Author: bugman Date: Thu Mar 8 12:32:00 2012 New Revision: 15465 URL: http://svn.gna.org/viewcvs/relax?rev=15465&view=rev Log: Merged revisions 15459-15464 via svnmerge from svn+ssh://bugman@xxxxxxxxxxx/svn/relax/1.3 ........ r15459 | bugman | 2012-03-08 10:44:57 +0100 (Thu, 08 Mar 2012) | 5 lines Added a reference implementation to the multi-processor package. This is to demonstrate to users of the package how an implementation is created. ........ r15460 | bugman | 2012-03-08 11:26:19 +0100 (Thu, 08 Mar 2012) | 5 lines Documentation improvements to the multi-processor test implementation. This should make it very easy to see what is happening at all stages. ........ r15461 | bugman | 2012-03-08 12:15:15 +0100 (Thu, 08 Mar 2012) | 3 lines Added some documentation about master processors on Linux 2.6 eating 100% of one CPU core. ........ r15462 | bugman | 2012-03-08 12:16:01 +0100 (Thu, 08 Mar 2012) | 3 lines The profiling print out is now sorted by cumulative time for better debugging. ........ r15463 | bugman | 2012-03-08 12:18:42 +0100 (Thu, 08 Mar 2012) | 6 lines Updated the test implementation timing numbers. This is after using 'echo "1" > /proc/sys/kernel/sched_compat_yield' to fix the Linux OpenMPI problem. ........ r15464 | bugman | 2012-03-08 12:21:13 +0100 (Thu, 08 Mar 2012) | 5 lines Copied the OpenMPI + Linux 2.6 100% master problem to the API documentation. The text has also been updated a little. ........ Added: branches/frame_order_testing/multi/test_implementation.py - copied unchanged from r15464, 1.3/multi/test_implementation.py Modified: branches/frame_order_testing/ (props changed) branches/frame_order_testing/multi/__init__.py Propchange: branches/frame_order_testing/ ------------------------------------------------------------------------------ --- svnmerge-integrated (original) +++ svnmerge-integrated Thu Mar 8 12:32:00 2012 @@ -1,1 +1,1 @@ -/1.3:1-15457 +/1.3:1-15464 Modified: branches/frame_order_testing/multi/__init__.py URL: http://svn.gna.org/viewcvs/relax/branches/frame_order_testing/multi/__init__.py?rev=15465&r1=15464&r2=15465&view=diff ============================================================================== --- branches/frame_order_testing/multi/__init__.py (original) +++ branches/frame_order_testing/multi/__init__.py Thu Mar 8 12:32:00 2012 @@ -100,6 +100,22 @@ To execute the calculations, the final part of the calculation code on the master must feature a call to:: processor_box.processor.run_queue(). + + +4 Example +========= + +See the script 'test_implementation.py' for a basic example of a reference, and full, implementation of the multi-processor package. + + +5 Issues +======== + +For multi-core systems and Linux 2.6, the following might be required to prevent the master processor from taking 100% of one CPU core while waiting for the slaves: + +# echo "1" > /proc/sys/kernel/sched_compat_yield + +This appears to be an OpenMPI problem with late 2.6 Linux kernels. """