ParaFROST stands for Parallel Formal Reasoning about SaTisfiability. It is a parallel SAT solver with GPU-accelerated inprocessing capable of harnessing NVIDIA CUDA-enabled GPUs in applying modern inprocessing tecnhiques in parallel. The CDCL search is built from scratch with various optimisations based on CaDiCaL heuristics (see our paper in FMSD'23). The inprocessing engine extends our previous work in SIGmA simplifier with new data structures, parallel garbage collection and more.

To install either the CPU or the GPU solvers, use the install.sh script which has the following usage:

  usage: install.sh [ <option> ... ]
  where <option> is one of the following

   -h or --help          print this usage summary
   -n or --less          print less verbose messages
   -q or --quiet         be quiet (make steps still be saved in the log)
   -c or --cpu           install CPU solver
   -g or --gpu           install GPU solver (if CUDA exists)
   -w or --wall          compile with '-Wall' flag
   -d or --debug         compile with debugging information
   -t or --assert        enable only code assertions
   -p or --pedantic      compile with '-pedantic' flag
   -l or --logging       enable logging (needed for verbosity level > 2)
   -s or --statistics    enable costly statistics (may impact runtime)
   -a or --all           enable all above flags except 'assert'
   --ncolors             disable colors in all solver outputs
   --clean=<target>      remove old installation of <cpu | gpu | all> solvers
   --standard=<n>        compile with <11 | 14 | 17> c++ standard
   --gextra="flags"      pass extra "flags" to the GPU compiler (nvcc)
   --cextra="flags"      pass extra "flags" to the CPU compiler (g++)

To build the GPU solver, make sure you have a CUDA-capable GPU with pre-installed NVIDIA driver and CUDA toolkit.

For installing CUDA v12.6, run the following commands on a Ubuntu 24.04:

wget https://developer.download.nvidia.com/compute/cuda/repos/ubuntu2404/x86_64/cuda-keyring_1.1-1_all.deb
sudo dpkg -i cuda-keyring_1.1-1_all.deb
sudo apt-get update
sudo apt-get -y install cuda-toolkit-12-6

The source code is also platform-compatible with Windows and WSL2. To install CUDA on those platforms, follow the installation guide in https://docs.nvidia.com/cuda/.

Now the GPU solver is ready to install by running the install script via the command ./install.sh -g. The parafrost binary, the library libparafrost.a, and the main include file solver.hpp will be created by default in the build directory.

To build a CPU-only version of the solver, run ./install.sh -c.

Add -t argument with the install command to enable assertions or -d to collect debugging information for both the CPU and GPU solvers.

The solver can be used via the command parafrost [<option> ...][<infile>.<cnf>][<option> ...].
For more options, type parafrost -h or parafrost --helpmore.

ParaFROST supports incremental solving to add/remove variables or clauses incrementally while solving with assumptions if needed. A fully configurable interface to integrate ParaFROST with CBMC model checker is created here (https://github.com/muhos/gpu4bmc). A similar interface can be created to work with ParaFROST in any SAT-based bounded model checker.

Please cite our latest paper FMSD'23 when using ParaFROST.