Handling missing data to improve HPC performance portability scores
Handling Missing Data in Performance Portability Studies
Distributed, Parallel, and Cluster Computing
Summary
Missing information is common in datasets used to measure how well software runs across different computers. This problem makes it hard or impossible to get accurate scores on software performance portability. The authors study different ways to fill in these missing data points, showing how each method affects accuracy, reliability, and computing effort. Their work helps people understand which approaches work best for different situations in high-performance computing.
What this means in practice
- •For hpc performance analysts: Improve analysis of software performance portability by selecting appropriate methods to complete missing benchmarking data.
- •For software performance engineers: Make more reliable performance portability score calculations despite incomplete hardware benchmarking or implementation gaps.
Authors
Ami Marowka
Abstract
Missing data is a common challenge in many real-world datasets, often leading to biased results or reduced accuracy. Missing data is a pervasive problem in performance efficiency datasets, particularly in high-performance computing (HPC) performance portability studies. Performance portability scores rely on complete sets of performance efficiencies as inputs to their respective metrics; however, missing values can arise due to incomplete benchmarking, hardware constraints, or implementation gaps. Many established performance portability metrics lack mechanisms for handling missing inputs, which can result in incomplete analyses or the inability to compute scores altogether. In this study, we evaluate a range of imputation methods and algorithms for addressing missing performance efficiencies. We present illustrative examples for each approach and apply them to real-world HPC performance portability datasets. Our results highlight the trade-offs between accuracy, robustness, and computational cost, offering practical recommendations for researchers seeking to mitigate the impact of missing data in performance portability assessments.