detecting …

… iterative detection in the context of these applications. As a result of this (sometimes tiring)
experience, we developed a different perspective on iterative detection … iterative detection as …

… We will focus on this model, although other applications can also fit within the formalism
discussed here. One can view the configuration of Figure 2 as a serially concatenated code, in …

… When an error occurs during an application execution, if the affected process stops working,
… error detection every d iterations and perform periodical checkpoint every cd iterations. Both …

… even applicable for certain applications. For convergent iterative applications, we observe
… of SDC, which is specific to an application but independent of the input dataset size. Based …

… In addition to enabling a concrete understanding of soft error behavior for an important
application class, this analysis has the potential to aid the design of soft error detectors and …

… to the time of their application with respect to the normal system operation. … error detection
are that they can detect transient error and recovery can be initiated before fault-caused error …