Requires Python3, and the following dependencies, installed via pip:
pip3 install matplotlib pandas seaborn
Run make to produce all the figures (produced as PDFs).
The plotting scripts read from all the .csv files in this directory,
that need to manually be moved in here from the benchmark
outputs:
./data.csvand./prep_time.csvcontain outputs from../benchmarks/../microbench-mem-accesscontains output from../microbenchmark-compare-read-arr/. You will likely need to tweak themachinesmap in./microbench-mem-access.pyfor it to pick up the data from your machine. Currently it is set to the machine names that were used for evaluation in the paper.
The raw data files used to produce the graphs in the paper are already included in this directory.