Help with CLOP configuration

[dylhunn]

I am trying to run Remi Coulom's CLOP to optimize my evaluation constants.

However, I can't get it to work. This is my experiment log output:

Code: Select all

2017-05-22 17:48:46 ===== Starting: DragonExperiment =====
2017-05-22 17:48:46 Reading DragonExperiment-old.dat...
2017-05-22 17:48:46 Backup done.
2017-05-22 17:48:46 Samples = 1; Results = 0; Mismatches = 0.
2017-05-22 17:48:46 Run
2017-05-22 17:48:46 id = 0 (cpu1); Seed = 0; Starting from setInProgress
2017-05-22 17:48:46 id = 0 (cpu1); Seed = 0; Error: execvp: Permission denied
2017-05-22 17:48:46 id = 0 stopped. RunningProcessors = 0
2017-05-22 17:48:46 Stop

And this is my experiment.clop file:

Code: Select all

# Name (used for .log and .dat files)
Name DragonExperiment

# Script for running a game. See DummyScript.py for details.
Script ./cutechess-clop.py

# Parameter(s) to be optimized
# <parameter_type> <name> <min> <max>
# <parameter_type> may be:
#  LinearParameter
#  IntegerParameter
#  GammaParameter
#  IntegerGammaParameter
# For GammaParameter, quadratic regression is performed on log(x)
# Warning: 123 and not 123.0 should be used for IntegerParameter
IntegerParameter BishopPairBonus  0 100
IntegerParameter DiagonalMobilityBonus  0 50
IntegerParameter OrthogonalMobilityBonus  0 50
IntegerParameter DoubledPawnPenalty  0 100
IntegerParameter PassedPawnBonus  0 100
IntegerParameter IsolatedPawnPenalty  0 100

# This could be the list of machine names in a distributed experiment.
# In order to run 4 games in parallel, 2 on machine1, 2 on machine2: 
Processor cpu1
#Processor machine2
#Processor machine2

# Call the script "Replications" times in a row with identical parameter values
# Replications may be used to alternate colors, for instance.
# Seed % Replications would indicate color.
Replications 2

# Parameters of statistical model of outcome
# For binary outcome (Win/Loss, no draws), use "DrawElo 0"
# For chess, use "DrawElo 100"
DrawElo 100

# Regression parameters
# H 3 is recommended (it is the default value)
# Correlations may be "all" (default) or "none"
# Even if variables are not correlated "all" should work well. The problem is
# that the regression might become very costly if the number of variables is
# high. So use "Correlations none" only if you are certain parameters are
# independent or you have so many variables that "all" is too costly.
H 3
Correlations all

What should I put on the processor line to get this to work? I'm running it on Linux.

Thanks in advance!

[dylhunn]

I got this working, so I'm posting an update for future readers. The following changes were necessary:

• I modified the cutechess source code to actually print out the result of each game (QT Debug printing didn't really work on my Fedora system)
• Be sure to use the included .sh script to invoke cutechess, DON'T directly invoke the executable
• Be sure your execution permissions are set (ie chmod 777 *)