Re: [OMPI devel] Infinite Loop: ompi_free_list_wait

2009-03-25 Thread Jeff Squyres
George -- correct me if I'm wrong -- we went through and audited ob1 and the relevant BTLs. The only places where WAIT remains are places that are guaranteed to not be problematic. So you shouldn't need to edit ob1 at all. If you're working with Open MPI, you might want to investigate using

[OMPI devel] Error in the versions 1.3 and 1.3.1 of OpenMPI when using SLURM_OVERCOMMIT=1

2009-03-25 Thread Hartmut Häfner
Dear Support, there is a problem with OpenMPI in version 1.3 and version 1.3.1 when using our batch system Slurm. On our parallel computer there are 2 queues - one with exclusive usage of slots (cores) (SLURM_OVERCOMMIT=0) within nodes and one without shared usage of slots (SLURM_OVERCOMMIT=1)

Re: [OMPI devel] Error in the versions 1.3 and 1.3.1 of OpenMPI when using SLURM_OVERCOMMIT=1

2009-03-25 Thread Ralph Castain
The answer is simple: the SLURM environmental variables when you set SLURM_OVERCOMMIT=1 are telling us that only one slot/node is available for your use. This is done by the SLURM_TASKS_PER_NODE envar. So we can only launch 1 proc/node as this is all SLURM is allowing us to do. Ralph On

Re: [OMPI devel] OMPI 1.3 - PERUSE peruse_comm_spec_t peer Negative Value

2009-03-25 Thread Kiril Dichev
Hi, at least for the specific test program I used, the negative values for the peer attribute disappeared after George's modifications in 20844. One remark: after installation, I had to remove the '#include "ompi_config.h"' line in the "include/peruse.h" header to get PERUSE applications to comp

Re: [OMPI devel] OMPI 1.3 - PERUSE peruse_comm_spec_t peer Negative Value

2009-03-25 Thread George Bosilca
I did a quick check and I think there is no other possibility to get the peer equal to -1. Well, at least in theory. Please let me know if you find any other problems with our PERUSE interface. Meanwhile, I will take a look at the peruse.h header. Thanks, george. On Mar 25, 2009, at 1

Re: [OMPI devel] OMPI 1.3 - PERUSE peruse_comm_spec_t peer Negative Value

2009-03-25 Thread Samuel K. Gutierrez
Hi All, George - I really appreciate the quick response. > Hi, > > at least for the specific test program I used, the negative values for the peer attribute disappeared after George's modifications in 20844. Same here for my profiling library - tested with openmpi-1.3.2a1r20855. > > One remark:

Re: [OMPI devel] OMPI 1.3 - PERUSE peruse_comm_spec_t peer Negative Value

2009-03-25 Thread George Bosilca
This was the result of an unfortunate commit. Starting from revision r20884 the header file ompi_config.h is not included anymore. Thanks for the report. george. On Mar 25, 2009, at 13:29 , Kiril Dichev wrote: One remark: after installation, I had to remove the '#include "ompi_config.h"'

[OMPI devel] MCA component dependency

2009-03-25 Thread Aurélien Bouteiller
Hi everyone, I'm trying to state that a particular component depends on another that should therefore be dlopened automatically when it is loaded. I found some code doing exactly that in mca_base_component_find:open_component, but can't find any example of how to actually trigger that cod