Reinforcement Learning-based Adaptive Mitigation of Uncorrected DRAM Errors in the Field

TL;DR

This paper introduces a reinforcement learning-based adaptive approach to mitigate uncorrected DRAM errors, significantly reducing lost compute time in supercomputers by predicting and acting on error likelihoods.

Contribution

It presents the first adaptive, RL-based method for uncorrected DRAM error mitigation, optimizing mitigation timing based on error prediction and cost-benefit analysis.

Findings

Reduces lost compute time by 54% on supercomputer logs

Achieves performance within 6% of an optimal Oracle method

Uses only two user-defined parameters for mitigation decision-making

Abstract

Scaling to larger systems, with current levels of reliability, requires cost-effective methods to mitigate hardware failures. One of the main causes of hardware failure is an uncorrected error in memory, which terminates the current job and wastes all computation since the last checkpoint. This paper presents the first adaptive method for triggering uncorrected error mitigation. It uses a prediction approach that considers the likelihood of an uncorrected error and its current potential cost. The method is based on reinforcement learning, and the only user-defined parameters are the mitigation cost and whether the job can be restarted from a mitigation point. We evaluate our method using classical machine learning metrics together with a cost-benefit analysis, which compares the cost of mitigation actions with the benefits from mitigating some of the errors. On two years of production logs from the MareNostrum supercomputer, our method reduces lost compute time by 54% compared with no mitigation and is just 6% below the optimal Oracle method. All source code is open source.