This was a research project for a graduate level course at Rice University: COMP 580, probabilistic data structures and algorithms.

The experiment code is written in bash and can be found in the ./experiments directory. The code to generate the plots can be found in ./plots. The experiment scripts automatically call these scripts to generate the plots.

The actual bloom filters are implemented in C and can be found in the ./code/src directory.

When compiled, the main, standard bloom filter, and hierarchical bloom filter are compiled seperately and linked to gather afterwords. Optimization is disable as well, so this should minimize any performance increases due to different compilations.

Simply run make run_fast for a quick execution of all the tests. These are not the settings used in the final report, but are used to validate the functionaltiy of the scripts. To generate data used in the reportm run make run_full. Both of these commands will generate .dat files in ./experiments/dats and plots in ./plots. Note, only one of each plot will be preserved, while the .dats will be preserved until deleted.

The PDF of the writeup as well sa the source tex files for can be found in the ./writeup directory.