November 2010 - mpiwg-ft - mailman3.cels.anl.gov

Re: [Mpi3-ft] SC10 Informal Meeting
by George R Carr Jr 18 Nov '10

18 Nov '10

At 5:33 PM -0600 11/16/10, Joshua Hursey wrote: >We are meeting at 12:30 on Thursday 11/18 at the SC10 Information booth. > >We are scheduled to meet for an hour, and will likely just find some >chairs in the general area. We will try to keep it small. For those >that cannot attend, I'll report back the the group on anything that >we discuss that may be of interest. > >-- Josh > I've attached an extract from my project document (in progress). -- George R Carr Jr U.S. Department of Commerce NOAA/OAR/GSD Tel: 303-497-4714 Email: George.Carr(a)noaa.gov

1 0

Re: [Mpi3-ft] SC10 Informal Meeting
by Joshua Hursey 16 Nov '10

16 Nov '10

We are meeting at 12:30 on Thursday 11/18 at the SC10 Information booth. We are scheduled to meet for an hour, and will likely just find some chairs in the general area. We will try to keep it small. For those that cannot attend, I'll report back the the group on anything that we discuss that may be of interest. -- Josh On Nov 13, 2010, at 5:26 PM, Joshua Hursey wrote: > Reminder. Please fill out the poll by Sunday morning, so we can set a time to meet. > > http://www.doodle.com/fhru658inp9v9f6c > > -- Josh > > On Nov 12, 2010, at 2:16 PM, Joshua Hursey wrote: > >> Given the complex schedules each of us tend to keep at SC, we will likely only be able to gather a few members of the group at any given time. I started a doodle poll with a few times that might work to meet. Anyone that would like to attend this meeting, fill this out by Sunday morning so we can set a time. >> >> http://www.doodle.com/fhru658inp9v9f6c >> >> A quick glance shows that probably Sunday, Monday before the Gala, or Thursday afternoon probably work best for me. >> >> Best, >> Josh >> >> >> On Nov 10, 2010, at 4:49 PM, George R Carr Jr wrote: >> >>> Per our conf call from two weeks ago, I have some usage cases to >>> discuss. I would like to discover if your test bed might be enough >>> for the simplest case or two. I'm adding/correcting a writeup that I >>> will be sending out soon. >>> >>> There was a question on when the group might be meeting during SC10. >>> Please let me know so that I might be there. >>> >>> Regards, >>> George >>> -- >>> George R Carr Jr >>> U.S. Department of Commerce >>> NOAA/OAR/GSD >>> Tel: 303-497-4714 >>> Email: George.Carr(a)noaa.gov >>> _______________________________________________ >>> mpi3-ft mailing list >>> mpi3-ft(a)lists.mpi-forum.org >>> http://lists.mpi-forum.org/mailman/listinfo.cgi/mpi3-ft >>> >> >> ------------------------------------ >> Joshua Hursey >> Postdoctoral Research Associate >> Oak Ridge National Laboratory >> http://users.nccs.gov/~jjhursey >> >> >> _______________________________________________ >> mpi3-ft mailing list >> mpi3-ft(a)lists.mpi-forum.org >> http://lists.mpi-forum.org/mailman/listinfo.cgi/mpi3-ft >> > > ------------------------------------ > Joshua Hursey > Postdoctoral Research Associate > Oak Ridge National Laboratory > http://users.nccs.gov/~jjhursey > > > _______________________________________________ > mpi3-ft mailing list > mpi3-ft(a)lists.mpi-forum.org > http://lists.mpi-forum.org/mailman/listinfo.cgi/mpi3-ft > ------------------------------------ Joshua Hursey Postdoctoral Research Associate Oak Ridge National Laboratory http://users.nccs.gov/~jjhursey

1 0

Re: [Mpi3-ft] SC10 Informal Meeting
by Joshua Hursey 13 Nov '10

13 Nov '10

Reminder. Please fill out the poll by Sunday morning, so we can set a time to meet. http://www.doodle.com/fhru658inp9v9f6c -- Josh On Nov 12, 2010, at 2:16 PM, Joshua Hursey wrote: > Given the complex schedules each of us tend to keep at SC, we will likely only be able to gather a few members of the group at any given time. I started a doodle poll with a few times that might work to meet. Anyone that would like to attend this meeting, fill this out by Sunday morning so we can set a time. > > http://www.doodle.com/fhru658inp9v9f6c > > A quick glance shows that probably Sunday, Monday before the Gala, or Thursday afternoon probably work best for me. > > Best, > Josh > > > On Nov 10, 2010, at 4:49 PM, George R Carr Jr wrote: > >> Per our conf call from two weeks ago, I have some usage cases to >> discuss. I would like to discover if your test bed might be enough >> for the simplest case or two. I'm adding/correcting a writeup that I >> will be sending out soon. >> >> There was a question on when the group might be meeting during SC10. >> Please let me know so that I might be there. >> >> Regards, >> George >> -- >> George R Carr Jr >> U.S. Department of Commerce >> NOAA/OAR/GSD >> Tel: 303-497-4714 >> Email: George.Carr(a)noaa.gov >> _______________________________________________ >> mpi3-ft mailing list >> mpi3-ft(a)lists.mpi-forum.org >> http://lists.mpi-forum.org/mailman/listinfo.cgi/mpi3-ft >> > > ------------------------------------ > Joshua Hursey > Postdoctoral Research Associate > Oak Ridge National Laboratory > http://users.nccs.gov/~jjhursey > > > _______________________________________________ > mpi3-ft mailing list > mpi3-ft(a)lists.mpi-forum.org > http://lists.mpi-forum.org/mailman/listinfo.cgi/mpi3-ft > ------------------------------------ Joshua Hursey Postdoctoral Research Associate Oak Ridge National Laboratory http://users.nccs.gov/~jjhursey

1 0

[Mpi3-ft] SC10 Informal Meeting
by Joshua Hursey 12 Nov '10

12 Nov '10

Given the complex schedules each of us tend to keep at SC, we will likely only be able to gather a few members of the group at any given time. I started a doodle poll with a few times that might work to meet. Anyone that would like to attend this meeting, fill this out by Sunday morning so we can set a time. http://www.doodle.com/fhru658inp9v9f6c A quick glance shows that probably Sunday, Monday before the Gala, or Thursday afternoon probably work best for me. Best, Josh On Nov 10, 2010, at 4:49 PM, George R Carr Jr wrote: > Per our conf call from two weeks ago, I have some usage cases to > discuss. I would like to discover if your test bed might be enough > for the simplest case or two. I'm adding/correcting a writeup that I > will be sending out soon. > > There was a question on when the group might be meeting during SC10. > Please let me know so that I might be there. > > Regards, > George > -- > George R Carr Jr > U.S. Department of Commerce > NOAA/OAR/GSD > Tel: 303-497-4714 > Email: George.Carr(a)noaa.gov > _______________________________________________ > mpi3-ft mailing list > mpi3-ft(a)lists.mpi-forum.org > http://lists.mpi-forum.org/mailman/listinfo.cgi/mpi3-ft > ------------------------------------ Joshua Hursey Postdoctoral Research Associate Oak Ridge National Laboratory http://users.nccs.gov/~jjhursey

1 0

Re: [Mpi3-ft] FT Working Group Teleconf - Canceled
by George R Carr Jr 10 Nov '10

10 Nov '10

>We were scheduled to have an FT working group teleconf today. >However, I am out sick and would like to cancel it today. I think >most people could probably use the time back since it is the week >before SC. > >Our next meeting will be in CA at the MPI Forum meeting. > >-- Josh > Per our conf call from two weeks ago, I have some usage cases to discuss. I would like to discover if your test bed might be enough for the simplest case or two. I'm adding/correcting a writeup that I will be sending out soon. There was a question on when the group might be meeting during SC10. Please let me know so that I might be there. Regards, George -- George R Carr Jr U.S. Department of Commerce NOAA/OAR/GSD Tel: 303-497-4714 Email: George.Carr(a)noaa.gov

1 0

[Mpi3-ft] FT Working Group Teleconf - Canceled
by Joshua Hursey 10 Nov '10

10 Nov '10

We were scheduled to have an FT working group teleconf today. However, I am out sick and would like to cancel it today. I think most people could probably use the time back since it is the week before SC. Our next meeting will be in CA at the MPI Forum meeting. -- Josh

1 0

[Mpi3-ft] Proposal Update: MPI_ANY_SOURCE, MPI_Abort, MPI_Kill
by Joshua Hursey 02 Nov '10

02 Nov '10

I updated the run-through stabilization proposal with some of the issues that were brought up during the past two meetings (MPI Forum meeting in Chicago, and teleconf last week). Below is a high-level list of the changes, and below a few topics to discuss: * Added definition of 'fail-stop' and 'transient' process failure * Cleaned up the failure detection text to make it more precise. * Started working on the text clarification for MPI_ANY_SOURCE (see first paragraph of section on 'Ch.3') * Touched up the MPI_Gather example * Added note on MPI_Abort for further discussion * Added start of an interface for MPI_Kill Proposal is here: https://svn.mpi-forum.org/trac/mpi-forum-web/wiki/ft/run_through_stabilizat… MPI_ANY_SOURCE -------------- I made a first attempt at some language for the MPI_ANY_SOURCE specification: https://svn.mpi-forum.org/trac/mpi-forum-web/wiki/ft/run_through_stabilizat… After the teleconf last week it seemed that we probably cannot explicitly say much about the order of delivery of pending messages versus an error class when a process fails on the communicator. What do people think about this? If it helps for comparison, FT-MPI has a couple message modes that they specified (my summaries from [1]): FTMPI_MSG_MODE_RESET: All messages from any process dropped after a failure is detected. FTMPI_MSG_MODE_CONT : Only messages from the failed process are dropped after a failure is detected. MPI_Abort --------- https://svn.mpi-forum.org/trac/mpi-forum-web/wiki/ft/run_through_stabilizat… It was mentioned during the last call that the intention of MPI_Abort is unclear in the case where you specify MPI_COMM_SELF as the argument. I added a couple notes to the wiki about this. It is possible that MPI_Abort should be more carefully defined beyond just this case to, for example, clarify the intention of the communicator argument and its impact of all derived {inter|intra}communicators. MPI_Kill -------- https://svn.mpi-forum.org/trac/mpi-forum-web/wiki/ft/run_through_stabilizat… If MPI_Abort is meant to be a distributed 'exit' option for the MPI application to terminate the whole job, we probably want the ability to remove a specific process from the computation. The MPI_Kill command allows us to specify a specific process to terminate identified by communicator and rank. The process will be excluded from all communicators that it is a part of as if it failed. Do people like this option? Have suggestions on further semantics? Thoughts? -- Josh [1] http://icl.cs.utk.edu/projectsfiles/ftmpi/pubs/isc2004-FT-MPI.pdf ------------------------------------ Joshua Hursey Postdoctoral Research Associate Oak Ridge National Laboratory http://users.nccs.gov/~jjhursey

1 0