- mpiwg-coll - mailman3.cels.anl.gov

Re: [mpi3-coll] March Teleconference
by Jesper Larsson Traeff 05 Mar '09

05 Mar '09

Hi all, should work for me Jesper On Wed, Mar 04, 2009 at 04:12:27PM -0500, Torsten Hoefler wrote: > Hello Group, > given Bronis' constraints, I'll move the telecon to the 19th 11am EST > (8am PST, 17h German time). > > Please let me know if this does not work? > > Best, > Torsten > > > > I'd like to propose this Thursday, the 5th or the 19th as dates for our > > > monthly teleconference. The time would be noon EST as usual. > > > > This Thursday does not work for me. > > > > > Does anybody object? If not, the I'll reserve this Thursday (deadline > > > for objections is Wednesday). We can also choose another time or day in > > > those weeks. > > > > Wednesday of this week at 10AM Pacific would work for me. > > I can make 8AM this Wednesday work also. > > > > Next week would be better for me. You were out of phase with my > > regular biweekly 9AM Thursday telecon but your proposal puts > > you in phase, which would mean I cannot make it. > > > > The following week I could make Thursday (March 19) at 8AM Pacific. > > > > Bronis > > > > > > > I would propose the following two important agenda items: > > > > > > 1) wrapup NBC proposal (this should be really short as we don't have > > > many changes) > > > - issues with MPI_IN_PLACE and #45 in the proposal > > > - intercommunicator proof of concept implementation > > > > > > 2) fixing the graph topology interface in MPI-2.2 (time-critical!) > > > - discuss proposal #33 > > > > > > Anything else? > > > > > > Thanks & Best, > > > Torsten > > > > > > -- > > > bash$ :(){ :|:&};: --------------------- http:// www. unixer.de/ ----- > > > Torsten Hoefler | Postdoctoral Researcher > > > Open Systems Lab | Indiana University > > > 150 S. Woodlawn Ave. | Bloomington, IN, 474045, USA > > > Lindley Hall Room 135 | +01 (812) 855-3608 > > > _______________________________________________ > > > mpi3-coll mailing list > > > mpi3-coll(a)lists.mpi-forum.org > > > http:// lists.mpi-forum.org/mailman/listinfo.cgi/mpi3-coll > > > > > > > > > > -- > bash$ :(){ :|:&};: --------------------- http://www.unixer.de/ ----- > Torsten Hoefler | Postdoctoral Researcher > Open Systems Lab | Indiana University > 150 S. Woodlawn Ave. | Bloomington, IN, 474045, USA > Lindley Hall Room 135 | +01 (812) 855-3608 > _______________________________________________ > mpi3-coll mailing list > mpi3-coll(a)lists.mpi-forum.org > http://lists.mpi-forum.org/mailman/listinfo.cgi/mpi3-coll

1 0

Re: [mpi3-coll] March Teleconference
by Torsten Hoefler 04 Mar '09

04 Mar '09

On Wed, Mar 04, 2009 at 03:03:15PM -0800, Bronis R. de Supinski wrote: > > Torsten: > > Works great for me! Thanks. What's the call information? this will be announced as soon as we decided on a date (comments are not closed yet ;). I'll usually announce it at the beginning of the week on this list. Best, Torsten -- bash$ :(){ :|:&};: --------------------- http://www.unixer.de/ ----- Torsten Hoefler | Postdoctoral Researcher Open Systems Lab | Indiana University 150 S. Woodlawn Ave. | Bloomington, IN, 474045, USA Lindley Hall Room 135 | +01 (812) 855-3608

1 0

Re: [mpi3-coll] March Teleconference
by Bronis R. de Supinski 04 Mar '09

04 Mar '09

Torsten: Works great for me! Thanks. What's the call information? Bronis On Wed, 4 Mar 2009, Torsten Hoefler wrote: > Hello Group, > given Bronis' constraints, I'll move the telecon to the 19th 11am EST > (8am PST, 17h German time). > > Please let me know if this does not work? > > Best, > Torsten > > > > I'd like to propose this Thursday, the 5th or the 19th as dates for our > > > monthly teleconference. The time would be noon EST as usual. > > > > This Thursday does not work for me. > > > > > Does anybody object? If not, the I'll reserve this Thursday (deadline > > > for objections is Wednesday). We can also choose another time or day in > > > those weeks. > > > > Wednesday of this week at 10AM Pacific would work for me. > > I can make 8AM this Wednesday work also. > > > > Next week would be better for me. You were out of phase with my > > regular biweekly 9AM Thursday telecon but your proposal puts > > you in phase, which would mean I cannot make it. > > > > The following week I could make Thursday (March 19) at 8AM Pacific. > > > > Bronis > > > > > > > I would propose the following two important agenda items: > > > > > > 1) wrapup NBC proposal (this should be really short as we don't have > > > many changes) > > > - issues with MPI_IN_PLACE and #45 in the proposal > > > - intercommunicator proof of concept implementation > > > > > > 2) fixing the graph topology interface in MPI-2.2 (time-critical!) > > > - discuss proposal #33 > > > > > > Anything else? > > > > > > Thanks & Best, > > > Torsten > > > > > > -- > > > bash$ :(){ :|:&};: --------------------- http:// www. unixer.de/ ----- > > > Torsten Hoefler | Postdoctoral Researcher > > > Open Systems Lab | Indiana University > > > 150 S. Woodlawn Ave. | Bloomington, IN, 474045, USA > > > Lindley Hall Room 135 | +01 (812) 855-3608 > > > _______________________________________________ > > > mpi3-coll mailing list > > > mpi3-coll(a)lists.mpi-forum.org > > > http:// lists.mpi-forum.org/mailman/listinfo.cgi/mpi3-coll > > > > > > > > > > -- > bash$ :(){ :|:&};: --------------------- http:// www. unixer.de/ ----- > Torsten Hoefler | Postdoctoral Researcher > Open Systems Lab | Indiana University > 150 S. Woodlawn Ave. | Bloomington, IN, 474045, USA > Lindley Hall Room 135 | +01 (812) 855-3608 > _______________________________________________ > mpi3-coll mailing list > mpi3-coll(a)lists.mpi-forum.org > http:// lists.mpi-forum.org/mailman/listinfo.cgi/mpi3-coll > >

1 0

Re: [mpi3-coll] March Teleconference
by Torsten Hoefler 04 Mar '09

04 Mar '09

Hello Group, given Bronis' constraints, I'll move the telecon to the 19th 11am EST (8am PST, 17h German time). Please let me know if this does not work? Best, Torsten > > I'd like to propose this Thursday, the 5th or the 19th as dates for our > > monthly teleconference. The time would be noon EST as usual. > > This Thursday does not work for me. > > > Does anybody object? If not, the I'll reserve this Thursday (deadline > > for objections is Wednesday). We can also choose another time or day in > > those weeks. > > Wednesday of this week at 10AM Pacific would work for me. > I can make 8AM this Wednesday work also. > > Next week would be better for me. You were out of phase with my > regular biweekly 9AM Thursday telecon but your proposal puts > you in phase, which would mean I cannot make it. > > The following week I could make Thursday (March 19) at 8AM Pacific. > > Bronis > > > > I would propose the following two important agenda items: > > > > 1) wrapup NBC proposal (this should be really short as we don't have > > many changes) > > - issues with MPI_IN_PLACE and #45 in the proposal > > - intercommunicator proof of concept implementation > > > > 2) fixing the graph topology interface in MPI-2.2 (time-critical!) > > - discuss proposal #33 > > > > Anything else? > > > > Thanks & Best, > > Torsten > > > > -- > > bash$ :(){ :|:&};: --------------------- http:// www. unixer.de/ ----- > > Torsten Hoefler | Postdoctoral Researcher > > Open Systems Lab | Indiana University > > 150 S. Woodlawn Ave. | Bloomington, IN, 474045, USA > > Lindley Hall Room 135 | +01 (812) 855-3608 > > _______________________________________________ > > mpi3-coll mailing list > > mpi3-coll(a)lists.mpi-forum.org > > http:// lists.mpi-forum.org/mailman/listinfo.cgi/mpi3-coll > > > > > -- bash$ :(){ :|:&};: --------------------- http://www.unixer.de/ ----- Torsten Hoefler | Postdoctoral Researcher Open Systems Lab | Indiana University 150 S. Woodlawn Ave. | Bloomington, IN, 474045, USA Lindley Hall Room 135 | +01 (812) 855-3608

1 0

Re: [mpi3-coll] Question on reduction operation in intercommunicators
by Jesper Larsson Traeff 04 Mar '09

04 Mar '09

Hi Adam, I think that in general, there should be no essential distinctions about requirements and restrictions on predefined and user-defined op's. The standard might need some clearing up in this respect, although I will not propose anything for now It would probably be reasonable for intercomms only to require that the same op is given between processes in the same local group (the intention of intercomms being (?) that as much as possible is done locally, and that communication between the two groups is applied parsimonously). However, a) it would complicate the semantics even more b) I'm not sure what really the benefit would be, if any (would be good if you could support with a strong example) c) it would disallow implementations where data would be communicated to the other group and (partly) reduced there So unless there's a really compelleing case for this, I'd rather leave as is (maybe the text needs clearing up?) best, jesper On Fri, Feb 13, 2009 at 12:57:10PM -0800, Adam Moody wrote: > Hi everyone, > The spec is a bit fuzzy regarding the op parameter for reduction-type > collectives on intercommunicators. There are a few lines in the > intracommunicator text under MPI_REDUCE, which says: > > page 161, lines 11-15: > > "Furthermore, the datatype and op given for predefined operators > must be the same on all processes." > "Note, that it is possible for users to supply different > user-defined operations to MPI_REDUCE in each process. MPI does not > define which operations are used on which operands in this case." > > However, nothing is explicitly said along such lines for > intercommunicators. I assume by default the above statements must > extend to intercommunicators as well. > > I have two questions. > > 1) For intracommunicators, I understand the restriction on having > the same op specified by all processes for predefined operators -- > performance. I could see a similar argument being made that all > processes (from both groups) in an intercomm also specify the same op > for predefined ops. However, I also see utility in forcing all > processes within a group to specify the same op, but allowing each group > to specify different ops (e.g., one may want the maximum of values from > group A and the minimum of values from group B). The downside in this > case is that processes of one group don't know the operation used by > processes in the other group without some communication. Also, we'd > need to be precise in specifying what data the two ops apply to (i.e., > do I apply the op specified in group A to the data from group A or the > data from group B?). > > 2) A similar question arises for user-defined operators, since here > the current spec says that different processes may specify different > ops. It says MPI can apply which ever ops to which ever operands in > such a case for intracommunicators. Again, I could see value in > enabling the user to specify the same user-defined op within a group, > but different ops across groups. > > Any thoughts? > -Adam > _______________________________________________ > mpi3-coll mailing list > mpi3-coll(a)lists.mpi-forum.org > http://lists.mpi-forum.org/mailman/listinfo.cgi/mpi3-coll

1 0

Re: [mpi3-coll] March Teleconference
by Jesper Larsson Traeff 03 Mar '09

03 Mar '09

I'm ok with both dates, but better with 19th agenda is good, the topology extension should (in my opinion) find it's way into MPI 2.2, so it would be good to start serious discussion on the proposal Jesper On Mon, Mar 02, 2009 at 03:06:21PM -0500, Torsten Hoefler wrote: > Hello Coll-WG, > I'd like to propose this Thursday, the 5th or the 19th as dates for our > monthly teleconference. The time would be noon EST as usual. > > Does anybody object? If not, the I'll reserve this Thursday (deadline > for objections is Wednesday). We can also choose another time or day in > those weeks. > > I would propose the following two important agenda items: > > 1) wrapup NBC proposal (this should be really short as we don't have > many changes) > - issues with MPI_IN_PLACE and #45 in the proposal > - intercommunicator proof of concept implementation > > 2) fixing the graph topology interface in MPI-2.2 (time-critical!) > - discuss proposal #33 > > Anything else? > > Thanks & Best, > Torsten > > -- > bash$ :(){ :|:&};: --------------------- http://www.unixer.de/ ----- > Torsten Hoefler | Postdoctoral Researcher > Open Systems Lab | Indiana University > 150 S. Woodlawn Ave. | Bloomington, IN, 474045, USA > Lindley Hall Room 135 | +01 (812) 855-3608 > _______________________________________________ > mpi3-coll mailing list > mpi3-coll(a)lists.mpi-forum.org > http://lists.mpi-forum.org/mailman/listinfo.cgi/mpi3-coll

1 0

Re: [mpi3-coll] March Teleconference
by Supalov, Alexander 03 Mar '09

03 Mar '09

Hi, I'm for 19th. Best regards. Alexander -----Original Message----- From: mpi3-coll-bounces(a)lists.mpi-forum.org [mailto:[email protected]] On Behalf Of Torsten Hoefler Sent: Monday, March 02, 2009 9:06 PM To: mpi3-coll(a)lists.mpi-forum.org Subject: [mpi3-coll] March Teleconference Hello Coll-WG, I'd like to propose this Thursday, the 5th or the 19th as dates for our monthly teleconference. The time would be noon EST as usual. Does anybody object? If not, the I'll reserve this Thursday (deadline for objections is Wednesday). We can also choose another time or day in those weeks. I would propose the following two important agenda items: 1) wrapup NBC proposal (this should be really short as we don't have many changes) - issues with MPI_IN_PLACE and #45 in the proposal - intercommunicator proof of concept implementation 2) fixing the graph topology interface in MPI-2.2 (time-critical!) - discuss proposal #33 Anything else? Thanks & Best, Torsten -- bash$ :(){ :|:&};: --------------------- http://www.unixer.de/ ----- Torsten Hoefler | Postdoctoral Researcher Open Systems Lab | Indiana University 150 S. Woodlawn Ave. | Bloomington, IN, 474045, USA Lindley Hall Room 135 | +01 (812) 855-3608 _______________________________________________ mpi3-coll mailing list mpi3-coll(a)lists.mpi-forum.org http://lists.mpi-forum.org/mailman/listinfo.cgi/mpi3-coll --------------------------------------------------------------------- Intel GmbH Dornacher Strasse 1 85622 Feldkirchen/Muenchen Germany Sitz der Gesellschaft: Feldkirchen bei Muenchen Geschaeftsfuehrer: Douglas Lusk, Peter Gleissner, Hannes Schwaderer Registergericht: Muenchen HRB 47456 Ust.-IdNr. VAT Registration No.: DE129385895 Citibank Frankfurt (BLZ 502 109 00) 600119052 This e-mail and any attachments may contain confidential material for the sole use of the intended recipient(s). Any review or distribution by others is strictly prohibited. If you are not the intended recipient, please contact the sender and delete all copies.

1 0

[mpi3-coll] March Teleconference
by Torsten Hoefler 02 Mar '09

02 Mar '09

Hello Coll-WG, I'd like to propose this Thursday, the 5th or the 19th as dates for our monthly teleconference. The time would be noon EST as usual. Does anybody object? If not, the I'll reserve this Thursday (deadline for objections is Wednesday). We can also choose another time or day in those weeks. I would propose the following two important agenda items: 1) wrapup NBC proposal (this should be really short as we don't have many changes) - issues with MPI_IN_PLACE and #45 in the proposal - intercommunicator proof of concept implementation 2) fixing the graph topology interface in MPI-2.2 (time-critical!) - discuss proposal #33 Anything else? Thanks & Best, Torsten -- bash$ :(){ :|:&};: --------------------- http://www.unixer.de/ ----- Torsten Hoefler | Postdoctoral Researcher Open Systems Lab | Indiana University 150 S. Woodlawn Ave. | Bloomington, IN, 474045, USA Lindley Hall Room 135 | +01 (812) 855-3608

1 0

Re: [mpi3-coll] NBC Draft Revision 5
by Adam Moody 26 Feb '09

26 Feb '09

Alright, that's enough votes for me. I'll close ticket #131 and we should update the NBC text, as well. -Adam Torsten Hoefler wrote: >I'd also back option #2 > >Best, > Torsten > > >

1 0

Re: [mpi3-coll] NBC Draft Revision 5
by Torsten Hoefler 26 Feb '09

26 Feb '09

I'd also back option #2 Best, Torsten -- bash$ :(){ :|:&};: --------------------- http://www.unixer.de/ ----- Torsten Hoefler | Postdoctoral Researcher Open Systems Lab | Indiana University 150 S. Woodlawn Ave. | Bloomington, IN, 474045, USA Lindley Hall Room 135 | +01 (812) 855-3608

1 0