SAT Competition 2002: Benchmark Submission Guidelines

Submission format

The submission can be a set of instances or a generator of instances. The set of instances should be representative of the problem to be solved at various scale. Generators should provide some parameters to scale the instances.

Generator format

The generator is a program to be launched on a Linux environment with some scaling parameters and a random seed if applicable. (for instance, a random 3-SAT generator will have two scaling parameters (nbvar,nbclauses) and a random seed parameter). It will output the instance on the standard output, using the file format below. The command line parameters will appear in the first commented lines of the instance.

File format

The benchmark file format will be in a simplified version of the DIMACS format:
c start with comments
p cnf 5 3
1 -5 4 0
-1 5 3 4 0
-3 -4 0


The benchmarks will be submitted in one of the following categories: Each instance should be submitted as SATISFIABLE, UNSATISFIABLE or UNKNOWN.


This category will include usual random 3-SAT, backbone controlled, etc. Only generator submissions are allowed here. Any generator must be able to produce many essentially different benchmarks for the same scaling parameters (given a different random seed).

Hand Made (crafted)

The benchmarks especially made to give a hard time to the solver. There will be an award for the smallest instance that cannot be solved by any solver.

Both instances or instance generators can be submitted. Here, no UNKNOWN instances. For both satisfiable and unsatisfiable instances, a proof must be submitted (e.g., a reference to a paper where the corresponding theorem is proved).


Here we should find instances from various applications, such as model checking, planning, encryption, etc.

These are series of instances, but NOT generators. The instances here must encode REAL problems.