relax SVN archive

The specific_analyses.relax_disp.parameters.loop_parameters() function can now run on a slave.

This is to fix a problem with r21250.

Shifted the optimisation printouts for the dispersion analysis out of the memo.

This improved the ordering of the printed out messages when running on a cluster. Instead of having
multiple optimisation printouts followed by a list of the corresponding optimised values, now they
are interleaved as they should be.

Fix for commit r21247.

Missing import.

Proper handling of the back-calculated dispersion data for the new 'MMQ 2-site' model.

Merged revisions 21245 via svnmerge from
svn+ssh://bugman@.../svn/relax/trunk

........
r21245 | bugman | 2013-10-27 14:39:57 +0100 (Sun, 27 Oct 2013) | 6 lines

Fix for bug #21233 (https://gna.org/bugs/?21233) - the missing mpi4py multi-processor messages.

When multiple commands were being sent to one slave, the captured IO was being overwritten by each
executed command. Therefore the slave would only return the printouts from the last command.
........

Fix for bug #21233 (https://gna.org/bugs/?21233) - the missing mpi4py multi-processor messages.

When multiple commands were being sent to one slave, the captured IO was being overwritten by each
executed command. Therefore the slave would only return the printouts from the last command.

Added the cluster graphic to the cluster section of the dispersion chapter and improved the text.

Added a 600x600 pixel graphic for the spin cluster for use in the user manual.

Removed most of the \clearpage commands in the dispersion chapter of the manual.

There was far too much whitespace.