THE F INAL RESULTS ! Daniel le Berre, Olivier Roussel, Laurent Simon - - PowerPoint PPT Presentation

the f inal results
SMART_READER_LITE
LIVE PREVIEW

THE F INAL RESULTS ! Daniel le Berre, Olivier Roussel, Laurent Simon - - PowerPoint PPT Presentation

Rules Participants First Stage Winners Conclusion Certified UNSAT AIG T HE SAT07 C ONTEST THE F INAL RESULTS ! Daniel le Berre, Olivier Roussel, Laurent Simon { leberre,roussel } @cril.univ-artois.fr, simon@lri.fr May, 2007 The SAT


slide-1
SLIDE 1

Rules Participants First Stage Winners Conclusion Certified UNSAT AIG

THE SAT’07 CONTEST

THE FINAL RESULTS!

Daniel le Berre, Olivier Roussel, Laurent Simon

{leberre,roussel}@cril.univ-artois.fr, simon@lri.fr

May, 2007 The SAT conference

Lisboa, Portugal

slide-2
SLIDE 2

Rules Participants First Stage Winners Conclusion Certified UNSAT AIG

ROAD MAP

1 HISTORY, RULES AND GOALS 2 THE PARTICIPANTS 3 THE FIRST STAGE 4 THE WINNERS 5 CONCLUSION AND NEXT CONTESTS 6 CERTIFIED UNSAT SPECIAL TRACK 7 AND INVERTER GRAPH SPECIAL TRACK

slide-3
SLIDE 3

Rules Participants First Stage Winners Conclusion Certified UNSAT AIG

MY SOLVER DID IT? DID WHAT?

slide-4
SLIDE 4

Rules Participants First Stage Winners Conclusion Certified UNSAT AIG

MY SOLVER DID IT? DID WHAT?

YES, YOUR SOLVERS MUST HAVE DONE IT... My Solver

slide-5
SLIDE 5

Rules Participants First Stage Winners Conclusion Certified UNSAT AIG

MY SOLVER DID IT? DID WHAT?

YES, YOUR SOLVERS MUST HAVE DONE IT... My Solver win the contest

slide-6
SLIDE 6

Rules Participants First Stage Winners Conclusion Certified UNSAT AIG

MY SOLVER DID IT? DID WHAT?

YES, YOUR SOLVERS MUST HAVE DONE IT... My Solver win the contest My Solver did

slide-7
SLIDE 7

Rules Participants First Stage Winners Conclusion Certified UNSAT AIG

MY SOLVER DID IT? DID WHAT?

YES, YOUR SOLVERS MUST HAVE DONE IT... My Solver win the contest My Solver did compile (at least once)

slide-8
SLIDE 8

Rules Participants First Stage Winners Conclusion Certified UNSAT AIG

MY SOLVER DID IT? DID WHAT?

YES, YOUR SOLVERS MUST HAVE DONE IT... My Solver win the contest My Solver did compile (at least once) My Solver did

slide-9
SLIDE 9

Rules Participants First Stage Winners Conclusion Certified UNSAT AIG

MY SOLVER DID IT? DID WHAT?

YES, YOUR SOLVERS MUST HAVE DONE IT... My Solver win the contest My Solver did compile (at least once) My Solver did core dumps

slide-10
SLIDE 10

Rules Participants First Stage Winners Conclusion Certified UNSAT AIG

MY SOLVER DID IT? DID WHAT?

YES, YOUR SOLVERS MUST HAVE DONE IT... My Solver win the contest My Solver did compile (at least once) My Solver did core dumps My Solver did

slide-11
SLIDE 11

Rules Participants First Stage Winners Conclusion Certified UNSAT AIG

MY SOLVER DID IT? DID WHAT?

YES, YOUR SOLVERS MUST HAVE DONE IT... My Solver win the contest My Solver did compile (at least once) My Solver did core dumps My Solver did spend my free-time looking for bugs

slide-12
SLIDE 12

Rules Participants First Stage Winners Conclusion Certified UNSAT AIG

MY SOLVER DID IT? DID WHAT?

YES, YOUR SOLVERS MUST HAVE DONE IT... My Solver win the contest My Solver did compile (at least once) My Solver did core dumps My Solver did spend my free-time looking for bugs My Solver did

slide-13
SLIDE 13

Rules Participants First Stage Winners Conclusion Certified UNSAT AIG

MY SOLVER DID IT? DID WHAT?

YES, YOUR SOLVERS MUST HAVE DONE IT... My Solver win the contest My Solver did compile (at least once) My Solver did core dumps My Solver did spend my free-time looking for bugs My Solver did travel by email

slide-14
SLIDE 14

Rules Participants First Stage Winners Conclusion Certified UNSAT AIG

MY SOLVER DID IT? DID WHAT?

YES, YOUR SOLVERS MUST HAVE DONE IT... My Solver win the contest My Solver did compile (at least once) My Solver did core dumps My Solver did spend my free-time looking for bugs My Solver did travel by email My Solver did

slide-15
SLIDE 15

Rules Participants First Stage Winners Conclusion Certified UNSAT AIG

MY SOLVER DID IT? DID WHAT?

YES, YOUR SOLVERS MUST HAVE DONE IT... My Solver win the contest My Solver did compile (at least once) My Solver did core dumps My Solver did spend my free-time looking for bugs My Solver did travel by email My Solver did contribute to the global warming

slide-16
SLIDE 16

Rules Participants First Stage Winners Conclusion Certified UNSAT AIG

MY SOLVER DID IT? DID WHAT?

YES, YOUR SOLVERS MUST HAVE DONE IT... My Solver win the contest My Solver did compile (at least once) My Solver did core dumps My Solver did spend my free-time looking for bugs My Solver did travel by email My Solver did contribute to the global warming My Solver did

slide-17
SLIDE 17

Rules Participants First Stage Winners Conclusion Certified UNSAT AIG

MY SOLVER DID IT? DID WHAT?

YES, YOUR SOLVERS MUST HAVE DONE IT... My Solver win the contest My Solver did compile (at least once) My Solver did core dumps My Solver did spend my free-time looking for bugs My Solver did travel by email My Solver did contribute to the global warming My Solver did something sometime...

slide-18
SLIDE 18

Rules Participants First Stage Winners Conclusion Certified UNSAT AIG

ROAD MAP

1 HISTORY, RULES AND GOALS

Rules

2

The participants

3

The First Stage

4

The Winners

5

Conclusion and next contests

6

Certified UNSAT Special Track

7

And Inverter Graph Special Track

slide-19
SLIDE 19

Rules Participants First Stage Winners Conclusion Certified UNSAT AIG

THE SAT-“007” AGENTS

THREE JUDGES THAT HAVE LICENCES TO KILL... SOLVERS

EWALD SPECKENMEYER Institut f¨ ur Informatik, K¨

  • ln

GEOFF SUTCLIFF Department of Computer Science, University of Miami LINTAO ZHANG Microsoft Research, Silicon Valley

slide-20
SLIDE 20

Rules Participants First Stage Winners Conclusion Certified UNSAT AIG

THE SAT-“007” AGENTS

THREE JUDGES THAT HAVE LICENCES TO KILL... SOLVERS

EWALD SPECKENMEYER Institut f¨ ur Informatik, K¨

  • ln

GEOFF SUTCLIFF Department of Computer Science, University of Miami LINTAO ZHANG Microsoft Research, Silicon Valley All the decisions were taken in agreement with the judges.

slide-21
SLIDE 21

Rules Participants First Stage Winners Conclusion Certified UNSAT AIG

THE SPECIAL TRACKS

CERTIFIED UNSAT solvers must output a certificate of unsatisfiability. Proof format / checker provided by Allen van Gelder. See Allen’s poster, wait for his presentation AND-INVERTER GRAPHS INPUT Using Armin Biere’s And-Inverter Graphs as input format (AIGER). May be used for compact encoding of circuits. A 2-slides results will be given.

slide-22
SLIDE 22

Rules Participants First Stage Winners Conclusion Certified UNSAT AIG

DIVIDE (PRICES) TO CONQUER (GLORY)

TWO MAIN DIVISIONS: COMPETITION AND DEMONSTRATION

COMPETITION the source code of the solver must be available after the competition. DEMONSTRATION a binary version of the solver must be available for research purpose.

slide-23
SLIDE 23

Rules Participants First Stage Winners Conclusion Certified UNSAT AIG

DIVIDE (PRICES) TO CONQUER (GLORY)

TWO MAIN DIVISIONS: COMPETITION AND DEMONSTRATION

COMPETITION the source code of the solver must be available after the competition. DEMONSTRATION a binary version of the solver must be available for research purpose. IDEAS Participation to the competition must benefit to the community By providing source code, binary or benchmarks By supporting the conference and the competition

slide-24
SLIDE 24

Rules Participants First Stage Winners Conclusion Certified UNSAT AIG

DIVIDE (PRICES) TO CONQUER (GLORY)

TWO MAIN DIVISIONS: COMPETITION AND DEMONSTRATION

COMPETITION the source code of the solver must be available after the competition. DEMONSTRATION a binary version of the solver must be available for research purpose. IDEAS Participation to the competition must benefit to the community By providing source code, binary or benchmarks By supporting the conference and the competition Some “suspicious” solvers were moved to demonstration during the

  • contest. Results of fixed versions of those solvers are available as

well in the demonstration division, first stage.

slide-25
SLIDE 25

Rules Participants First Stage Winners Conclusion Certified UNSAT AIG

THE SCORING SCHEME

BENCHMARK PURSE to be divided equally among the solvers able to solve it. SPEED PURSE to be divided unequally among the solvers able to solve a given benchmark. SERIES an extra credit is given for each series solved. SOLVER his score is the sum of the credits obtained per benchmarks solved.

slide-26
SLIDE 26

Rules Participants First Stage Winners Conclusion Certified UNSAT AIG

WHO CAN WIN?

ONLY PARTICIPANTS THAT WEAR CONTEST T-SHIRT CAN WIN

Three categories: industrial, crafted and random Three specialties: SAT, UNSAT and SAT+UNSAT Three medals: gold, silver and bronze So we have 27 awards.

slide-27
SLIDE 27

Rules Participants First Stage Winners Conclusion Certified UNSAT AIG

INVARIANTS RULES AND DETAILS

Only 3 solvers per submitter can enter the first stage, competition division. Only 1 solver per submitter can enter the second stage, competition division.

slide-28
SLIDE 28

Rules Participants First Stage Winners Conclusion Certified UNSAT AIG

ROAD MAP

1

History, rules and Goals

2 THE PARTICIPANTS

Solvers Benchmarks

3

The First Stage

4

The Winners

5

Conclusion and next contests

6

Certified UNSAT Special Track

7

And Inverter Graph Special Track

slide-29
SLIDE 29

Rules Participants First Stage Winners Conclusion Certified UNSAT AIG

LOCAL SEARCH SOLVERS

adaptg2wsat0, adaptg2wsatp Chu-Min Li, Wanxia Wei and Harry Zhang Decreasing Variable Scoring adaptnovelty Holger Hoos and Dave Tompkins SAT 2004 winner adaptg2wsat+, FH, Hybrid1, Wanxia Wei, Chu-Min Li and Harry Zhang Switching between 2 or 4 heuristics, auto-adaptative gnovelty+ Duc Nghia Pham and Charles Gretton SAPS, SAPSRT Dave A. D. Tompkins, Frank Hutter, and Holger H. Hoos Dynamic Local Search Ranov Duc Nghia Pham and Anbu- lagan

slide-30
SLIDE 30

Rules Participants First Stage Winners Conclusion Certified UNSAT AIG

COMPLETE, CDCL SOLVERS

CMUSAT, CMUSAT-base Himanshu Jain and Edmund Clarke CDCL, Standard Template Lib. Minisat-2007 Niklas Sorensson and Niklas Een Who don’t know Minisat? Picosat Armin Biere Rsat Knot Pipatsrisawat and Ad- nan Darwiche MXC David R. Bregman and David

  • G. Mitchell

Best student solver (sat-race 06) Mmisat Monahov Ivan Tinisat, TiniSatELite Jinbo Huang pronounce “teeny sat” SAT7 Christian Kern, Moham- mad Khaleghi, Stefan Kugele, Christian Schallhart, Michael Tautschnig and Andreas Weis

slide-31
SLIDE 31

Rules Participants First Stage Winners Conclusion Certified UNSAT AIG

COMPLETE, NON-CDCL SOLVERS

dewSatz-1a Anbulagan LA Saturation with restriction in satz kcnfs-2004, kcnfs-2006, kcnfs-smp Gilles Dequen and Olivier Dubois march-ks Marijn Heule and Hans van Maaren tts-4.0 Ivor Spence stands for “Ternary Tree Solver” UnitMarch Marijn Heule, Denis de Leeuw Duarte, and Hans van Maaren multi-bit assignements

slide-32
SLIDE 32

Rules Participants First Stage Winners Conclusion Certified UNSAT AIG

OTHER SOLVERS

SATzilla-CRAFTED, SATzilla-RANDOM, SATzilla-ALL Lin Xu, Frank Hutter, Holger

  • H. Hoos and Kevin Leyton-

Brown Specialized Portfolio MiraXTv1, MiraXTv2, MiraXTv3 Tobias Schubert, Matthew Lewis, Natalia Kalinnik and Bernd Becker Multi-threaded solver

slide-33
SLIDE 33

Rules Participants First Stage Winners Conclusion Certified UNSAT AIG

DEMONSTRATION DIVISION

SUBMITTED IN COMPETITION DIVISION

barcelogic Robert Nieuwenhuis, Albert Oliveras and Tomas Lioret minimarch Siert Wieringa, Hans van Maaren and Marijn Heule dewSatz Anbulagan

slide-34
SLIDE 34

Rules Participants First Stage Winners Conclusion Certified UNSAT AIG

DEMONSTRATION DIVISION

adaptg2wsat Chu-Min Li, Wanxia Wei and Harry Zhang SAT4J-1.7 Daniel Le Berre Spear, Spear-FH, Spear-FHS Domagoj Babic Ornithorynque Olivier Roussel

slide-35
SLIDE 35

Rules Participants First Stage Winners Conclusion Certified UNSAT AIG

CRAFTED

Series Subseries contest02 (10) contest03 looksrandom (10), others (10) contest04 connamacher (10), others (10) contest05 counting-clq (10), counting-php (10), jarvisalo (10), others (20), pebbling (10), phnf (10), QG (10), sabharwal (10) Difficult contest-02-03-04 (9), contest05 (36) spence hard (6), medium (10)

slide-36
SLIDE 36

Rules Participants First Stage Winners Conclusion Certified UNSAT AIG

A VIEW OF CRAFTED BENCHMARKS

100 1000 10000 100000 1e+06 1e+07 10 100 1000 10000 100000 1e+06 Total Size #Variables contest02-mix contest03-looksrandom contest03-others contest04-connamacher contest04-others contest05-counting-clq contest05-counting-php contest05-jarvisalo contest05-others contest05-pebbling contest05-phnf contest05-QG contest05-sabharwal Difficult-contest-02-03-04 Difficult-contest05 spence-hard spence-medium

slide-37
SLIDE 37

Rules Participants First Stage Winners Conclusion Certified UNSAT AIG

RANDOM

ON TRESHOLD BENCHMARKS 5 SAT benchmarks, 5 UNSAT benchmarks per series 3SAT from 360 to 650 variables (7 series) 5SAT from 90 to 130 variables (7 series) 7SAT from 45 to 75 variables (7 series)

slide-38
SLIDE 38

Rules Participants First Stage Winners Conclusion Certified UNSAT AIG

RANDOM

ON TRESHOLD BENCHMARKS 5 SAT benchmarks, 5 UNSAT benchmarks per series 3SAT from 360 to 650 variables (7 series) 5SAT from 90 to 130 variables (7 series) 7SAT from 45 to 75 variables (7 series) 2+0.7SAT from 3500 to 6500 variables (4 series) 2+p0.8SAT from 1295 to 2405 variables (4 series) 2+p0.9SAT from 630 to 1170 variables (4 series)

slide-39
SLIDE 39

Rules Participants First Stage Winners Conclusion Certified UNSAT AIG

RANDOM

ON TRESHOLD BENCHMARKS 5 SAT benchmarks, 5 UNSAT benchmarks per series 3SAT from 360 to 650 variables (7 series) 5SAT from 90 to 130 variables (7 series) 7SAT from 45 to 75 variables (7 series) 2+0.7SAT from 3500 to 6500 variables (4 series) 2+p0.8SAT from 1295 to 2405 variables (4 series) 2+p0.9SAT from 630 to 1170 variables (4 series) LARGE BENCHMARKS : 10 (hopefully) SAT benchmarks per series 3SAT from 4000 to 19000 variables (6 series) 5SAT from 600 to 1100 variables (6 series) 7SAT from 140 to 240 variables (6 series)

slide-40
SLIDE 40

Rules Participants First Stage Winners Conclusion Certified UNSAT AIG

A VIEW OF RANDOM BENCHMARKS

1000 10000 100000 1e+06 10 100 1000 10000 100000 Total Size #Variables 2+p-p0.7-v3500 2+p-p0.7-v4500 2+p-p0.7-v5500 2+p-p0.7-v6500 2+p-p0.8-v1295 2+p-p0.8-v1665 2+p-p0.8-v2035 2+p-p0.8-v2405 2+p-p0.9-v1170 2+p-p0.9-v630 2+p-p0.9-v810 2+p-p0.9-v990 LargeSize-3SAT-v10000 LargeSize-3SAT-v13000 LargeSize-3SAT-v16000 LargeSize-3SAT-v19000 LargeSize-3SAT-v4000 LargeSize-3SAT-v7000 LargeSize-5SAT-v1000 LargeSize-5SAT-v1100 LargeSize-5SAT-v600 LargeSize-5SAT-v700 LargeSize-5SAT-v800 LargeSize-5SAT-v900 LargeSize-7SAT-v140 LargeSize-7SAT-v160 LargeSize-7SAT-v180 LargeSize-7SAT-v200 LargeSize-7SAT-v220 LargeSize-7SAT-v240 OnTreshold-3SAT-v360 OnTreshold-3SAT-v400 OnTreshold-3SAT-v450 OnTreshold-3SAT-v500 OnTreshold-3SAT-v550 OnTreshold-3SAT-v600 OnTreshold-3SAT-v650 OnTreshold-5SAT-v100 OnTreshold-5SAT-v110 OnTreshold-5SAT-v120 OnTreshold-5SAT-v130 OnTreshold-5SAT-v70 OnTreshold-5SAT-v80 OnTreshold-5SAT-v90 OnTreshold-7SAT-v45 OnTreshold-7SAT-v50 OnTreshold-7SAT-v55 OnTreshold-7SAT-v60 OnTreshold-7SAT-v65 OnTreshold-7SAT-v70 OnTreshold-7SAT-v75

slide-41
SLIDE 41

Rules Participants First Stage Winners Conclusion Certified UNSAT AIG

INDUSTRIAL

Series Subseries anbulagan hard-sat (7), hard-unsat (24), medium-sat (10), medium-unsat (6) babic dspam (10), hsatv17 (10), xinetd (10) crypto (10) fuhs hard (6), medium (10) grieu (10) jarvisalo (7) manolios (10) narain (5) palacios hard (7), medium (10), uts (10) velev vliw-sat-4.0 (10), vliw-unsat-4.0 (10) zarpas IBM-FV-2002-13-rule-1 (19), IBM-FV-2002- 31-1-rule-1 (16), IBM-FV-2004-30 (18)

slide-42
SLIDE 42

Rules Participants First Stage Winners Conclusion Certified UNSAT AIG

A VIEW OF INDUSTRIAL BENCHMARKS

1000 10000 100000 1e+06 1e+07 1e+08 100 1000 10000 100000 1e+06 1e+07 Total Size #Variables anbulagan-hard-sat anbulagan-hard-unsat anbulagan-medium-sat anbulagan-medium-unsat babic-dspam babic-hsatv17 babic-xinetd crypto-crypto fuhs-hard fuhs-medium grieu-grieu jarvisalo-jarvisalo manolios-manolios narain-narain palacios-hard palacios-medium palacios-uts velev-vliw_sat_4.0 velev-vliw_unsat_2.0 zarpas-IBM_FV_2002_13_rule_1 zarpas-IBM_FV_2002_31_1_rule_1 zarpas-IBM_FV_2004_30

slide-43
SLIDE 43

Rules Participants First Stage Winners Conclusion Certified UNSAT AIG

ROAD MAP

1

History, rules and Goals

2

The participants

3 THE FIRST STAGE

Overall Pictures Remaining benchmarks

4

The Winners

5

Conclusion and next contests

6

Certified UNSAT Special Track

7

And Inverter Graph Special Track

slide-44
SLIDE 44

Rules Participants First Stage Winners Conclusion Certified UNSAT AIG

ALL SOLVERS, ALL BENCHMARKS

50 100 150 200 250 300 350 400 450 200 400 600 800 1000 1200 1400

#Solved CPU−Time needed (s)

← Mmisat−2007−02−08 ← TTS−4.0 ← ornithorynque−0.1−alpha ← SAT4J−−JVM−PARAM−CHANGED−SAT−2007 ← UnitMarch−2007−02−08 ← DEWSATZ−1A−2007−02−08 ← tinisat−2007−02−08 ← Spear−2007−02−12 ← TiniSatELite−2007−02−08 ← CMUSAT−BASE−2007−02−08 ← Spear−FH−1.0 ← SAT4J−SAT−2007 ← Barcelogic−Fixed−2007−04−13 ← sapsrt−2007−02−08 ← Spear−FHS−1.0 ← saps−2007−02−08 ← DEWSATZ−2007−04−26−(fixed) ← Rsat−2007−02−08 ← CMUSAT−2007−02−08 ← KCNFS−2006 ← gnovelty+−2007−02−08 ← MiraXT−v3 ← KCNFS−SMP ← SAT7−2007−02−08 ← picosat−535 ← KCNFS−2004 ← MiraXT−v1 ← adaptnovelty−2007−02−08 ← MiraXT−v2 ← ranov−2007−02−08 ← adaptg2wsatp−2007−02−08 ← FH−2007−02−08 ← adaptg2wsat−2007−02−08 ← Hybrid1−2007−02−08 ← adaptg2wsat+−2007−02−08 ← adaptg2wsat0−2007−02−08 ← MXC−2007−02−08 ← March−KS−2007−02−08 ← minisat−SAT−2007 ← minisat−SAT−2007−(with−assertions) ← minimarch−2007−04−26−(fixed) ← SATzilla−FULL ← SATzilla−RANDOM ← SATzilla−CRAFTED Mmisat−2007−02−08 (3) TTS−4.0 (40)

  • rnithorynque−0.1−alpha (53)

SAT4J−SAT−2007 (66) UnitMarch−2007−02−08 (69) DEWSATZ−1A−2007−02−08 (126) tinisat−2007−02−08 (145) Spear−2007−02−12 (151) TiniSatELite−2007−02−08 (162) CMUSAT−BASE−2007−02−08 (165) Spear−FH−1.0 (165) SAT4J−SAT−2007 (167) Barcelogic−Fixed−2007−04−13 (168) sapsrt−2007−02−08 (170) Spear−FHS−1.0 (171) saps−2007−02−08 (173) DEWSATZ−2007−04−26−(fixed) (176) Rsat−2007−02−08 (186) CMUSAT−2007−02−08 (187) KCNFS−2006 (190) gnovelty+−2007−02−08 (206) MiraXT−v3 (206) KCNFS−SMP (207) SAT7−2007−02−08 (210) picosat−535 (216) KCNFS−2004 (219) MiraXT−v1 (220) adaptnovelty−2007−02−08 (224) MiraXT−v2 (224) ranov−2007−02−08 (228) adaptg2wsatp−2007−02−08 (233) FH−2007−02−08 (235) adaptg2wsat−2007−02−08 (236) Hybrid1−2007−02−08 (241) adaptg2wsat+−2007−02−08 (242) adaptg2wsat0−2007−02−08 (242) MXC−2007−02−08 (242) March−KS−2007−02−08 (253) minisat−SAT−2007 (267) minisat−SAT−2007−(with−assertions) (269) minimarch−2007−04−26−(fixed) (280) SATzilla−FULL (325) SATzilla−RANDOM (339) SATzilla−CRAFTED (362)

slide-45
SLIDE 45

Rules Participants First Stage Winners Conclusion Certified UNSAT AIG

CLUSTERING OF SOLVERS

50 99 149 198 248 297 minisat−SAT−2007−(with−assertions) minisat−SAT−2007 minimarch−2007−04−26−(fixed) SAT7−2007−02−08 MXC−2007−02−08 SAT4J−SAT−2007 MiraXT−v3 MiraXT−v1 MiraXT−v2 TiniSatELite−2007−02−08 tinisat−2007−02−08 Spear−FHS−1.0 Spear−FH−1.0 Spear−2007−02−12 Rsat−2007−02−08 Barcelogic−Fixed−2007−04−13 CMUSAT−BASE−2007−02−08 CMUSAT−2007−02−08 picosat−535 UnitMarch−2007−02−08 TTS−4.0 Mmisat−2007−02−08

  • rnithorynque−0.1−alpha

SAT4J−−JVM−PARAM−CHANGED−SAT−2007 March−KS−2007−02−08 KCNFS−SMP KCNFS−2006 KCNFS−2004 DEWSATZ−1A−2007−02−08 DEWSATZ−2007−04−26−(fixed) SATzilla−RANDOM SATzilla−CRAFTED SATzilla−FULL sapsrt−2007−02−08 saps−2007−02−08 gnovelty+−2007−02−08 ranov−2007−02−08 adaptnovelty−2007−02−08 Hybrid1−2007−02−08 FH−2007−02−08 adaptg2wsat+−2007−02−08 adaptg2wsat0−2007−02−08 adaptg2wsat−2007−02−08 adaptg2wsatp−2007−02−08

Solvers Distance

267 269 280 210 242 167 224 220 206 162 145 171 165 151 186 168 165 187 216 69 40 3 53 66 253 190 219 207 126 176 325 339 362 170 173 206 228 224 241 235 242 242 236 233

269, 267 179, 157 226, 200 252, 225 240, 212 186, 157 261, 218 268, 211 273, 208 257, 207 41, 2 175, 132 244, 200 287, 190 232, 172 253, 183 222, 153 193, 127 255, 197 271, 172 85, 2 210, 144 209, 143 308, 240 292, 152 220, 98 188, 114 245, 84 126, 0 320, 179 258, 62 274, 62 397, 304 269, 136 338, 123 185, 0 414, 262 286, 90 367, 56 383, 0 431, 85 501, 0 617, 0

slide-46
SLIDE 46

Rules Participants First Stage Winners Conclusion Certified UNSAT AIG

ALL SOLVERS ON CRAFTED BENCHMARKS

10 20 30 40 50 60 70 80 200 400 600 800 1000 1200 1400

#Solved CPU−Time needed (s)

← gnovelty+−2007−02−08 ← sapsrt−2007−02−08 ← Mmisat−2007−02−08 ← saps−2007−02−08 ← adaptnovelty−2007−02−08 ← UnitMarch−2007−02−08 ← adaptg2wsat0−2007−02−08 ← adaptg2wsatp−2007−02−08 ← adaptg2wsat−2007−02−08 ← adaptg2wsat+−2007−02−08 ← ranov−2007−02−08 ← FH−2007−02−08 ← Hybrid1−2007−02−08 ← ornithorynque−0.1−alpha ← KCNFS−2006 ← KCNFS−2004 ← DEWSATZ−2007−04−26−(fixed) ← KCNFS−SMP ← DEWSATZ−1A−2007−02−08 ← tinisat−2007−02−08 ← TiniSatELite−2007−02−08 ← Spear−2007−02−12 ← SAT4J−SAT−2007 ← Spear−FHS−1.0 ← March−KS−2007−02−08 ← Spear−FH−1.0 ← CMUSAT−BASE−2007−02−08 ← Rsat−2007−02−08 ← TTS−4.0 ← Barcelogic−Fixed−2007−04−13 ← SATzilla−FULL ← picosat−535 ← SAT7−2007−02−08 ← CMUSAT−2007−02−08 ← MXC−2007−02−08 ← MiraXT−v3 ← MiraXT−v1 ← MiraXT−v2 ← minimarch−2007−04−26−(fixed) ← SATzilla−CRAFTED ← SATzilla−RANDOM ← minisat−SAT−2007−(with−assertions) ← minisat−SAT−2007 gnovelty+−2007−02−08 (1) sapsrt−2007−02−08 (1) Mmisat−2007−02−08 (2) saps−2007−02−08 (2) adaptnovelty−2007−02−08 (3) UnitMarch−2007−02−08 (3) adaptg2wsat0−2007−02−08 (4) adaptg2wsatp−2007−02−08 (4) adaptg2wsat−2007−02−08 (6) adaptg2wsat+−2007−02−08 (6) ranov−2007−02−08 (6) FH−2007−02−08 (7) Hybrid1−2007−02−08 (8)

  • rnithorynque−0.1−alpha (9)

KCNFS−2006 (14) KCNFS−2004 (14) DEWSATZ−2007−04−26−(fixed) (15) KCNFS−SMP (15) DEWSATZ−1A−2007−02−08 (20) tinisat−2007−02−08 (25) TiniSatELite−2007−02−08 (27) Spear−2007−02−12 (31) SAT4J−SAT−2007 (32) Spear−FHS−1.0 (32) March−KS−2007−02−08 (33) Spear−FH−1.0 (33) CMUSAT−BASE−2007−02−08 (37) Rsat−2007−02−08 (37) TTS−4.0 (37) Barcelogic−Fixed−2007−04−13 (40) SATzilla−FULL (44) picosat−535 (46) SAT7−2007−02−08 (46) CMUSAT−2007−02−08 (52) MXC−2007−02−08 (53) MiraXT−v3 (55) MiraXT−v1 (55) MiraXT−v2 (57) minimarch−2007−04−26−(fixed) (67) SATzilla−CRAFTED (69) SATzilla−RANDOM (69) minisat−SAT−2007−(with−assertions) (71) minisat−SAT−2007 (71)

slide-47
SLIDE 47

Rules Participants First Stage Winners Conclusion Certified UNSAT AIG

SCORES OF SOLVERS ON CRAFTED BENCHMAKRS

Solver #Solv.

  • Sc. All
  • Sc. SAT
  • Sc. UNS

TTS 4.0 37 (2 35) 36458 252 32300 SATzilla CRAFTED 69 (24 45) 30337 8239 22054 March-KS 2007-02-08 33 (15 18) 29134 9621 14048 minisat SAT-2007 71 (22 49) 24601 8652 15905 MXC 2007-02-08 53 (15 38) 15504 2133 13596 MiraXT v3 57 (18 39) 14739 6012 8683 MiraXT v1 55 (15 40) 13793 5318 8431 CMUSAT 2007-02-08 52 (15 37) 13728 5779 9405 SATzilla RANDOM 44 (22 22) 13113 8117 5667 MiraXT v2 55 (15 40) 10923 2837 8543 picosat 535 46 (20 26) 10514 4838 5841 Rsat 2007-02-08 37 (11 26) 9164 2583 6574 SAT7 2007-02-08 46 (16 30) 7353 2398 5216 CMUSAT-BASE 2007-02-08 37 (14 23) 5276 1694 4074 DEWSATZ-1A 2007-02-08 20 (4 16) 4609 434 4258 TiniSatELite 2007-02-08 27 (9 18) 3574 1135 2854 tinisat 2007-02-08 25 (10 15) 3134 1131 2418 KCNFS SMP 15 (2 13) 2222 272 2015 KCNFS 14 2149 275 1939

slide-48
SLIDE 48

Rules Participants First Stage Winners Conclusion Certified UNSAT AIG

SCORES OF SOLVERS ON CRAFTED BENCHMAKRS

Solver #Solv.

  • Sc. All
  • Sc. SAT
  • Sc. UNS

TTS 4.0 37 (2 35) 36458 252 32300 SATzilla CRAFTED 69 (24 45) 30337 8239 22054 March-KS 2007-02-08 33 (15 18) 29134 9621 14048 minisat SAT-2007 71 (22 49) 24601 8652 15905 MXC 2007-02-08 53 (15 38) 15504 2133 13596 MiraXT v3 57 (18 39) 14739 6012 8683 MiraXT v1 55 (15 40) 13793 5318 8431 CMUSAT 2007-02-08 52 (15 37) 13728 5779 9405 SATzilla RANDOM 44 (22 22) 13113 8117 5667 MiraXT v2 55 (15 40) 10923 2837 8543 picosat 535 46 (20 26) 10514 4838 5841 Rsat 2007-02-08 37 (11 26) 9164 2583 6574 SAT7 2007-02-08 46 (16 30) 7353 2398 5216 CMUSAT-BASE 2007-02-08 37 (14 23) 5276 1694 4074 DEWSATZ-1A 2007-02-08 20 (4 16) 4609 434 4258 TiniSatELite 2007-02-08 27 (9 18) 3574 1135 2854 tinisat 2007-02-08 25 (10 15) 3134 1131 2418 KCNFS SMP 15 (2 13) 2222 272 2015 KCNFS 14 2149 275 1939

slide-49
SLIDE 49

Rules Participants First Stage Winners Conclusion Certified UNSAT AIG

THE TTS CASE

SCORING NOT SO STRANGE BEHAVIOR

How can TTS have a so great score while solving so few benchmarks?

slide-50
SLIDE 50

Rules Participants First Stage Winners Conclusion Certified UNSAT AIG

THE TTS CASE

SCORING NOT SO STRANGE BEHAVIOR

How can TTS have a so great score while solving so few benchmarks? TTS IS THE ONLY ONE TO SOLVE THE FOLLOWING BENCHMARKS spence/hard/s101-100 spence/hard/s97-100 Hard/contest05/counting-php/harder-fphp-016-015 Hard/contest05/counting-php/easier-fphp-020-015 Hard/contest05/counting-php/harder-fphp-018-017 Difficult/contest05/sabharwal/counting-easier-fphp-014-012 Difficult/contest05/sabharwal/counting-harder-php-018-017 Difficult/contest05/sabharwal/counting-easier-php-018-014 Difficult/contest05/sabharwal/counting-harder-php-014-013 Difficult/contest05/jarvisalo/mod2c-3cage-unsat-10-3 Difficult/contest05/jarvisalo/mod2c-3cage-unsat-10-2

slide-51
SLIDE 51

Rules Participants First Stage Winners Conclusion Certified UNSAT AIG

ALL SOLVERS ON RANDOM BENCHMARKS

50 100 150 200 250 200 400 600 800 1000 1200 1400

#Solved CPU−Time needed (s)

← Mmisat−2007−02−08 ← ornithorynque−0.1−alpha ← Barcelogic−Fixed−2007−04−13 ← tinisat−2007−02−08 ← TiniSatELite−2007−02−08 ← Spear−2007−02−12 ← Spear−FH−1.0 ← Spear−FHS−1.0 ← Rsat−2007−02−08 ← CMUSAT−2007−02−08 ← CMUSAT−BASE−2007−02−08 ← UnitMarch−2007−02−08 ← picosat−535 ← MiraXT−v2 ← MiraXT−v3 ← MiraXT−v1 ← SAT4J−SAT−2007 ← DEWSATZ−1A−2007−02−08 ← SAT7−2007−02−08 ← minisat−SAT−2007−(with−assertions) ← minisat−SAT−2007 ← MXC−2007−02−08 ← minimarch−2007−04−26−(fixed) ← DEWSATZ−2007−04−26−(fixed) ← KCNFS−SMP ← sapsrt−2007−02−08 ← saps−2007−02−08 ← KCNFS−2006 ← SATzilla−CRAFTED ← KCNFS−2004 ← gnovelty+−2007−02−08 ← SATzilla−FULL ← March−KS−2007−02−08 ← SATzilla−RANDOM ← ranov−2007−02−08 ← adaptnovelty−2007−02−08 ← FH−2007−02−08 ← adaptg2wsat−2007−02−08 ← adaptg2wsatp−2007−02−08 ← Hybrid1−2007−02−08 ← adaptg2wsat+−2007−02−08 ← adaptg2wsat0−2007−02−08 Mmisat−2007−02−08 (1)

  • rnithorynque−0.1−alpha (11)

Barcelogic−Fixed−2007−04−13 (29) tinisat−2007−02−08 (30) TiniSatELite−2007−02−08 (32) Spear−2007−02−12 (38) Spear−FH−1.0 (39) Spear−FHS−1.0 (40) Rsat−2007−02−08 (43) CMUSAT−2007−02−08 (46) CMUSAT−BASE−2007−02−08 (48) UnitMarch−2007−02−08 (62) picosat−535 (67) MiraXT−v2 (68) MiraXT−v3 (74) MiraXT−v1 (78) SAT4J−SAT−2007 (78) DEWSATZ−1A−2007−02−08 (84) SAT7−2007−02−08 (95) minisat−SAT−2007−(with−assertions) (100) minisat−SAT−2007 (101) MXC−2007−02−08 (105) minimarch−2007−04−26−(fixed) (125) DEWSATZ−2007−04−26−(fixed) (142) KCNFS−SMP (165) sapsrt−2007−02−08 (167) saps−2007−02−08 (169) KCNFS−2006 (189) SATzilla−CRAFTED (189) KCNFS−2004 (191) gnovelty+−2007−02−08 (203) SATzilla−FULL (205) March−KS−2007−02−08 (208) SATzilla−RANDOM (215) ranov−2007−02−08 (216) adaptnovelty−2007−02−08 (218) FH−2007−02−08 (220) adaptg2wsat−2007−02−08 (222) adaptg2wsatp−2007−02−08 (222) Hybrid1−2007−02−08 (226) adaptg2wsat+−2007−02−08 (227) adaptg2wsat0−2007−02−08 (231)

slide-52
SLIDE 52

Rules Participants First Stage Winners Conclusion Certified UNSAT AIG

SCORES OF SOLVERS ON RANDOM BENCHMARKS

Solver #Solv.

  • Sc. All
  • Sc. SAT
  • Sc. UNS

gnovelty+ 2007-02-08 203 (203 0) 69639 65013 adaptg2wsat0 2007-02-08 231 (231 0) 68244 68285 adaptg2wsat+ 2007-02-08 227 (227 0) 63732 63773 Hybrid1 2007-02-08 226 (226 0) 61927 61968 adaptnovelty 2007-02-08 218 (218 0) 59476 57517 adaptg2wsatp 2007-02-08 222 (222 0) 57506 56880 FH 2007-02-08 220 (220 0) 57102 56477 March-KS 2007-02-08 208 (121 87) 48845 14386 41953 ranov 2007-02-08 216 (216 0) 45428 45469 KCNFS 2004 191 (106 85) 40314 12362 34447 KCNFS 2006 189 (105 84) 38672 11618 33549 sapsrt 2007-02-08 167 (167 0) 37641 37682 SATzilla RANDOM 215 (137 78) 37616 16678 27432 KCNFS SMP 165 (80 85) 37031 7862 34827 saps 2007-02-08 169 (169 0) 32931 32972 SATzilla CRAFTED 189 (143 46) 23670 17606 8358 MXC 2007-02-08 105 (63 42) 10480 5924 6671 minisat SAT-2007 101 (59 42) 9727 5097 6745 SAT7 95 8387 4940 5187

slide-53
SLIDE 53

Rules Participants First Stage Winners Conclusion Certified UNSAT AIG

SCORES OF SOLVERS ON RANDOM BENCHMARKS

Solver #Solv.

  • Sc. All
  • Sc. SAT
  • Sc. UNS

gnovelty+ 2007-02-08 203 (203 0) 69639 65013 adaptg2wsat0 2007-02-08 231 (231 0) 68244 68285 adaptg2wsat+ 2007-02-08 227 (227 0) 63732 63773 Hybrid1 2007-02-08 226 (226 0) 61927 61968 adaptnovelty 2007-02-08 218 (218 0) 59476 57517 adaptg2wsatp 2007-02-08 222 (222 0) 57506 56880 FH 2007-02-08 220 (220 0) 57102 56477 March-KS 2007-02-08 208 (121 87) 48845 14386 41953 ranov 2007-02-08 216 (216 0) 45428 45469 KCNFS 2004 191 (106 85) 40314 12362 34447 KCNFS 2006 189 (105 84) 38672 11618 33549 sapsrt 2007-02-08 167 (167 0) 37641 37682 SATzilla RANDOM 215 (137 78) 37616 16678 27432 KCNFS SMP 165 (80 85) 37031 7862 34827 saps 2007-02-08 169 (169 0) 32931 32972 SATzilla CRAFTED 189 (143 46) 23670 17606 8358 MXC 2007-02-08 105 (63 42) 10480 5924 6671 minisat SAT-2007 101 (59 42) 9727 5097 6745 SAT7 95 8387 4940 5187

slide-54
SLIDE 54

Rules Participants First Stage Winners Conclusion Certified UNSAT AIG

ALL SOLVERS ON INDUSTRIAL BENCHMARKS

20 40 60 80 100 120 200 400 600 800 1000 1200 1400

#Solved CPU−Time needed (s)

← gnovelty+−2007−02−08 ← KCNFS−2004 ← saps−2007−02−08 ← sapsrt−2007−02−08 ← adaptnovelty−2007−02−08 ← TTS−4.0 ← UnitMarch−2007−02−08 ← ranov−2007−02−08 ← adaptg2wsat0−2007−02−08 ← adaptg2wsatp−2007−02−08 ← Hybrid1−2007−02−08 ← adaptg2wsat−2007−02−08 ← FH−2007−02−08 ← adaptg2wsat+−2007−02−08 ← KCNFS−2006 ← March−KS−2007−02−08 ← KCNFS−SMP ← DEWSATZ−2007−04−26−(fixed) ← DEWSATZ−1A−2007−02−08 ← ornithorynque−0.1−alpha ← SAT4J−SAT−2007 ← SAT4J−−JVM−PARAM−CHANGED−SAT−2007 ← SATzilla−RANDOM ← SAT7−2007−02−08 ← CMUSAT−BASE−2007−02−08 ← SATzilla−CRAFTED ← Spear−2007−02−12 ← MiraXT−v1 ← MXC−2007−02−08 ← minimarch−2007−04−26−(fixed) ← SATzilla−FULL ← CMUSAT−2007−02−08 ← MiraXT−v3 ← tinisat−2007−02−08 ← MiraXT−v2 ← Spear−FH−1.0 ← minisat−SAT−2007−(with−assertions) ← minisat−SAT−2007 ← Barcelogic−Fixed−2007−04−13 ← Spear−FHS−1.0 ← picosat−535 ← TiniSatELite−2007−02−08 ← Rsat−2007−02−08 gnovelty+−2007−02−08 (2) KCNFS−2004 (2) saps−2007−02−08 (2) sapsrt−2007−02−08 (2) adaptnovelty−2007−02−08 (3) TTS−4.0 (3) UnitMarch−2007−02−08 (4) ranov−2007−02−08 (6) adaptg2wsat0−2007−02−08 (7) adaptg2wsatp−2007−02−08 (7) Hybrid1−2007−02−08 (7) adaptg2wsat−2007−02−08 (8) FH−2007−02−08 (8) adaptg2wsat+−2007−02−08 (9) KCNFS−2006 (10) March−KS−2007−02−08 (12) KCNFS−SMP (16) DEWSATZ−2007−04−26−(fixed) (19) DEWSATZ−1A−2007−02−08 (22)

  • rnithorynque−0.1−alpha (33)

SAT4J−SAT−2007 (57) SAT4J−SAT−2007 (66) SATzilla−RANDOM (66) SAT7−2007−02−08 (69) CMUSAT−BASE−2007−02−08 (80) SATzilla−CRAFTED (81) Spear−2007−02−12 (82) MiraXT−v1 (83) MXC−2007−02−08 (84) minimarch−2007−04−26−(fixed) (88) SATzilla−FULL (88) CMUSAT−2007−02−08 (89) MiraXT−v3 (89) tinisat−2007−02−08 (90) MiraXT−v2 (91) Spear−FH−1.0 (93) minisat−SAT−2007−(with−assertions) (96) minisat−SAT−2007 (97) Barcelogic−Fixed−2007−04−13 (99) Spear−FHS−1.0 (99) picosat−535 (103) TiniSatELite−2007−02−08 (103) Rsat−2007−02−08 (106)

slide-55
SLIDE 55

Rules Participants First Stage Winners Conclusion Certified UNSAT AIG

SCORES OF SOLVERS ON INDUSTRIAL BENCHMARKS

Solver #Solv.

  • Sc. All
  • Sc. SAT
  • Sc. UNS

Rsat 2007-02-08 106 (47 59) 44626 19422 22801 TiniSatELite 2007-02-08 103 (43 60) 30305 10911 18991 minisat SAT-2007 97 (37 60) 29721 13467 15850 MiraXT v3 89 (35 54) 26595 13292 11359 picosat 535 103 (51 52) 25513 16749 9360 CMUSAT 2007-02-08 89 (28 61) 25339 7275 18186 MiraXT v1 91 (36 55) 23850 12138 11958 tinisat 2007-02-08 90 (40 50) 20241 10264 9274 MXC 2007-02-08 84 (37 47) 19175 10609 9163 MiraXT v2 83 (27 56) 18824 6708 12513 CMUSAT-BASE 2007-02-08 80 (38 42) 18730 10138 8714 SATzilla CRAFTED 81 (35 46) 16727 8006 9317 SATzilla RANDOM 66 (28 38) 14937 6484 8103 SAT7 2007-02-08 69 (30 39) 11454 5883 5898 KCNFS 2006 16 (1 15) 3030 277 2952 KCNFS SMP 10 (0 10) 2877 2877 DEWSATZ-1A 2007-02-08 22 (3 19) 2591 249 2342 FH 2007-02-08 8 (5 3) 1871 1542 329 March-KS 12 1848 217 1630

slide-56
SLIDE 56

Rules Participants First Stage Winners Conclusion Certified UNSAT AIG

SCORES OF SOLVERS ON INDUSTRIAL BENCHMARKS

Solver #Solv.

  • Sc. All
  • Sc. SAT
  • Sc. UNS

Rsat 2007-02-08 106 (47 59) 44626 19422 22801 TiniSatELite 2007-02-08 103 (43 60) 30305 10911 18991 minisat SAT-2007 97 (37 60) 29721 13467 15850 MiraXT v3 89 (35 54) 26595 13292 11359 picosat 535 103 (51 52) 25513 16749 9360 CMUSAT 2007-02-08 89 (28 61) 25339 7275 18186 MiraXT v1 91 (36 55) 23850 12138 11958 tinisat 2007-02-08 90 (40 50) 20241 10264 9274 MXC 2007-02-08 84 (37 47) 19175 10609 9163 MiraXT v2 83 (27 56) 18824 6708 12513 CMUSAT-BASE 2007-02-08 80 (38 42) 18730 10138 8714 SATzilla CRAFTED 81 (35 46) 16727 8006 9317 SATzilla RANDOM 66 (28 38) 14937 6484 8103 SAT7 2007-02-08 69 (30 39) 11454 5883 5898 KCNFS 2006 16 (1 15) 3030 277 2952 KCNFS SMP 10 (0 10) 2877 2877 DEWSATZ-1A 2007-02-08 22 (3 19) 2591 249 2342 FH 2007-02-08 8 (5 3) 1871 1542 329 March-KS 12 1848 217 1630

slide-57
SLIDE 57

Rules Participants First Stage Winners Conclusion Certified UNSAT AIG

SOTAC SOLVERS

OVER ALL BENCHMARKS, FIRST STAGE

Solver #SOTAC

TTS

11

gnovelty+

5

March KS

4

adaptg2wsatp

3

SATzilla

3

adaptnovelty

3

adaptg2wsat+

3

adaptg2wsat0

2

adaptg2wsat

2

Rsat

2 Solver #SOTAC

Hybrid1

2

MiraXT

2

CMUSAT

2

MiraXT

1

picosat

1

DEWSATZ

1

SAT4J

1

minimarch

1

Spear

1

MXC

1

slide-58
SLIDE 58

Rules Participants First Stage Winners Conclusion Certified UNSAT AIG

SOLVED CRAFTED BENCHMARKS, FIRST STAGE

Series SubSeries #Solved #Tot contest02 mix 3 10 contest03 looksrandom 10 contest03

  • thers

1 10 contest04 connamacher 2 10 contest04

  • thers

1 10 contest05 counting-clq 10 contest05 counting-php 3 10 contest05 jarvisalo 10 10 contest05

  • thers

14 20 contest05 pebbling 10 10 contest05 phnf 2 10 contest05 QG 3 10 contest05 sabharwal 10 10 Difficult contest-02-03-04 9 9 Difficult contest05 31 36 spence hard 2 6 spence medium 10 10

slide-59
SLIDE 59

Rules Participants First Stage Winners Conclusion Certified UNSAT AIG

SOLVED RANDOM BENCHMARKS, FIRST STAGE

Series SubSeries #Solved #Tot 2+p-p0.7 v3500 10 10 2+p-p0.7 v6500 6 10 2+p-p0.8 v1295 10 10 2+p-p0.8 v2405 6 11 2+p-p0.9 v1170 5 10 2+p-p0.9 v990 6 10 LargeSize-3SAT v7000 10 10 LargeSize-3SAT v19000 1 10 LargeSize-5SAT v600 10 10 LargeSize-5SAT v1100 7 10 LargeSize-7SAT v140 10 10 LargeSize-7SAT v240 10 OnTreshold-3SAT v360 10 10 OnTreshold-3SAT v650 5 10 OnTreshold-5SAT v70 10 10 OnTreshold-5SAT v130 5 10 OnTreshold-7SAT v45 10 10 OnTreshold-7SAT v75 6 10

slide-60
SLIDE 60

Rules Participants First Stage Winners Conclusion Certified UNSAT AIG

SOLVED INDUSTRIAL BENCHMARKS, FIRST STAGE

Series SubSeries #Solved #Tot anbulagan hard-sat 1 7 anbulagan hard-unsat 1 24 anbulagan medium-sat 10 10 anbulagan medium-unsat 5 6 babic dspam 10 10 babic xinetd 10 10 crypto crypto 10 10 fuhs hard 1 6 fuhs medium 10 10 grieu grieu 10 10 jarvisalo jarvisalo 5 7 manolios manolios 10 10 narain narain 4 5 palacios hard 2 7 palacios uts 10 10 velev vliw-sat-4.0 10 10 velev vliw-unsat-2.0 4 9 zarpas IBM-FV-2002-13-rule-1 19 zarpas IBM-FV-2002-31-1-rule-1 2 16 zarpas IBM-FV-2004-30 10 18

slide-61
SLIDE 61

Rules Participants First Stage Winners Conclusion Certified UNSAT AIG

ROAD MAP

1

History, rules and Goals

2

The participants

3

The First Stage

4 THE WINNERS

Smallest unsolved benchmarks The crafted winners The random winners The industrial winners

5

Conclusion and next contests

6

Certified UNSAT Special Track

7

And Inverter Graph Special Track

slide-62
SLIDE 62

Rules Participants First Stage Winners Conclusion Certified UNSAT AIG

SMALLEST UNSOLVED CRAFTED BENCHMARKS

Benchmark size #clauses #variables spence/hard/s117-100 732 244 117 Hard/contest03/looksrand/hgen8-n260 904 399 212 Hard/contest03/looksrand/hgen8-n320-01 1102 486 260 Hard/contest03/looksrand/hgen8-n320-03 1108 489 260 Hard/contest03/looksrand/hgen8-n320-02 1114 492 260

slide-63
SLIDE 63

Rules Participants First Stage Winners Conclusion Certified UNSAT AIG

SMALLEST UNSOLVED RANDOM BENCHMARKS

Benchmark size #clauses #var OnTres/3SAT/v550/unif-k3-r4.26-v550-c2343-07.UNSAT 7029 2343 550 OnTres/3SAT/v550/unif-k3-r4.26-v550-c2343-03.UNSAT 7029 2343 550 OnTres/3SAT/v550/unif-k3-r4.26-v550-c2343-15.UNSAT 7029 2343 550 OnTres/3SAT/v550/unif-k3-r4.26-v550-c2343-01.UNSAT 7029 2343 550 OnTres/3SAT/v550/unif-k3-r4.26-v550-c2343-20.UNSAT 7029 2343 550

slide-64
SLIDE 64

Rules Participants First Stage Winners Conclusion Certified UNSAT AIG

SMALLEST UNSOLVED INDUSTRIAL BENCHMARKS

Benchmark size #clauses #variables jarvisalo/eq.atree.braun.12.unsat 14874 5726 1694 jarvisalo/eq.atree.braun.13.unsat 17668 6802 2010 fuhs/hard/AProVE07-01 76290 28770 7502 fuhs/hard/AProVE07-25 83706 31884 8920 fuhs/hard/AProVE07-26 211276 79766 21734

slide-65
SLIDE 65

Rules Participants First Stage Winners Conclusion Certified UNSAT AIG

AND NOW...

The results for the crafted category...

slide-66
SLIDE 66

Rules Participants First Stage Winners Conclusion Certified UNSAT AIG

CATEGORY CRAFTED, SAT+UNS SPECIALTY

THE WINNER!

WERE QUALIFIED

CMUSAT 2007-02-08, MXC 2007-02-08, MiraXT v3, Rsat 2007-02-08, SATzilla CRAFTED, minisat SAT-2007, picosat 535

And the winner is...

slide-67
SLIDE 67

Rules Participants First Stage Winners Conclusion Certified UNSAT AIG

CATEGORY CRAFTED, SAT+UNS SPECIALTY

THE WINNER!

WERE QUALIFIED

CMUSAT 2007-02-08, MXC 2007-02-08, MiraXT v3, Rsat 2007-02-08, SATzilla CRAFTED, minisat SAT-2007, picosat 535

And the winner is...

SATzilla CRAFTED

slide-68
SLIDE 68

Rules Participants First Stage Winners Conclusion Certified UNSAT AIG

CATEGORY CRAFTED, SAT+UNS SPECIALTY

THE WINNER!

WERE QUALIFIED

CMUSAT 2007-02-08, MXC 2007-02-08, MiraXT v3, Rsat 2007-02-08, SATzilla CRAFTED, minisat SAT-2007, picosat 535

And the winner is...

SATzilla CRAFTED

minisat SAT-2007

slide-69
SLIDE 69

Rules Participants First Stage Winners Conclusion Certified UNSAT AIG

CATEGORY CRAFTED, SAT+UNS SPECIALTY

THE WINNER!

WERE QUALIFIED

CMUSAT 2007-02-08, MXC 2007-02-08, MiraXT v3, Rsat 2007-02-08, SATzilla CRAFTED, minisat SAT-2007, picosat 535

And the winner is...

SATzilla CRAFTED

minisat SAT-2007

MXC 2007-02-08

slide-70
SLIDE 70

Rules Participants First Stage Winners Conclusion Certified UNSAT AIG

CATEGORY CRAFTED, SAT+UNS SPECIALTY

THE DETAILS!

Solver Score #SAT #Uns

SATzilla CRAFTED 74469 27 67 minisat SAT-2007 63371 26 72 MXC 2007-02-08 39848 20 57 MiraXT v3 34236 24 54 CMUSAT 2007-02-08 26461 21 45 Rsat 2007-02-08 19532 15 40 picosat 535 19081 22 38

slide-71
SLIDE 71

Rules Participants First Stage Winners Conclusion Certified UNSAT AIG

CATEGORY CRAFTED, SAT SPECIALTY

THE WINNER!

WERE QUALIFIED

CMUSAT 2007-02-08, MXC 2007-02-08, March-KS 2007-02-08, MiraXT v3, Rsat

2007-02-08, SATzilla CRAFTED, TTS 4.0, minisat SAT-2007, picosat 535

And the winner is...

slide-72
SLIDE 72

Rules Participants First Stage Winners Conclusion Certified UNSAT AIG

CATEGORY CRAFTED, SAT SPECIALTY

THE WINNER!

WERE QUALIFIED

CMUSAT 2007-02-08, MXC 2007-02-08, March-KS 2007-02-08, MiraXT v3, Rsat

2007-02-08, SATzilla CRAFTED, TTS 4.0, minisat SAT-2007, picosat 535

And the winner is...

March-KS 2007-02-08

slide-73
SLIDE 73

Rules Participants First Stage Winners Conclusion Certified UNSAT AIG

CATEGORY CRAFTED, SAT SPECIALTY

THE WINNER!

WERE QUALIFIED

CMUSAT 2007-02-08, MXC 2007-02-08, March-KS 2007-02-08, MiraXT v3, Rsat

2007-02-08, SATzilla CRAFTED, TTS 4.0, minisat SAT-2007, picosat 535

And the winner is...

March-KS 2007-02-08

SATzilla CRAFTED

slide-74
SLIDE 74

Rules Participants First Stage Winners Conclusion Certified UNSAT AIG

CATEGORY CRAFTED, SAT SPECIALTY

THE WINNER!

WERE QUALIFIED

CMUSAT 2007-02-08, MXC 2007-02-08, March-KS 2007-02-08, MiraXT v3, Rsat

2007-02-08, SATzilla CRAFTED, TTS 4.0, minisat SAT-2007, picosat 535

And the winner is...

March-KS 2007-02-08

SATzilla CRAFTED

minisat SAT-2007

slide-75
SLIDE 75

Rules Participants First Stage Winners Conclusion Certified UNSAT AIG

CATEGORY CRAFTED, SAT SPECIALTY

THE DETAILS!

Solver Score #SAT #Uns

March-KS 2007-02-08 16323 18 SATzilla CRAFTED 14275 27 minisat SAT-2007 13785 26 MiraXT v3 11601 24 CMUSAT 2007-02-08 8093 21 picosat 535 7153 22 MXC 2007-02-08 5136 20 Rsat 2007-02-08 4475 15 TTS 4.0 1155 3

slide-76
SLIDE 76

Rules Participants First Stage Winners Conclusion Certified UNSAT AIG

CATEGORY CRAFTED, UNSAT SPECIALTY

THE WINNER!

WERE QUALIFIED

CMUSAT 2007-02-08, MXC 2007-02-08, March-KS 2007-02-08, MiraXT v3, Rsat

2007-02-08, SATzilla CRAFTED, TTS 4.0, minisat SAT-2007, picosat 535

And the winner is...

slide-77
SLIDE 77

Rules Participants First Stage Winners Conclusion Certified UNSAT AIG

CATEGORY CRAFTED, UNSAT SPECIALTY

THE WINNER!

WERE QUALIFIED

CMUSAT 2007-02-08, MXC 2007-02-08, March-KS 2007-02-08, MiraXT v3, Rsat

2007-02-08, SATzilla CRAFTED, TTS 4.0, minisat SAT-2007, picosat 535

And the winner is...

SATzilla CRAFTED

slide-78
SLIDE 78

Rules Participants First Stage Winners Conclusion Certified UNSAT AIG

CATEGORY CRAFTED, UNSAT SPECIALTY

THE WINNER!

WERE QUALIFIED

CMUSAT 2007-02-08, MXC 2007-02-08, March-KS 2007-02-08, MiraXT v3, Rsat

2007-02-08, SATzilla CRAFTED, TTS 4.0, minisat SAT-2007, picosat 535

And the winner is...

SATzilla CRAFTED

TTS 4.0

slide-79
SLIDE 79

Rules Participants First Stage Winners Conclusion Certified UNSAT AIG

CATEGORY CRAFTED, UNSAT SPECIALTY

THE WINNER!

WERE QUALIFIED

CMUSAT 2007-02-08, MXC 2007-02-08, March-KS 2007-02-08, MiraXT v3, Rsat

2007-02-08, SATzilla CRAFTED, TTS 4.0, minisat SAT-2007, picosat 535

And the winner is...

SATzilla CRAFTED

TTS 4.0

minisat SAT-2007

slide-80
SLIDE 80

Rules Participants First Stage Winners Conclusion Certified UNSAT AIG

CATEGORY CRAFTED, UNSAT SPECIALTY

THE DETAILS!

Solver Score #SAT #Uns

SATzilla CRAFTED 39922 67 TTS 4.0 38950 39 minisat SAT-2007 38090 72 MXC 2007-02-08 26151 57 March-KS 2007-02-08 19684 29 MiraXT v3 18379 54 CMUSAT 2007-02-08 16437 45 Rsat 2007-02-08 13373 40 picosat 535 11011 38

slide-81
SLIDE 81

Rules Participants First Stage Winners Conclusion Certified UNSAT AIG

ALL SOLVERS ON CRAFTED BENCHMARKS

10 20 30 40 50 60 70 80 90 100 1000 2000 3000 4000 5000 6000

#Solved CPU−Time needed (s)

← TTS−4.0 ← March−KS−2007−02−08 ← Rsat−2007−02−08 ← picosat−535 ← CMUSAT−2007−02−08 ← MXC−2007−02−08 ← MiraXT−v3 ← SATzilla−CRAFTED ← minisat−SAT−2007 TTS−4.0 (42) March−KS−2007−02−08 (47) Rsat−2007−02−08 (55) picosat−535 (60) CMUSAT−2007−02−08 (66) MXC−2007−02−08 (77) MiraXT−v3 (78) SATzilla−CRAFTED (94) minisat−SAT−2007 (98)

slide-82
SLIDE 82

Rules Participants First Stage Winners Conclusion Certified UNSAT AIG

AND NOW...

The results for the random category...

slide-83
SLIDE 83

Rules Participants First Stage Winners Conclusion Certified UNSAT AIG

CATEGORY RANDOM, SAT+UNS SPECIALTY

THE WINNER!

WERE QUALIFIED

DEWSATZ-1A 2007-02-08, KCNFS 2004, MXC 2007-02-08, March-KS 2007-02-08, MiraXT v3, SAT7 2007-02-08, SATzilla RANDOM, minisat SAT-2007

And the winner is...

slide-84
SLIDE 84

Rules Participants First Stage Winners Conclusion Certified UNSAT AIG

CATEGORY RANDOM, SAT+UNS SPECIALTY

THE WINNER!

WERE QUALIFIED

DEWSATZ-1A 2007-02-08, KCNFS 2004, MXC 2007-02-08, March-KS 2007-02-08, MiraXT v3, SAT7 2007-02-08, SATzilla RANDOM, minisat SAT-2007

And the winner is...

SATzilla RANDOM

slide-85
SLIDE 85

Rules Participants First Stage Winners Conclusion Certified UNSAT AIG

CATEGORY RANDOM, SAT+UNS SPECIALTY

THE WINNER!

WERE QUALIFIED

DEWSATZ-1A 2007-02-08, KCNFS 2004, MXC 2007-02-08, March-KS 2007-02-08, MiraXT v3, SAT7 2007-02-08, SATzilla RANDOM, minisat SAT-2007

And the winner is...

SATzilla RANDOM

March-KS 2007-02-08

slide-86
SLIDE 86

Rules Participants First Stage Winners Conclusion Certified UNSAT AIG

CATEGORY RANDOM, SAT+UNS SPECIALTY

THE WINNER!

WERE QUALIFIED

DEWSATZ-1A 2007-02-08, KCNFS 2004, MXC 2007-02-08, March-KS 2007-02-08, MiraXT v3, SAT7 2007-02-08, SATzilla RANDOM, minisat SAT-2007

And the winner is...

SATzilla RANDOM

March-KS 2007-02-08

KCNFS 2004

slide-87
SLIDE 87

Rules Participants First Stage Winners Conclusion Certified UNSAT AIG

CATEGORY RANDOM, SAT+UNS SPECIALTY

THE DETAILS!

Solver Score #SAT #Uns

SATzilla RANDOM 189835 147 101 March-KS 2007-02-08 167430 146 111 KCNFS 2004 124280 130 107 minisat SAT-2007 36387 83 57 MXC 2007-02-08 35538 82 53 DEWSATZ-1A 2007-02-08 30754 73 48 SAT7 2007-02-08 30282 74 48 MiraXT v3 26491 69 37

slide-88
SLIDE 88

Rules Participants First Stage Winners Conclusion Certified UNSAT AIG

CATEGORY RANDOM, SAT SPECIALTY

THE WINNER!

WERE QUALIFIED

DEWSATZ-1A 2007-02-08, KCNFS 2004, MXC 2007-02-08, March-KS 2007-02-08, MiraXT v3, SAT7 2007-02-08, SATzilla RANDOM, adaptg2wsat+ 2007-02-08, adaptg2wsat0 2007-02-08, adaptnovelty 2007-02-08, gnovelty+ 2007-02-08, minisat SAT-2007, ranov 2007-02-08, sapsrt 2007-02-08

And the winner is...

slide-89
SLIDE 89

Rules Participants First Stage Winners Conclusion Certified UNSAT AIG

CATEGORY RANDOM, SAT SPECIALTY

THE WINNER!

WERE QUALIFIED

DEWSATZ-1A 2007-02-08, KCNFS 2004, MXC 2007-02-08, March-KS 2007-02-08, MiraXT v3, SAT7 2007-02-08, SATzilla RANDOM, adaptg2wsat+ 2007-02-08, adaptg2wsat0 2007-02-08, adaptnovelty 2007-02-08, gnovelty+ 2007-02-08, minisat SAT-2007, ranov 2007-02-08, sapsrt 2007-02-08

And the winner is...

gnovelty+ 2007-02-08

slide-90
SLIDE 90

Rules Participants First Stage Winners Conclusion Certified UNSAT AIG

CATEGORY RANDOM, SAT SPECIALTY

THE WINNER!

WERE QUALIFIED

DEWSATZ-1A 2007-02-08, KCNFS 2004, MXC 2007-02-08, March-KS 2007-02-08, MiraXT v3, SAT7 2007-02-08, SATzilla RANDOM, adaptg2wsat+ 2007-02-08, adaptg2wsat0 2007-02-08, adaptnovelty 2007-02-08, gnovelty+ 2007-02-08, minisat SAT-2007, ranov 2007-02-08, sapsrt 2007-02-08

And the winner is...

gnovelty+ 2007-02-08

adaptg2wsat0 2007-02-08

slide-91
SLIDE 91

Rules Participants First Stage Winners Conclusion Certified UNSAT AIG

CATEGORY RANDOM, SAT SPECIALTY

THE WINNER!

WERE QUALIFIED

DEWSATZ-1A 2007-02-08, KCNFS 2004, MXC 2007-02-08, March-KS 2007-02-08, MiraXT v3, SAT7 2007-02-08, SATzilla RANDOM, adaptg2wsat+ 2007-02-08, adaptg2wsat0 2007-02-08, adaptnovelty 2007-02-08, gnovelty+ 2007-02-08, minisat SAT-2007, ranov 2007-02-08, sapsrt 2007-02-08

And the winner is...

gnovelty+ 2007-02-08

adaptg2wsat0 2007-02-08

adaptg2wsat+ 2007-02-08

slide-92
SLIDE 92

Rules Participants First Stage Winners Conclusion Certified UNSAT AIG

CATEGORY RANDOM, SAT SPECIALTY

THE DETAILS!

Solver Score #SAT #Uns

gnovelty+ 2007-02-08 122500 242 adaptg2wsat0 2007-02-08 114109 248 adaptg2wsat+ 2007-02-08 112877 252 adaptnovelty 2007-02-08 96497 240 ranov 2007-02-08 86647 242 sapsrt 2007-02-08 68218 188 SATzilla RANDOM 29895 147 March-KS 2007-02-08 26721 146 KCNFS 2004 23786 130 MXC 2007-02-08 13308 82 minisat SAT-2007 12397 83 SAT7 2007-02-08 11594 74 DEWSATZ-1A 2007-02-08 10832 73 MiraXT v3 10614 69

slide-93
SLIDE 93

Rules Participants First Stage Winners Conclusion Certified UNSAT AIG

CATEGORY RANDOM, UNSAT SPECIALTY

THE WINNER!

WERE QUALIFIED

DEWSATZ-1A 2007-02-08, KCNFS 2004, MXC 2007-02-08, March-KS 2007-02-08, SAT7

2007-02-08, SATzilla RANDOM, minisat SAT-2007, MiraXT v3

And the winner is...

slide-94
SLIDE 94

Rules Participants First Stage Winners Conclusion Certified UNSAT AIG

CATEGORY RANDOM, UNSAT SPECIALTY

THE WINNER!

WERE QUALIFIED

DEWSATZ-1A 2007-02-08, KCNFS 2004, MXC 2007-02-08, March-KS 2007-02-08, SAT7

2007-02-08, SATzilla RANDOM, minisat SAT-2007, MiraXT v3

And the winner is...

March-KS 2007-02-08

slide-95
SLIDE 95

Rules Participants First Stage Winners Conclusion Certified UNSAT AIG

CATEGORY RANDOM, UNSAT SPECIALTY

THE WINNER!

WERE QUALIFIED

DEWSATZ-1A 2007-02-08, KCNFS 2004, MXC 2007-02-08, March-KS 2007-02-08, SAT7

2007-02-08, SATzilla RANDOM, minisat SAT-2007, MiraXT v3

And the winner is...

March-KS 2007-02-08

KCNFS 2004

slide-96
SLIDE 96

Rules Participants First Stage Winners Conclusion Certified UNSAT AIG

CATEGORY RANDOM, UNSAT SPECIALTY

THE WINNER!

WERE QUALIFIED

DEWSATZ-1A 2007-02-08, KCNFS 2004, MXC 2007-02-08, March-KS 2007-02-08, SAT7

2007-02-08, SATzilla RANDOM, minisat SAT-2007, MiraXT v3

And the winner is...

March-KS 2007-02-08

KCNFS 2004

SATzilla RANDOM

slide-97
SLIDE 97

Rules Participants First Stage Winners Conclusion Certified UNSAT AIG

CATEGORY RANDOM, UNSAT SPECIALTY

THE DETAILS!

Solver Score #SAT #Uns

March-KS 2007-02-08 88041 111 KCNFS 2004 73087 107 SATzilla RANDOM 61008 101 minisat SAT-2007 16870 57 MXC 2007-02-08 14997 53 DEWSATZ-1A 2007-02-08 12794 48 SAT7 2007-02-08 12628 48 MiraXT v3 9573.3 37 37

slide-98
SLIDE 98

Rules Participants First Stage Winners Conclusion Certified UNSAT AIG

ALL SOLVERS ON RANDOM BENCHMARKS

50 100 150 200 250 300 1000 2000 3000 4000 5000 6000

#Solved CPU−Time needed (s)

← MiraXT−v3 ← DEWSATZ−1A−2007−02−08 ← SAT7−2007−02−08 ← MXC−2007−02−08 ← minisat−SAT−2007 ← sapsrt−2007−02−08 ← KCNFS−2004 ← adaptnovelty−2007−02−08 ← gnovelty+−2007−02−08 ← ranov−2007−02−08 ← adaptg2wsat0−2007−02−08 ← SATzilla−RANDOM ← adaptg2wsat+−2007−02−08 ← March−KS−2007−02−08 MiraXT−v3 (106) DEWSATZ−1A−2007−02−08 (121) SAT7−2007−02−08 (122) MXC−2007−02−08 (135) minisat−SAT−2007 (140) sapsrt−2007−02−08 (188) KCNFS−2004 (237) adaptnovelty−2007−02−08 (240) gnovelty+−2007−02−08 (242) ranov−2007−02−08 (242) adaptg2wsat0−2007−02−08 (248) SATzilla−RANDOM (248) adaptg2wsat+−2007−02−08 (252) March−KS−2007−02−08 (257)

slide-99
SLIDE 99

Rules Participants First Stage Winners Conclusion Certified UNSAT AIG

AND NOW...

The results for the industrial category...

slide-100
SLIDE 100

Rules Participants First Stage Winners Conclusion Certified UNSAT AIG

CATEGORY INDUSTRIAL, SAT+UNS SPECIALTY

THE WINNER!

WERE QUALIFIED

CMUSAT 2007-02-08, MXC 2007-02-08, MiraXT v3, Rsat 2007-02-08, SAT7 2007-02-08, SATzilla CRAFTED, TiniSatELite 2007-02-08, minisat SAT-2007, picosat 535

And the winner is...

slide-101
SLIDE 101

Rules Participants First Stage Winners Conclusion Certified UNSAT AIG

CATEGORY INDUSTRIAL, SAT+UNS SPECIALTY

THE WINNER!

WERE QUALIFIED

CMUSAT 2007-02-08, MXC 2007-02-08, MiraXT v3, Rsat 2007-02-08, SAT7 2007-02-08, SATzilla CRAFTED, TiniSatELite 2007-02-08, minisat SAT-2007, picosat 535

And the winner is...

Rsat 2007-02-08

slide-102
SLIDE 102

Rules Participants First Stage Winners Conclusion Certified UNSAT AIG

CATEGORY INDUSTRIAL, SAT+UNS SPECIALTY

THE WINNER!

WERE QUALIFIED

CMUSAT 2007-02-08, MXC 2007-02-08, MiraXT v3, Rsat 2007-02-08, SAT7 2007-02-08, SATzilla CRAFTED, TiniSatELite 2007-02-08, minisat SAT-2007, picosat 535

And the winner is...

Rsat 2007-02-08

picosat 535

slide-103
SLIDE 103

Rules Participants First Stage Winners Conclusion Certified UNSAT AIG

CATEGORY INDUSTRIAL, SAT+UNS SPECIALTY

THE WINNER!

WERE QUALIFIED

CMUSAT 2007-02-08, MXC 2007-02-08, MiraXT v3, Rsat 2007-02-08, SAT7 2007-02-08, SATzilla CRAFTED, TiniSatELite 2007-02-08, minisat SAT-2007, picosat 535

And the winner is...

Rsat 2007-02-08

picosat 535

minisat SAT-2007

slide-104
SLIDE 104

Rules Participants First Stage Winners Conclusion Certified UNSAT AIG

CATEGORY INDUSTRIAL, SAT+UNS SPECIALTY

THE DETAILS!

Solver Score #SAT #Uns

Rsat 2007-02-08 55452 63 76 picosat 535 51639 72 67 minisat SAT-2007 50398 53 79 TiniSatELite 2007-02-08 49560 55 75 CMUSAT 2007-02-08 41197 46 77 MXC 2007-02-08 38004 47 69 MiraXT v3 37700 52 74 SATzilla CRAFTED 31439 47 67 SAT7 2007-02-08 27607 51 63

slide-105
SLIDE 105

Rules Participants First Stage Winners Conclusion Certified UNSAT AIG

CATEGORY INDUSTRIAL, SAT SPECIALTY

THE WINNER!

WERE QUALIFIED

CMUSAT 2007-02-08, MXC 2007-02-08, MiraXT v3, Rsat 2007-02-08, SAT7 2007-02-08, SATzilla CRAFTED, TiniSatELite 2007-02-08, minisat SAT-2007, picosat 535

And the winner is...

slide-106
SLIDE 106

Rules Participants First Stage Winners Conclusion Certified UNSAT AIG

CATEGORY INDUSTRIAL, SAT SPECIALTY

THE WINNER!

WERE QUALIFIED

CMUSAT 2007-02-08, MXC 2007-02-08, MiraXT v3, Rsat 2007-02-08, SAT7 2007-02-08, SATzilla CRAFTED, TiniSatELite 2007-02-08, minisat SAT-2007, picosat 535

And the winner is...

picosat 535

slide-107
SLIDE 107

Rules Participants First Stage Winners Conclusion Certified UNSAT AIG

CATEGORY INDUSTRIAL, SAT SPECIALTY

THE WINNER!

WERE QUALIFIED

CMUSAT 2007-02-08, MXC 2007-02-08, MiraXT v3, Rsat 2007-02-08, SAT7 2007-02-08, SATzilla CRAFTED, TiniSatELite 2007-02-08, minisat SAT-2007, picosat 535

And the winner is...

picosat 535

Rsat 2007-02-08

slide-108
SLIDE 108

Rules Participants First Stage Winners Conclusion Certified UNSAT AIG

CATEGORY INDUSTRIAL, SAT SPECIALTY

THE WINNER!

WERE QUALIFIED

CMUSAT 2007-02-08, MXC 2007-02-08, MiraXT v3, Rsat 2007-02-08, SAT7 2007-02-08, SATzilla CRAFTED, TiniSatELite 2007-02-08, minisat SAT-2007, picosat 535

And the winner is...

picosat 535

Rsat 2007-02-08

minisat SAT-2007

slide-109
SLIDE 109

Rules Participants First Stage Winners Conclusion Certified UNSAT AIG

CATEGORY INDUSTRIAL, SAT SPECIALTY

THE DETAILS!

Solver Score #SAT #Uns

picosat 535 33812 72 Rsat 2007-02-08 24850 63 minisat SAT-2007 20639 53 TiniSatELite 2007-02-08 19977 55 MiraXT v3 18288 52 CMUSAT 2007-02-08 17571 46 MXC 2007-02-08 17006 47 SATzilla CRAFTED 13491 47 SAT7 2007-02-08 12363 51

slide-110
SLIDE 110

Rules Participants First Stage Winners Conclusion Certified UNSAT AIG

CATEGORY INDUSTRIAL, UNSAT SPECIALTY

THE WINNER!

WERE QUALIFIED

CMUSAT 2007-02-08, MXC 2007-02-08, MiraXT v3, Rsat 2007-02-08, SAT7 2007-02-08, SATzilla CRAFTED, TiniSatELite 2007-02-08, minisat SAT-2007, picosat 535

And the winner is...

slide-111
SLIDE 111

Rules Participants First Stage Winners Conclusion Certified UNSAT AIG

CATEGORY INDUSTRIAL, UNSAT SPECIALTY

THE WINNER!

WERE QUALIFIED

CMUSAT 2007-02-08, MXC 2007-02-08, MiraXT v3, Rsat 2007-02-08, SAT7 2007-02-08, SATzilla CRAFTED, TiniSatELite 2007-02-08, minisat SAT-2007, picosat 535

And the winner is...

Rsat 2007-02-08

slide-112
SLIDE 112

Rules Participants First Stage Winners Conclusion Certified UNSAT AIG

CATEGORY INDUSTRIAL, UNSAT SPECIALTY

THE WINNER!

WERE QUALIFIED

CMUSAT 2007-02-08, MXC 2007-02-08, MiraXT v3, Rsat 2007-02-08, SAT7 2007-02-08, SATzilla CRAFTED, TiniSatELite 2007-02-08, minisat SAT-2007, picosat 535

And the winner is...

Rsat 2007-02-08

minisat SAT-2007

slide-113
SLIDE 113

Rules Participants First Stage Winners Conclusion Certified UNSAT AIG

CATEGORY INDUSTRIAL, UNSAT SPECIALTY

THE WINNER!

WERE QUALIFIED

CMUSAT 2007-02-08, MXC 2007-02-08, MiraXT v3, Rsat 2007-02-08, SAT7 2007-02-08, SATzilla CRAFTED, TiniSatELite 2007-02-08, minisat SAT-2007, picosat 535

And the winner is...

Rsat 2007-02-08

minisat SAT-2007

TiniSatELite 2007-02-08

slide-114
SLIDE 114

Rules Participants First Stage Winners Conclusion Certified UNSAT AIG

CATEGORY INDUSTRIAL, UNSAT SPECIALTY

THE DETAILS!

Solver Score #SAT #Uns

Rsat 2007-02-08 30193 76 minisat SAT-2007 29683 79 TiniSatELite 2007-02-08 29459 75 CMUSAT 2007-02-08 22146 77 MXC 2007-02-08 20922 69 MiraXT v3 19086 74 SATzilla CRAFTED 17873 67 picosat 535 17203 67 SAT7 2007-02-08 15430 63

slide-115
SLIDE 115

Rules Participants First Stage Winners Conclusion Certified UNSAT AIG

ALL SOLVERS ON INDUSTRIAL BENCHMARKS

20 40 60 80 100 120 140 2000 4000 6000 8000 10000 12000

#Solved CPU−Time needed (s)

← SAT7−2007−02−08 ← SATzilla−CRAFTED ← MXC−2007−02−08 ← Siege−V4 ← CMUSAT−2007−02−08 ← MiraXT−v3 ← TiniSatELite−2007−02−08 ← minisat−SAT−2007 ← picosat−535 ← Rsat−2007−02−08 SAT7−2007−02−08 (114) SATzilla−CRAFTED (114) MXC−2007−02−08 (116) Siege−V4 (122) CMUSAT−2007−02−08 (123) MiraXT−v3 (126) TiniSatELite−2007−02−08 (130) minisat−SAT−2007 (132) picosat−535 (139) Rsat−2007−02−08 (139)

slide-116
SLIDE 116

Rules Participants First Stage Winners Conclusion Certified UNSAT AIG

DETAILED RESULTS AVAILABLE ON THE WEB

Go to

http://www.cril.univ-artois.fr/SAT07

slide-117
SLIDE 117

Rules Participants First Stage Winners Conclusion Certified UNSAT AIG

ROAD MAP

1

History, rules and Goals

2

The participants

3

The First Stage

4

The Winners

5 CONCLUSION AND NEXT CONTESTS

Conclusion

6

Certified UNSAT Special Track

7

And Inverter Graph Special Track

slide-118
SLIDE 118

Rules Participants First Stage Winners Conclusion Certified UNSAT AIG

LET’S GIVE AUTHORS A WARM APPLAUSE

They deserve it.

slide-119
SLIDE 119

Rules Participants First Stage Winners Conclusion Certified UNSAT AIG

SUBJECT TO DISCUSSION

SCORING MECHANISM Some solvers were ranked before other solvers that solved more

  • instances. However, it guarantees that benchmarks have different

hardness. Solving a hard benchmark gives more credit than an easy one. One solution may be not to compute the score of benchmarks during the contest, but before the contest, based on previous solvers (like in CASC). Any suggestion welcome!

slide-120
SLIDE 120

Rules Participants First Stage Winners Conclusion Certified UNSAT AIG

SUBJECT TO DISCUSSION

SCORING MECHANISM Some solvers were ranked before other solvers that solved more

  • instances. However, it guarantees that benchmarks have different

hardness. Solving a hard benchmark gives more credit than an easy one. One solution may be not to compute the score of benchmarks during the contest, but before the contest, based on previous solvers (like in CASC). Any suggestion welcome! Other hot topic: Taking multi-core architecture into account for time measures. Real time vs CPU time?

slide-121
SLIDE 121

Rules Participants First Stage Winners Conclusion Certified UNSAT AIG

NEXT COMPETITION?

Next year, there will be a competitive event (race or competition). New organizers are (almost) ready :)

slide-122
SLIDE 122

Rules Participants First Stage Winners Conclusion Certified UNSAT AIG

NEXT COMPETITION?

Next year, there will be a competitive event (race or competition). New organizers are (almost) ready :)

slide-123
SLIDE 123

Rules Participants First Stage Winners Conclusion Certified UNSAT AIG

ROAD MAP

1

History, rules and Goals

2

The participants

3

The First Stage

4

The Winners

5

Conclusion and next contests

6 CERTIFIED UNSAT SPECIAL TRACK 7

And Inverter Graph Special Track

slide-124
SLIDE 124

Rules Participants First Stage Winners Conclusion Certified UNSAT AIG

THE RULES

The first part consisted of the well-studied Pigeon-Hole formulas because there is a known proof scheme that produces what are believed to be the shortest possible proofs (see “Theory” in the table), due to S. Cook in 1971, and rediscovered by A. Hakin about 1985, and again by A. Van Gelder about 2003. The formula for P pigeons has P(P − 1) variables and P(P2 − 2P + 3)/2 clauses. The “Theory” derivation has (P − 1)(P + 2)2P−3 clauses. Note that the tts solver produces proofs that grow close to the

  • ptimum ratio!

The second section consisted of 17 benchmarks from the regular track, industrial category, of the competition. Resources: 1 CPU hour at 2.6 GHz, 8 GB real memory, about 6 GB swap, and file size limit 36 GB (most cases) on a local disk.

slide-125
SLIDE 125

Rules Participants First Stage Winners Conclusion Certified UNSAT AIG

THE CONTESTANTS

Program Proof Format Authors booleforce %RES

  • A. Biere

picosat %RES

  • A. Biere

picosat %RUP

  • A. Biere

tts %RPT

  • I. Spence

zchaffSE07 %RES

  • Z. Fu, L. Zhang, (others?)

zchaff 0 %RES

  • A. Van Gelder, Z. Fu, L. Zhang, (others?)
slide-126
SLIDE 126

Rules Participants First Stage Winners Conclusion Certified UNSAT AIG

WHO ARE THE WINNERS?

It was a mixed bag. tts was the clear winner on pigeon-hole formulas, but did very poorly on industrial formulas. booleforce edged out zchaff on industrial formulas, but had a few programming problems, and did poorly on pigeon-hole formulas. zchaff had the best balance between the two sections. picosat with the RUP format produced the most proofs, but many RUP proofs could not be verified within the allocated time and disk space.

slide-127
SLIDE 127

Rules Participants First Stage Winners Conclusion Certified UNSAT AIG

ROAD MAP

1

History, rules and Goals

2

The participants

3

The First Stage

4

The Winners

5

Conclusion and next contests

6

Certified UNSAT Special Track

7 AND INVERTER GRAPH SPECIAL TRACK

slide-128
SLIDE 128

Rules Participants First Stage Winners Conclusion Certified UNSAT AIG

AND INVERTER GRAPHS SPECIAL TRACK

THE IDEA

In 2005, the special ”non clausal” track using the eDimacs format provided by Fahiem Bacchus and Toby Walsh received no submission. In 2006, Armin Biere proposed the And Inverter Graph format for ”structural/non clausal” input. Many translators from and to AIG are made available. In 2007, the first HardWare Model Checking Competition is

  • rganized in conjonction with CAV’07 using the AIGER format.

The idea is to experiment the AIG format for benchmarks in the industrial category that tend to be huge, with a structure lost during CNF translation.

slide-129
SLIDE 129

Rules Participants First Stage Winners Conclusion Certified UNSAT AIG

AND INVERTER GRAPHS SPECIAL TRACK

THE COMPETITORS

Only 3 submitters for that special track, for a total of 5 solvers.

AIG-CMUSAT from Himanshu Jain. Uses aigtocnf translator available

in AIGER distribution to feed CMUSAT SAT solver.

AIGSAT-SP from Himanshu Jain, C. Bartzis and E. Clarke. AIG to

NNF translation.

AIGSAT-WC from Himanshu Jain, C. Bartzis and E. Clarke. AIG to

NNF translation. BCP different from previous one. MINICIRC from Niklas E`

  • en. Uses a three stage preprocessing

from AIG to CNF to feed MiniSAT 2.0 SAT solver.

PICOAIGERSAT from Armin Biere. Uses a basic AIG to CNF

translation to feed picosat SAT solver. The benchmarks were provided by Armin Biere (available from AIGER web site) or translated from CNF benchmarks in the industrial category (IBM and Velev benchmarks too big to be translated with current translator developed by Armin Biere).

slide-130
SLIDE 130

Rules Participants First Stage Winners Conclusion Certified UNSAT AIG

AND INVERTER GRAPHS SPECIAL TRACK

THE RESULTS

Available in details from http://www.cril.univ-artois.fr/SAT07/ Need to be taken with care: no certificates/witnesses provided for SAT answers. Solvers received late: minor I/O and stability problems no fixed for some solvers (e.g. MiniCirc) on current results. Conclusion: Check by yourself the results. Check the competition website for additional information in the next few weeks. Embrace the AIGER format.