Fig 4. Summarization memory savings.

Of all methods discussed in this paper, MzTree is the only one that can handle data sets that require more RAM than the application provides. MzTree’s summarization feature (shown here with limit set to 5,000 points) allows users to work with arbitrarily large data sets even in situations where available RAM is less than that required to hold the unsummarized set of returned points, such as system memory limits or visualization rendering limits.