Parallelization of transition counting for process mining on multi-core CPUs and GPUs

Abstract

Many process mining tools and techniques produce output models based on the counting of transitions between tasks or users in an event log. Although this counting can be performed in a forward pass through the event log, when analyzing large event logs according to different perspectives it may become impractical or time-consuming to perform multiple such passes. In this work, we show how transition counting can be parallelized by taking advantage of CPU multi-threading and GPU-accelerated computing. We describe the parallelization strategies, together with a set of experiments to illustrate the performance gains that can be expected with such parallelizations.