Artificial neural networks for online error detection

TL;DR

This paper introduces a low-cost, online error detection method using artificial neural networks trained with fault injection, effectively identifying silent data corruptions with minimal performance overhead and high fault detection accuracy.

Contribution

It demonstrates the effectiveness of ANNs for online error detection in hardware, outperforming existing methods like Topaz in detection coverage and efficiency.

Findings

ANNs detect 94.85% of faults with 6.45% CPU overhead

ANN-based detection outperforms Topaz in accuracy and performance

Overclocked ARM CPUs used for validation show practical applicability

Abstract

Hardware reliability is adversely affected by the downscaling of semiconductor devices and the scale-out of systems necessitated by modern applications. Apart from crashes, this unreliability often manifests as silent data corruptions (SDCs), affecting application output. Therefore, we need low-cost and low-human-effort solutions to reduce the incidence rate and the effects of SDCs on the quality of application outputs. We propose Artificial Neural Networks (ANNs) as an effective mechanism for online error detection. We train ANNs using software fault injection. We find that the average overhead of our approach, followed by a costly error correction by re-execution, is 6.45% in terms of CPU cycles. We also report that ANNs discover 94.85% of faults thereby resulting in minimal output quality degradation. To validate our approach we overclock ARM Cortex A53 CPUs, execute benchmarks on them and record the program outputs. ANNs prove to be an efficient error detection mechanism, better than a state of the art approximate error detection mechanism (Topaz), both in terms of performance (12.81% CPU overhead) and quality of application output (94.11% detection coverage).