November 2013 - mpi-forum - mailman3.cels.anl.gov

Re: [Mpi-forum] New Logo for the MPI Forum
by Jeff Squyres (jsquyres) 13 Nov '13

13 Nov '13

Yes! We are in dire need of a real logo. If you have real artists who can supply some proposals, that would be fabulous. On Nov 13, 2013, at 4:43 PM, Jerome Vienne <viennej(a)tacc.utexas.edu> wrote: > Hi Martin, > > A "Communication, Media, & Design" team has been created at TACC recently, and I will be happy to work with them to bring few logo proposals before the next meeting. > Are you still looking for a new logo ? > > Thanks, > Jerome > -- > Jerome Vienne, Ph.D > HPC Software Tools Group > Texas Advanced Computing Center (TACC) > viennej(a)tacc.utexas.edu | Phone: (512) 475-9322 > Office: ROC 1.455B | Fax: (512) 475-9445 > > > On Sat, Sep 14, 2013 at 12:55 PM, Martin Schulz <schulz6(a)llnl.gov> wrote: > Hi all, > > It's long overdue that we use a new logo for the MPI forum - in case you haven't looked at the forum website in a while, it still says MPI-2. At the forum meeting in Madrid we decided on the following process (to make sure we drive this forward). Anyone who has a suggestion for a new logo, please send it to me (until 2 weeks before the meetings). I will collect all of them and then send them to everyone for comments. We will then vote at the next forum meeting in Chicago and pick one. Currently, we have two proposals on the table (created by the UTK folks a while ago - thanks!) - those are attached - but the more choices we have the better. > > Get creative … :) > > Martin > > > _______________________________________________ > mpi-forum mailing list > mpi-forum(a)lists.mpi-forum.org > http://lists.mpi-forum.org/mailman/listinfo.cgi/mpi-forum > > _______________________________________________ > mpi-forum mailing list > mpi-forum(a)lists.mpi-forum.org > http://lists.mpi-forum.org/mailman/listinfo.cgi/mpi-forum -- Jeff Squyres jsquyres(a)cisco.com For corporate legal information go to: http://www.cisco.com/web/about/doing_business/legal/cri/

1 0

Re: [Mpi-forum] New Logo for the MPI Forum
by Jerome Vienne 13 Nov '13

13 Nov '13

Hi Martin, A "Communication, Media, & Design" team has been created at TACC recently, and I will be happy to work with them to bring few logo proposals before the next meeting. Are you still looking for a new logo ? Thanks, Jerome -- Jerome Vienne, Ph.D HPC Software Tools Group Texas Advanced Computing Center (TACC) viennej(a)tacc.utexas.edu | Phone: (512) 475-9322 Office: ROC 1.455B | Fax: (512) 475-9445 On Sat, Sep 14, 2013 at 12:55 PM, Martin Schulz <schulz6(a)llnl.gov> wrote: > Hi all, > > It's long overdue that we use a new logo for the MPI forum - in case you > haven't looked at the forum website in a while, it still says MPI-2. At the > forum meeting in Madrid we decided on the following process (to make sure > we drive this forward). Anyone who has a suggestion for a new logo, please > send it to me (until 2 weeks before the meetings). I will collect all of > them and then send them to everyone for comments. We will then vote at the > next forum meeting in Chicago and pick one. Currently, we have two > proposals on the table (created by the UTK folks a while ago - thanks!) - > those are attached - but the more choices we have the better. > > Get creative … :) > > Martin > > > _______________________________________________ > mpi-forum mailing list > mpi-forum(a)lists.mpi-forum.org > http://lists.mpi-forum.org/mailman/listinfo.cgi/mpi-forum >

1 0

Re: [Mpi-forum] Question about the semantics of MPI_Comm_disconnect
by George Bosilca 13 Nov '13

13 Nov '13

Disconnect doesn’t work at the physical connection level, it is supposed to mark two previously connected world as disconnected in the MPI sense. In MPI there is this transitive notion of connected processes, which can only be reversed by the usage of disconnect. Basically MPI_Finalize is collective over all connected processes, which is a burden in some cases. Disconnect is here to allow the disruption of the connected notion. Thus, Disconnect only has a special meaning when called on a communicator returned by SPAWN, ACCEPT/CONNECT and JOIN. Any other usage is similar to MPI_Comm_Free. This in well explained in 10.5.4 once at the beginning of the section and then again in the first advice to users. George. PS: That being said if instead of calling Disconnect one release all communicators with MPI_Comm_free the processes are still connected, and the MPI_Finalize remains global with a scope outside a single MPI_WORLD. On Nov 13, 2013, at 22:00 , Jeff Squyres (jsquyres) <jsquyres(a)cisco.com> wrote: > On Nov 12, 2013, at 5:40 PM, Rajeev Thakur <thakur(a)mcs.anl.gov> wrote: > >> Think of Comm_disconnect as "whatever connection was there between client and server is gone". > > > Rajeev -- > > This sentence has been bugging me, and I just figured out why. > > I've always thought of COMM_DISCONNECT just as you describe: whatever connection was there between client and server is gone. But what about other types of MPI handles -- like MPI_File and MPI_Window? For example: > > ----- > MPI_Comm_spawn(..., &child); > MPI_Intercomm_merge(child, 0, &child_intra); > MPI_File_open(child_intra, ..., &file); > MPI_Comm_disconnect(&child_intra); > MPI_Comm_disconnect(&child); > ----- > > At this point, all communicators between parent and child are gone, but there's still a File (the same example could be done with an RMA window). > > So the processes are *not* actually disconnected (in the MPI sense of the word "disconnected). Indeed, in the above example, if there was no File, you still have to do *2* disconnects before the parent and child are truly disconnected. > > Hence, I think that this is a very poorly named function -- it *doesn't* necessarily disconnect MPI processes. It really performs and operation that is something like MPI_COMM_FREE_WAIT. > > I wonder if this is worth an AtoU... > > -- > Jeff Squyres > jsquyres(a)cisco.com > For corporate legal information go to: http://www.cisco.com/web/about/doing_business/legal/cri/ > > _______________________________________________ > mpi-forum mailing list > mpi-forum(a)lists.mpi-forum.org > http://lists.mpi-forum.org/mailman/listinfo.cgi/mpi-forum

1 0

Re: [Mpi-forum] Question about the semantics of MPI_Comm_disconnect
by Jeff Squyres (jsquyres) 13 Nov '13

13 Nov '13

On Nov 12, 2013, at 5:40 PM, Rajeev Thakur <thakur(a)mcs.anl.gov> wrote: > Think of Comm_disconnect as "whatever connection was there between client and server is gone". Rajeev -- This sentence has been bugging me, and I just figured out why. I've always thought of COMM_DISCONNECT just as you describe: whatever connection was there between client and server is gone. But what about other types of MPI handles -- like MPI_File and MPI_Window? For example: ----- MPI_Comm_spawn(..., &child); MPI_Intercomm_merge(child, 0, &child_intra); MPI_File_open(child_intra, ..., &file); MPI_Comm_disconnect(&child_intra); MPI_Comm_disconnect(&child); ----- At this point, all communicators between parent and child are gone, but there's still a File (the same example could be done with an RMA window). So the processes are *not* actually disconnected (in the MPI sense of the word "disconnected). Indeed, in the above example, if there was no File, you still have to do *2* disconnects before the parent and child are truly disconnected. Hence, I think that this is a very poorly named function -- it *doesn't* necessarily disconnect MPI processes. It really performs and operation that is something like MPI_COMM_FREE_WAIT. I wonder if this is worth an AtoU... -- Jeff Squyres jsquyres(a)cisco.com For corporate legal information go to: http://www.cisco.com/web/about/doing_business/legal/cri/

1 0

December Meeting
by Schulz, Martin 13 Nov '13

13 Nov '13

Hi all, I updated the meetings page and included details on the next meeting in Chicago. For local logistics, I basically for now copied the information from our last meeting in Chicago (will show up within the hour). Rich will update this as soon as he has the registration link. Also, I am not sure if we have a room block this time - Rich, please let us know what has been arranged - thanks! As discussed during the last two forum meetings, we slightly adjusted the schedule - we will start at 2pm on Monday and stop before lunch on Thursday. To compensate for the lost time compared to the old schedule, we agreed to work longer hours if necessary. As for the agenda - I am hoping to again do working groups on Monday and Tuesday (potential 2 groups in parallel). Wednesday and Thursday is reserved for plenary items, incl. errata presentations and votes. Aside from that, I am aware of two large blocks of discussion (FT and Hybrid) and for now gave each topic a half day for discussions - this can of course be changed as need. Working group leads: please let me know how much time you would like for your group in the break-outs and whether you need plenary time or not. Please let me know of any tickets that are up for readings or votes. Thanks! Martin PS: Also, please note that the dates for next year are fixed now. ________________________________________________________________________ Martin Schulz, schulzm(a)llnl.gov, http://people.llnl.gov/schulzm CASC @ Lawrence Livermore National Laboratory, Livermore, USA

1 0

Re: [Mpi-forum] Question about the semantics of MPI_Comm_disconnect
by Jeff Squyres (jsquyres) 12 Nov '13

12 Nov '13

On Nov 12, 2013, at 7:24 PM, George Bosilca <bosilca(a)icl.utk.edu> wrote: >> I understand that it is desirable to be vague in some places. But this is not a case of being vague: it's a case of being *wrong*. And it should therefore be fixed. > > The challenge here was to explain a concept without referring to the notion of “unexpected” messages as it is lacking from the standard. I guess any reasonable user would understand that files must be closed, windows must be synchronized, but for pt2pt things are different because two sides are involved. Marking them as complete is only enough on one side, they should be matched in the other in order to really disappear. Thus, completed and matched make sense … Yes -- for p2p only. :-) (we're violently agreeing, I think) > Anyway, if you propose to replace this text with the one Rajeev mentioned in his email (from MPI_Finalize), I guess is a good solution. Yeah, or a reference to it. Nathan and I will cook up an errata proposal. -- Jeff Squyres jsquyres(a)cisco.com For corporate legal information go to: http://www.cisco.com/web/about/doing_business/legal/cri/

1 0

Re: [Mpi-forum] Question about the semantics of MPI_Comm_disconnect
by George Bosilca 12 Nov '13

12 Nov '13

On Nov 13, 2013, at 01:09 , Jeff Squyres (jsquyres) <jsquyres(a)cisco.com> wrote: > On Nov 12, 2013, at 6:55 PM, George Bosilca <bosilca(a)icl.utk.edu> wrote: > >> It is a sentence inherited from MPI 1, and it made sense in a world dominated by pt2pt. > > Sure, I assumed that. :-) > >> It is very literal and it does not apply “as it” to all concepts in the __current__ standard, but the intent is remain clear: prevent any form of unexpected messages. > > I understand that it is desirable to be vague in some places. But this is not a case of being vague: it's a case of being *wrong*. And it should therefore be fixed. The challenge here was to explain a concept without referring to the notion of “unexpected” messages as it is lacking from the standard. I guess any reasonable user would understand that files must be closed, windows must be synchronized, but for pt2pt things are different because two sides are involved. Marking them as complete is only enough on one side, they should be matched in the other in order to really disappear. Thus, completed and matched make sense … Anyway, if you propose to replace this text with the one Rajeev mentioned in his email (from MPI_Finalize), I guess is a good solution. George. > > -- > Jeff Squyres > jsquyres(a)cisco.com > For corporate legal information go to: http://www.cisco.com/web/about/doing_business/legal/cri/ > > _______________________________________________ > mpi-forum mailing list > mpi-forum(a)lists.mpi-forum.org > http://lists.mpi-forum.org/mailman/listinfo.cgi/mpi-forum

1 0

Re: [Mpi-forum] Question about the semantics of MPI_Comm_disconnect
by Jeff Squyres (jsquyres) 12 Nov '13

12 Nov '13

On Nov 12, 2013, at 6:55 PM, George Bosilca <bosilca(a)icl.utk.edu> wrote: > It is a sentence inherited from MPI 1, and it made sense in a world dominated by pt2pt. Sure, I assumed that. :-) > It is very literal and it does not apply “as it” to all concepts in the __current__ standard, but the intent is remain clear: prevent any form of unexpected messages. I understand that it is desirable to be vague in some places. But this is not a case of being vague: it's a case of being *wrong*. And it should therefore be fixed. -- Jeff Squyres jsquyres(a)cisco.com For corporate legal information go to: http://www.cisco.com/web/about/doing_business/legal/cri/

1 0

Re: [Mpi-forum] Next MPI Forum meeting (Chicago Dec 9-12)
by Schulz, Martin 12 Nov '13

12 Nov '13

Hi all, Rich and one of his admins are working on the logistics for meeting and we'll put that up as soon as possible. The logistics will be the same as usual, though, for Chicago. As for the agenda, I was planning on sending a request for input later tonight. On Nov 12, 2013, at 3:47 PM, Steve Oyanagi <sko(a)cray.com> wrote: > Hi, > > I am perhaps too early, but I note there are no logistics > or agenda for the Chicago meeting which now less than a > month away. > > Did I miss something and this meeting was canceled? > > I am beginning to make trip preparations, so knowing: > > 1. That the meeting is occurring, Yes. > 2. When we start on Monday, Monday 2pm > 3. Tentatively when we stop on Thursday (I realize that > might change), The plan is Thursday noon. > 4. Where is the meeting hotel and rate, I don't think we had meeting hotels or rates for a while now (Rich, please correct me if I am wrong) - the meeting will be at Microsoft in the Aon Center, 2nd floor. > would all be helpful. I assume the meeting is at the same > Microsoft location as the other Chicago meeting in downtown. Yes. Martin > I did not attend the last meeting in Madrid so I am unaware > of any arrangements made there, and I have not seen a > recap of next meeting info. > > Thanks much, > - Steve > _______________________________________________ > mpi-forum mailing list > mpi-forum(a)lists.mpi-forum.org > http://lists.mpi-forum.org/mailman/listinfo.cgi/mpi-forum ________________________________________________________________________ Martin Schulz, schulzm(a)llnl.gov, http://people.llnl.gov/schulzm CASC @ Lawrence Livermore National Laboratory, Livermore, USA

1 0

Re: [Mpi-forum] Next MPI Forum meeting (Chicago Dec 9-12)
by George Bosilca 12 Nov '13

12 Nov '13

+1, especially regarding the hotel. George. On Nov 13, 2013, at 00:47 , Steve Oyanagi <sko(a)cray.com> wrote: > Hi, > > I am perhaps too early, but I note there are no logistics > or agenda for the Chicago meeting which now less than a > month away. > > Did I miss something and this meeting was canceled? > > I am beginning to make trip preparations, so knowing: > > 1. That the meeting is occurring, > > 2. When we start on Monday, > > 3. Tentatively when we stop on Thursday (I realize that > might change), > > 4. Where is the meeting hotel and rate, > > would all be helpful. I assume the meeting is at the same > Microsoft location as the other Chicago meeting in downtown. > > I did not attend the last meeting in Madrid so I am unaware > of any arrangements made there, and I have not seen a > recap of next meeting info. > > Thanks much, > - Steve > _______________________________________________ > mpi-forum mailing list > mpi-forum(a)lists.mpi-forum.org > http://lists.mpi-forum.org/mailman/listinfo.cgi/mpi-forum

1 0