Dear Gary, I've finally managed to work out what was happening in the multi-processor branch with the loss of control of STDOUT and STDERR when errors occur. This now allows the test suite and user controlled IO redirection to function properly. This was one of the major hurdles for getting the branch merged. I've tested this for the uni-processor and it works. I'm now testing this out with mpi4py (using openmpi) on a dual core system to see if the code works. Would you be able to test this code by running your script and data from https://mail.gna.org/public/relax-devel/2007-04/msg00048.html? Would it be possible to attach this data to a task at https://gna.org/task/?group=relax so I can more easily get this code running again and merged? Cheers. Now only two more things remain prior to merging the branch back. The first is the breaking up of the model-free grid search within relax (and not the optimisation code as this is now external in the minfx package http://gna.org/projects/minfx/). This grid search is currently broken. The second is a good code clean up. There is old code that needs pruning, 44 FIXME comments, 29 TODO comments, 5 CHECKME comments, lots of commented out code, and a requirement to convert the code to the relax standards. If you could help a little with these, it would be much appreciated. I hope to do this one day, not too far off in the future. Do you still have plans of publishing this? For reference for those unaware of Gary's multi-processor branch, please look at the following two key references: https://mail.gna.org/public/relax-devel/2007-05/msg00000.html https://mail.gna.org/public/relax-devel/2007-04/msg00048.html Everything started with this post: https://mail.gna.org/public/relax-devel/2006-04/msg00023.html The branch is located at: svn co svn://svn.gna.org/svn/relax/branches/multi_processor_merge This is the updated, current code for the 1.3 line. The original multi-processor branch using the 1.2 relax designs (branched from around the relax-1.3.0 stage) is located at: svn co svn://svn.gna.org/svn/relax/branches/multi_processor Regards, Edward