Cost-Aware Prediction of Uncorrected DRAM Errors in the Field
Machine Learning, Deep Learning and Artificial Intelligence
Requirements, Performance, and Benchmarks
Reliability and Resiliency
TimeWednesday, 18 November 20201pm - 1:30pm EST
DescriptionThis paper presents and evaluates a method to predict DRAM uncorrected errors, a leading cause of hardware failures in large-scale HPC clusters. The method uses a random forest classifier, which was trained and evaluated using error logs from two years of production of the MareNostrum 3 supercomputer. By enabling the system to take measures to mitigate node failures, our method reduces lost compute time by up to 57%, a net saving of 21,000 node hours per year. We release all source code as open source.
We also discuss and clarify aspects of methodology that are essential for a DRAM prediction method to be useful in practice. We explain why standard evaluation metrics, such as precision and recall, are insufficient, and base the evaluation on a cost–benefit analysis. This methodology can help ensure that any DRAM error predictor is clear from training bias and has a clear cost–benefit calculation.