DINO: Divergent Node Cloning for Sustained Redundancy in HPC

Abstract

A plethora of resilience techniques have been investigated ranging from checkpoint/restart over redundancy to algorithm-based fault tolerance. Each technique works well for a different subset of application kernels, and depending on the kernel, has different overheads, resource requirements, and fault masking capabilities. If, however, such techniques are combined and they interact across kernels, new vulnerability windows are created.This work contributes the idea of end-to-end resilience by protecting windows of vulnerability between kernels guarded by different resilience techniques. It introduces the live vulnerability factor (LVF), a new metric that quantifies any lack of end-to-end protection for a given data structure. The work further promotes end-to-end application protection across kernels via a pragma-based specification, implemented as an extension to OpenMP, for diverse resilience schemes with minimal programming effort. This lifts the data protection burden from application programmers allowing them to focus solely on algorithms and performance while resilience is specified and subsequently embedded into the code through the compiler/library and supported by the runtime system. Two case studies demonstrate that end-to-end resilience meshes well with different execution paradigms and assess its overhead and effectiveness for different codes. In experiments, end-to-end resilience has an overhead over kernel-specific resilience of 1% on average.