Obtaining games

• Use generate.py to generate the random games (needs "rngame" and "stgame" in tools/)
• Use download.sh to download and extract the benchmarks by Keiren

Compiling Oink

• sources: https://www.github.com/trolando/oink
• build with CMake
• copy oink, rngame, stgame to tools/

Compiling SPGSolver

• sources: https://www.github.com/umbertomarotta/SPGSolver/
• g++ CPPSolver.cpp -std=c++11 -lboost_program_options -lpthread -O3 -o spgsolver
• copy spgsolver to tools/spgsolver

Compiling parallel-si

• sources: https://www.github.com/jfearnley/parallel-si
• build with CMake
• copy parsi to tools/

Compiling pbespgsolver

• sources: https://svn.win.tue.nl/repos/MCRL2/ revision 15046
• build with CMake, disable all the extra tooling
• copy pbespgsolver to tools/

Compiling PGSolver

• sources: https://github.com/tcsprojects/pgsolver
• build with OPAM
• copy pgsolver to tools/

To run the experiments

• in terminal: ./exp.py run 1
• with slurm: sbatch slurm.sh

To obtain the results to csv

• just the finished experiments:./exp.py csv > results.csv
• with PAR2 score (penalize timeouts x2): ./exp.py csvpar2 > par2.csv

Remarks

• List of games odd cycles: oddcycles
• Log files used for the paper: logs.tar
• The csv files used for the paper: results.csv, par2.csv
• Random games used for the paper: random.tar.bz2 (split into 100M files)
• Use "Release" builds in CMake