MaxCliquePara general information The following assumes you are running a console in a directory where you have copied/uncompressed the source files. 1. Compiling to compile sources run: make to clean (remove the binary) run: make clean 2. Running run the binary executable: ./maxClique [number_of_threads] [job_queue_size] where graph_file_name : either relative or absolute path to the file containing an undirected graph in DIMACS file format (http://prolland.free.fr/works/research/dsat/dimacs.html) number_of_threads (optional) : the number of threads to run the algorithm on. (0 - execute serially using serial code; 1 .. execute on 1 thread using parallel code; n .. execute on n threads using parallel code). If the argument is not specified, 0 is selected as default job_queue_size (optional) : size of the job queue. This argument is a remnant of development cycle in which a more complex code was used. Currently the only reasonable numbers are 0 and 1. 0 is the default and causes the algorithm to split jobs only when a thread is waiting (there is no queue of jobs erady for processing). 1 enables the queue of length 1, which means there is always a job on queue for a thread to take, when it becomes idle. This is the faster mode of operation and the one used in the article "Exact Parallel Maximum Clique Algorithm for General and Protein Graphs" The results are reported in text form on the standard output, listed in the following form: Loading DIMACS graph <|V|> vertices <|E|> edges -- pMC[ threads, jobs](int,bitstring based set,Graph,MCR sort for Bitstrings) search took ; steps Clique (<|Q|>): [Q1,Q2,...,Qq] Thread efficiency = where the variables filled by the algorithm are marked in <...>. Text of the third line starts with pMC if parallel code is used and MC if serial code is used. Time reported is for the clique search only (reading file from disk and transforming the graph into adjacency matrix are not counted)