This repository provides a Bokeh scripty that I used to visualize millions of allocation statistics generated by FOM-tools (link to FOM-tools https://github.com/FOM-Tools/FOM-Tools). This script is just an example for how powerful Bokeh is in order to explore interactively such datasets.
Allocation statistics were generated by FOM tools and the corresponding lifetime was computed in a postprocessing step. Two example output files can be found in the directory files . Each point in the Bokeh plot presents an allocated object and how long it lived. X-position indicates when the object was created after program start, y-position indicates the lifetime of the object (time from allocation to deallocation). The script uses Bokeh's tooltip to interactively show the corresponding stacktrace for each allocations.
Screenshot:
Bokeh has been extremally useful in order to find many interesting object lifetime patterns in the vast of amount of allocation data. See also http://iopscience.iop.org/article/10.1088/1742-6596/898/7/072031/pdf for more details.
In order to run this script: install bokeh and then execute "bokeh serve plot1.py", then open the URL http://localhost:5006/plot1