Qualification

There will be two qualification rounds for registered solvers, where successful participation in at least one of them is mandatory to be accredited for SAT-Race.

In each round, we check the following requirements:

Requirements

  • Conformance to I/O specification. Solvers have to respect the I/O requirements (CNF, AIG). In particular, for CNF-based solvers, the output must contain a solution line (starting with "s ") and—in case of a satisfiable instance—at least one values line (starting with "v "). Exit codes also must be set correctly (10 for satisfiable, 20 for unsatisfiable, 0 for unknown).
  • Correctness of results. The solvers are not allowed to produce wrong answers (e.g. 'satisfiable' for an unsatisfiable instance, but of course they may time-out).
  • Sufficient performance. Solvers must meet some minimal speed requirements. We will set the level of required performance particularly high for the first qualification round to stimulate improvement of solvers.

Cheating (e.g. pre-computing results) is not allowed. We will check all solvers on additional instances, and fraudulent solvers will be disqualified (the organizing committee will decide on this).

Dates

The dead-lines for the qualification rounds are February 10 for the first one, and March 12 for the second one.

First Qualification Round

The first qualification round has taken place on and after February 10. Each solver had to tackle 50 instances. In contrast to the SAT-Race itself, the run-time limit was set to 20 minutes per solver and instance. Solvers which were able to solve at least 40 instances are already qualified for SAT-Race 2008. These solvers are:

Main Track (Sequential, CNF)
SolverSubmitterAffiliation #Solved
preSAT Cédric Piette CRIL-CNRS, France 48
RsatKnot PipatsrisawatUCLA, USA 47
CMUSATHimanshu JainCMU, USA 46
MXCDavid BregmanSimon Fraser University, Canada 45
kwJohan AlfredssonOepir Consulting, Sweden 42
eSATSaid JabbourCRIL Lens, France 42
SATzillaLin XuUBC, Canada 42
LocalMinisatVadim RyvchinTechnion, Israel 42 / 41
picosatArmin BiereJKU Linz, Austria 41

Special Track 1 (Parallel, CNF)
SolverSubmitterAffiliation #Solved
pMiniSat Geoffrey Chu University of Melbourne, Australia 45

Special Track 2 (Sequential, AIG)
SolverSubmitterAffiliation #Solved
CMUSAT-AIG Himanshu Jain CMU, USA 49
NFLSATHimanshu JainCMU, USA 47
picoaigersatArmin BiereJKU Linz, Austria 46
MiniCircNiklas EenCadence Research Lab, USA 44 / 50
kw_aigerJohan AlfredssonOepir Consulting, Sweden 44

Two numbers in the '#Solved'-column indicate that the solver participated in both qualification rounds.

Second Qualification Round

The second qualification round has taken place on and after March 12. Each solver had to tackle 50 instances, the instances were the same as in the first qualification round. In contrast to the SAT-Race itself, the run-time limit was set to 20 minutes per solver and instance. Solvers which were able to solve at least 20 instances (in either the first or second qualification round) are qualified for SAT-Race 2008. These solvers are (in addition to the ones that already passed the first qualification round):

Main Track (Sequential, CNF)
SolverSubmitterAffiliation #Solved
MiniSat Niklas Sörensson Independent, Sweden - / 43
BarcelogicRobert Nieuwenhuis Tech. Univ. Catalonia, Spain- / 39
SpearDomagoj BabicUBC, Canada- / 34
ClaspTorsten Schaub University of Potsdam, Germany33 / -
TinisatJinbo HuangNICTA, Australia 30 / 30
EurekaVadim RyvchinIntel, Israel 26 / 27
HinotosFlorian Letombe University of Southampton, UK24 / 24
SAT4J 2.0Daniel Le BerreCRIL-CNRS, France 21 / 20

Special Track 1 (Parallel, CNF)
SolverSubmitterAffiliation #Solved
ManySat Youssef Hamadi Microsoft Research, UK - / 46
MiraXTTobias Schubert University of Freiburg, Germany(45) / 39

Special Track 2 (Sequential, AIG)
SolverSubmitterAffiliation #Solved
MiniSat++ Niklas Sörensson Independent, Sweden - / 50

Two first number in the '#Solved'-column shows the result of the first, the second number of the
second qualification round. Numbers in parentheses indicate that the solver produced errors.