A Case Study of Application Structure Aware Resilience Through Differentiated State Saving and Recovery

摘要

Resilience is a growing concern for large-scale simulations. As failures become more frequent, alternatives to global checkpointing that limit the extent of needed recovery become more desirable. Additionally, platforms will differ in both error rates and types, therefore, a flexible and customizable recovery strategy will be extremely helpful to the applications running on these platforms. Applications often have structures that provide logical confinement spaces that can be exploited for this purpose. We investigate a customizable recovery strategy using Chombo, a structured adaptive mesh refinement (SAMR) library, as a case study. We exploit the inherent granularities and hierarchy in SAMR to limit the impact of faults for localized recovery, and identify tunable parameters for customizing the strategy depending upon the application and platform behavior. We use Global View Resilience (GVR) library, which provides global versioning arrays for application-controlled state saving as our resiliency interface.