Clustering simulation failures for triage and debugging

Abstract

Methods, systems and computer program products for clustering simulation failures are provided. Aspects include receiving simulation data comprising a plurality of simulation failure files, generating a token for each simulation failure file of the plurality of simulation failure files, determining a token score for each token for each simulation failure file of the plurality simulation failure files, normalizing each token score for each token in the plurality of simulation failure files utilizing a weighting scheme to create a normalized token score for each token, determining a set of groups for the plurality of simulation failure files, and assigning one or more simulation failure files from the plurality of simulation failure files into a group in the set of groups based at least in part on normalized token score.