SAT Competition 2018
Mandatory participation requirements for SAT solvers
- The source code of submitted SAT solvers must be made available (licensed for research purposes) except for the solvers participating only in the No-Limits Track. The open source policy is strict: any submission that contains binary code will be disqualified.
- A 1–2 page system description document (see below for details).
- The co-authors of a solver in the solver description must match the co-authors listed in the submission system.
- SAT solvers must conform to DIMACS input/output requirements (see SAT Competition 2009 for details). Printing models in case of a satisfiable instance is required for all tracks expect for "no-limits". Additionally, UNSAT certificates (proofs) are required for the main tracks.
- SAT solvers that are planning to participate in the Main track must be able to produce certificates for UNSAT instances in tracecheck or DRAT (Delete Resolution Asymmetric Tautologies) format which is backwards compatible with the DRUP format of SAT Competition 2013.
- Each Main Track participant (team) is required to submit 20 new benchmark instances (not seen in previous competitions). At least 10 of those benchmarks should be "interesting": not too easy (solvable by MiniSat in a minute) or too hard (unsolvable by the participants own solver within one hour on a computer similar to the nodes of the StarExec cluster). See the benchmarks page for more information.
System Description Document
Each entrant to the 2018 SAT Competition must include a short (at least 1, at most 2 pages) description of the system. This should include a list of all authors of the system and their present institutional affiliations. It should also provide details of any non-standard algorithmic techniques (e.g., heuristics, simplification/learning techniques, etc.) and data structures, as well as references to relevant literature (be they by the authors themselves or by others). The system description must be formatted in IEEE Proceedings style, and submitted as PDF. The system descriptions will be posted on the 2018 SAT Competition website. Furthermore, provided the quality and number of system descriptions is high enough, the organizers are considering publishing the collection of system description as a report under the report series of Department of Computer Science, University of Helsinki (with ISSN and ISBN numbers).
In each track, solvers are ranked using a PAR-2 score based on a 5000 seconds timeout. Regarding the main track: In case a proof of unsatisfiability cannot be validated within 20000 seconds, the benchmark is considered not solved.
Participation of Portfolios
By a portfolio SAT solver we mean a combination of two or more (core) SAT solvers developed by different groups of authors. A portfolio SAT solver may participate only in the "No-Limits" track of the competition.
There will be no qualification round, but only a testing round to ensure that solvers are running properly on the evaluation system. The test set of instances will consist in a random selection of instances from the last SAT Competitions. The organizers reserve the right to disqualify poorly performing solvers.
A SAT solver will be disqualified if the solver produces a wrong answer. Specifically, if a solver reports UNSAT on an instance that was proven to be SAT by some other solver, or it reports SAT and provides a wrong certificate. A solver disqualified from the competition is not eligible to win any award. Disqualified solvers will be marked as such on the competition results page.
Note that there is a dedicated period when the participants can check their results to ensure that no problems are caused by the competition framework.
A solver can be withdrawn from SAT Competition 2018 only before the deadline for the submission of the final versions. After this deadline no further changes or withdrawals of the solvers are possible.
Participation of the organizers
The organizers of SAT Competition 2018 are not allowed to compete.
Number of submissions
Each participant is restricted to be a co-author of at most four different sequential solvers, two different parallel solvers, and one Glucose hack-track submission. Two solvers are different as soon as their sources differ or the compilation options are different. Solvers are also different if they use different command line options. However, we make an exception for a command line option that enables emitting UNSAT proofs for the certified UNSAT tracks. Authors can select per solver whether to participate in a track or not.