What steps will reproduce the problem? 1. run analysis using automated search() 2. issue select(unique, optimal) command. This is supposed to be default, but a select() did not seem to work either 3. final reported trees had 8 trees, 7 of which were suboptimal

What is the expected output? What do you see instead? One optimal tree

What version of the product are you using? On what operating system?

POY 5.0.0 Black Sabbath Development build 4900d3888e9d+ compiled with OCAML 4.00.1, compiled for this system (not from the repositories) mpirun.mpich2 HYDRA: Version: 1.4.1 (2011 version from the repositories)

LSB Version: ...core-4.0-amd64:core-4.0-noarch Description: Ubuntu 12.04.2 LTS Release: 12.04 Codename: precise Dual Opteron, 2x4cores

Please provide any additional information below. command file: http://db.tt/N8lEBHMN final tree file: http://dl.dropbox.com/u/31824503/run6_noexc/run6.tre pdf of tree: http://dl.dropbox.com/u/31824503/run6_noexc/run6_tre.pdf other files are in the public dropbox directory, but not the data files can put them up on request

Incidently: POY4 does not compile under OCAML 4 because the config script doesn't recognise this version and claims it is old. Compiling POY5 ran without problems, although I didn't --enable-long-sequences

I forgot to mention that report("run6_tre.ps",graphtrees:collapse) does not report all trees in the file, only the last one, and that report("run6_con.ps",graphconsensus) does not make a consensus of the optimal and suboptimal trees in the file.

this has been resolved: problem caused by mpi compiler and mpirun binaries not matching (I think). See POY mail list.