Lennard-Jones 38 global minimum

SystemLennard-Jones
Atoms38
Runs100
Cap—1

1 The problem file does not state a cap.

The main metric is Force calls. Scoring rules.

Charts

Success rate on Lennard-Jones 38 global minimum, in rank order. The rate is the share of runs that finished. 5 ranked rows did not record a success rate and are not drawn.
Success rate on Lennard-Jones 38 global minimum, in rank order. The rate is the share of runs that finished. 5 ranked rows did not record a success rate and are not drawn. PDF.
Dolan-Moré profile for Basin hopping / global on Lennard-Jones 38 global minimum. The curve is the fraction of starts whose cost is at most tau times the best finite cost in this family. Tau is 1, 2, 4, 8, and 16. An unsolved start stays in the denominator and does not cross a finite tau. Dolan, E. D. and Moré, J. J., Math. Program. 91:201-213, 2002.
Dolan-Moré profile for Basin hopping / global on Lennard-Jones 38 global minimum. The curve is the fraction of starts whose cost is at most tau times the best finite cost in this family. Tau is 1, 2, 4, 8, and 16. An unsolved start stays in the denominator and does not cross a finite tau. Dolan, E. D. and Moré, J. J., Math. Program. 91:201-213, 2002. PDF.
Force calls on each recorded start for Lennard-Jones 38 global minimum. The box is the quartiles of those counts. A blue point is a success and a vermillion cross is a recorded failure (9). The horizontal position is the recorded count. The vertical offset only keeps marks from covering one another.
Force calls on each recorded start for Lennard-Jones 38 global minimum. The box is the quartiles of those counts. A blue point is a success and a vermillion cross is a recorded failure (9). The horizontal position is the recorded count. The vertical offset only keeps marks from covering one another. PDF.
Performance profile for Lennard-Jones 38 global minimum00.51124816Cost ratioFraction of startsGMIN (df8da599) · basin-hopping with continuous symmetry measuresAIRSS (0.9.4) · random structure searchGMIN (df8da599) · basin-hoppingGMIN (df8da599) · parallel tempering basin-hoppingeOn (3.5.0) · Basin hoppingEon svn (r2549) · Basin hopping
Performance profile. Higher means more starts solved within that cost ratio.
  • GMIN (df8da599) · basin-hopping with continuous symmetry measures
  • AIRSS (0.9.4) · random structure search
  • GMIN (df8da599) · basin-hopping
  • GMIN (df8da599) · parallel tempering basin-hopping
  • eOn (3.5.0) · Basin hopping
  • Eon svn (r2549) · Basin hopping
Cost distribution for Lennard-Jones 38 global minimum02465883.83 to 15190.1: 215190.1 to 39216: 139216 to 101243: 0101243 to 261376: 0261376 to 674789: 75883.83674789
Cost distribution of 10 ranked rows. Logarithmic bins, from 5883.83 to 674789. 1 ranked rows have no cost on this scale and are left off the chart.

Performance profile SVG Cost and success SVG

Performance profile for Lennard-Jones 38 global minimum
Performance profile. Select the plot for the full SVG.
Cost and success for Lennard-Jones 38 global minimum
Cost and success. Select the plot for the full SVG.
Best cost for Lennard-Jones 38 global minimumLennard-Jones 38 global minimum10³10⁴10⁵10⁶cost (force calls, log scale)GMIN: 5883.83GMIN5883.83eOn: 503146eOn503146Eon svn: 508011Eon svn508011pele: 521595pele521595
A shorter bar is a lower cost on a log scale shared by the codes of this problem.

Leaderboard

A shorter bar is a lower cost on the log scale for that problem. Select a column heading to sort.
RankCodeBestMedianSuccessPAR-2
1 GMIN 5883.83 144331 100.0% 5883.83
2 eOn 503146 503146 100.0% 503146
3 Eon svn 508011 553271 —1 —2
4 pele 521595 521595 100.0% —2

1 A dash in the success rate means that row did not record one.

2 A dash in PAR-2 means that row has no PAR-2 score.

The tables below list every ranked row, by method family and overall.

These archived rows rank on the value each one published, which leaves out any failure. The failure rule has no score for them, for the reason given with each.

  • eOn (r2156) · eon-basinhopping (the entry records no failure count)

Connect / double-ended

Connect / double-ended. One row is one code, one version, and one method. A lower cost is better. PAR-2 counts a finished run as the calls it used and a failed run as twice the cap. An archived row shows the average it published, and its PAR-2 applies that rule to the failures the entry records. Hessians is the mean analytic Hessian count. FD calls is that count times one force call per coordinate.
RankRowCostUnitHessiansFD callsRuns
1pele · hybrid eigenvector following521595force calls00100

Basin hopping / global

Basin hopping / global. One row is one code, one version, and one method. A lower cost is better. PAR-2 counts a finished run as the calls it used and a failed run as twice the cap. An archived row shows the average it published, and its PAR-2 applies that rule to the failures the entry records. Hessians is the mean analytic Hessian count. FD calls is that count times one force call per coordinate.
RankRowCostUnitHessiansFD callsRuns
1GMIN (df8da599) · basin-hopping with continuous symmetry measures5883.83force calls00100
2GMIN · basin-hopping with symmetrise csm moves6421force calls00100
3GMIN · basin-hopping with symmetrise moves22711force calls00100
4GMIN · basin-hopping265951force calls00100
5GMIN (df8da599) · basin-hopping432092force calls00100
6eOn (3.5.0) · Basin hopping503146force calls00100
7eOn (r2156) · eon-basinhopping508011force calls00100
8Eon svn (r2549) · Basin hopping598530force calls00100
9GMIN (df8da599) · parallel tempering basin-hopping674789force calls00100
10AIRSS (0.9.4) · random structure searchforce calls00100

Overall

One row is one code, one version, and one method. A lower cost is better. PAR-2 counts a finished run as the calls it used and a failed run as twice the cap. An archived row shows the average it published, and its PAR-2 applies that rule to the failures the entry records. Hessians is the mean analytic Hessian count. FD calls is that count times one force call per coordinate.
RowFamilyFamily rankUnitPAR-2PublishedMedian callsMean callsHessiansFD callsFailuresRuns
GMIN (df8da599) · basin-hopping with continuous symmetry measuresBasin hopping / global1force calls5883.833980.55883.8300crash 0, divergence 0, cap 0, wrong_stationary 0100
GMIN · basin-hopping with symmetrise csm movesBasin hopping / global2force calls642147716421000 failed100
GMIN · basin-hopping with symmetrise movesBasin hopping / global3force calls227111599822711000 failed100
GMIN · basin-hoppingBasin hopping / global4force calls265951202167265951000 failed100
GMIN (df8da599) · basin-hoppingBasin hopping / global5force calls43209230196241209200crash 0, divergence 0, cap 1, wrong_stationary 0100
eOn (3.5.0) · Basin hoppingBasin hopping / global6force calls50314636902850314600crash 0, divergence 0, cap 0, wrong_stationary 0100
eOn (r2156) · eon-basinhoppingBasin hopping / global7force calls50801150801100not recorded100
pele · hybrid eigenvector followingConnect / double-ended1force calls521595320504521595000 failed100
Eon svn (r2549) · Basin hoppingBasin hopping / global8force calls59853042322655853000crash 0, divergence 0, cap 2, wrong_stationary 0100
GMIN (df8da599) · parallel tempering basin-hoppingBasin hopping / global9force calls67478930729055478900crash 0, divergence 0, cap 6, wrong_stationary 0100
AIRSS (0.9.4) · random structure searchBasin hopping / global10force calls70635.511213400crash 0, divergence 0, cap 0, wrong_stationary 0100