On Jan 30, 2007, at 9:39 AM, Ralph H Castain wrote:
On 1/30/07 9:24 AM, "Greg Watson" <gwat...@lanl.gov> wrote:
Yes, we need the hostfile information before job execution. We call
setup_job() before a debug job to request a process allocation for
the application being debugged. We use spawn() to launch a non-debug
application. It sounds like I should just leave things the way they
currently are.
I think we've had the discussion about bproc before, but the reason
we still support 1.0.2 is that the registry *is* populated with node
information prior to launch. This was an agreed on feature that
OpenMPI was to provide for PTP. I haven't been able to test 1.2 on a
bproc machine (since I can't get it to work), but it sounds like the
changes removed this functionality. Frankly, this makes OpenMPI less
attractive to us, since we now have to go and get this information
ourselves. My thinking now is that in the future we probably won't
use OpenMPI for anything other than building and launching the
application.
Decisions such as that, of course, are up to you. Meantime, take a
gander at
the data in ORTE_BPROC_NODE_SEGMENT within the registry. I tried to
preserve
some of what was being done, even though the method used to
populate the
bproc data was problematic and not really correct. You may find
that the
info stuck in there meets your needs for the GUI.
My point, though, is that only bproc and hostfile would be
supported under
the best of current conditions, and we only get that by
circumscribing the
functional intent of several key frameworks. The general ability
(across all
systems) to obtain the node info prior to launch isn't built into
the code
at the moment, but is planned for ORTE 2.0 (and was built in a
separate
prototype branch). For reasons totally beyond my control, the
prototype ORTE
2.0 code has *not* been incorporated into Open MPI yet.
Sorry...I like that no more than you... :-/
I suppose it is our decision in the sense that we could decide not to
provide the functionality and hope that it is implemented in OpenMPI
sometime in the future. If we have to implement something ourselves
to provide this functionality, then we may as well do it in a generic
way that will work with any runtime systems.
Greg