This is a SAT solver based on MiniSAT, which currently implements basic CDCL solving.
It currently exposes 1 binaries, solve_dimacs which take as input a list
of DIMACS files. solve_dimacs uses
CDCL solving with the data-structures and algorithms used in MiniSAT in order to more efficiently
find a so find a solution.
In order to properly reproduce the results there are a couple of necessary dependencies:
Rust with version
rustc 1.42.0-nightly (859764425 2020-01-07). The dependencies will be installed at compilation
time.
In order to compile & run the executable on any CNF files:
$ cargo run --release <CNF files>The benchmarks used for the results in the paper are in $PROJECT_DIR/src/bin/data/bmc/.
In order to produce the output files(requires Ruby), run
$ cd $PROJECT_DIR/src/bin
$ ./test_sound.rbThis may take some time, and I also manually edited solve_dimacs.rs to the number of desired
cores.
In order to generate the graphs, modify analyze.py in $PROJECT_DIR/src/bin/ to use the
specified metric, and direct it to the set of output files from test_sound.rb or other CSV
files output by the solver.
In order to compile the paper, run the following commands:
pdflatex paper
bibtex paper
pdflatex paper
pdflatex paperThe source code is available in the $PROJECT_DIR/src