Model Counting Competition Submission of Ganak + Approxmc 2025

Scripts and binary. The binary shows all SHA1 hashes so you can rebuild it to exactly the same behaviour.

SHA hashes:

c o Ganak SHA1: 61286bc663e22987dfff25b4c01b1b4cff5beff3
c o Arjun SHA1: 39ef145587c791ac101c8b3b503aa62b5b756d76
c o SBVA SHA1: 0faa08cf3cc26ed855831c9dc16a3489c9ae010f
c o CMS SHA1: 6c866e338b3221c6ed036273268037f96397b092
c o ApproxMC SHA1: fb5dfbe46293e956cec7e963d9b649d0437160c1
c o BreakID SHA1: dee9744b7041cec373aa0489128b06a40fce43a1

CadiBack:
c o Version 0.2.1 9c42e308406e76a5e8ea224d65e971f4208c9c36

CadiCal:
c o CaDiCaL 2.0.0 0e8be0642ebf3e872a9df806fc0c1c1d8fa90084

All of the above is available online, in public repositories, at the time, and before the time of the competition, under meelgroup or msoos organizations/users on GitHub. CMS refers to CryptoMiniSat.

Static binary compiled with gcc (GCC) 15.1.1 20250425

How to rebuild

Clone all above repositories in their respective directories (i.e. cadical, cadiback, sbva, breakid, approxmc, arjun, ganak), then make sure that a build subdirectory exists for all except cadical and cadiback. Then go to the build subdirectory for each, and execute: ln -s ../scripts/*.sh . or ln -s ../scripts/build_scripts/*.sh . to get the build scripts into the build subdirectory for each. Then go into ganak/build and call ./rebuild_static_all_release.sh.

I have the GMP library recompiled to build both C++ and static libraries:

#!/bin/bash
make clean
CC=gcc-14 CXX=g++-14 ./configure --enable-shared --enable-static --enable-cxx
make clean
CC=gcc-14 CXX=g++-14 make -j14
sudo CC=gcc-14 CXX=g++-14 make -j14 install

And I have mpfr compiled for static libraries, and location to GMP:

#!/bin/bash
make clean
CC=gcc-14 CXX=g++-14 ./configure --enable-shared --enable-static --with-gmp=/usr/local
make clean
CC=gcc-14 CXX=g++-14 make -j14
sudo CC=gcc-14 CXX=g++-14 make -j14 install

Thanks

Many thanks are printed while running the binary. Authors' names, papers, etc. Each repository also has its own set of thanks and authors and papers. Science is a mountain, and this tool is built on a mountain of work and papers and books etc. It's a huge mountain, a beautiful montain, the best mountain.