February 2014 - mpiwg-ft - mailman3.cels.anl.gov

Re: [mpiwg-ft] Proposal Feedback
by Bland, Wesley B. 25 Feb '14

25 Feb '14

On Feb 23, 2014, at 7:20 AM, Jim Dinan <james.dinan(a)gmail.com> wrote: > Hi Aurelien, > > > (1) Are the semantics for the MPI probe operations (probe, iprobe, mprobe) well defined? Including when MPI_ANY_SOURCE is used? > > > Yes, they are exactly the same as for normal recv operations. We had a WG phone call issue on mprobe alone, sometime last september, to verify that it was not problematic. > > We state that non-blocking operations must not raise an error when they are initiated. We may need an exception to this rule for MPI_Iprobe to state that it has MPI_Test semantics. An Iprobe could be polling with MPI_ANY_SOURCE and lead to deadlock if it does not return an error. Hmm, that’s a good point. I guess the probe functions wouldn’t be covered under the text about the completion functions. I guess we do need to add text here. It’s inside the two week deadline so I’m not sure what the right thing to do here would be. > > > (2) For comm_shrink, what happens when the remote group in an intercommunicator becomes empty? Is it valid to get back an intercommunicator with an empty remote group? > That’s a pretty bizarre case. It is to be noted that the semantic for the shrink is well defined: the outcome is the same as if the user had called split with colors set to MPI_UNDEFINED at all remote group processes. From the definition of SPLIT stated page 247, the well defined outcome is that MPI_COMM_NULL is returned when a color is used on only one side of the intercomm. > > Ok, thanks! Just wanted to make sure that this was defined. > > > (3) In 15.2 (pg. 594:21), can we replace the sentence "always complete in a finite amount of time" with the simpler "eventually complete”? > I believe the original formulation carries the argument more visibly by being pedantic; but if you feel strongly about it, I don't. > > The wording isn't incorrect, this is just word smithing. Strong words "always" and "finite" imply something about when a thing will occur, when in fact the semantic is not any stronger than "eventually". "Eventually" is used for semantics like this in the RMA chapter. > > ~Jim. > _______________________________________________ > mpiwg-ft mailing list > mpiwg-ft(a)lists.mpi-forum.org > http://lists.mpi-forum.org/mailman/listinfo.cgi/mpiwg-ft

1 0

Re: [mpiwg-ft] Proposal Feedback
by Jim Dinan 23 Feb '14

23 Feb '14

Hi Aurelien, > (1) Are the semantics for the MPI probe operations (probe, iprobe, > mprobe) well defined? Including when MPI_ANY_SOURCE is used? > > > Yes, they are exactly the same as for normal recv operations. We had a WG > phone call issue on mprobe alone, sometime last september, to verify that > it was not problematic. We state that non-blocking operations must not raise an error when they are initiated. We may need an exception to this rule for MPI_Iprobe to state that it has MPI_Test semantics. An Iprobe could be polling with MPI_ANY_SOURCE and lead to deadlock if it does not return an error. > (2) For comm_shrink, what happens when the remote group in an > intercommunicator becomes empty? Is it valid to get back an > intercommunicator with an empty remote group? > That's a pretty bizarre case. It is to be noted that the semantic for the > shrink is well defined: the outcome is the same as if the user had called > split with colors set to MPI_UNDEFINED at all remote group processes. From > the definition of SPLIT stated page 247, the well defined outcome is that > MPI_COMM_NULL is returned when a color is used on only one side of the > intercomm. Ok, thanks! Just wanted to make sure that this was defined. > > (3) In 15.2 (pg. 594:21), can we replace the sentence "always complete > in a finite amount of time" with the simpler "eventually complete"? > I believe the original formulation carries the argument more visibly by > being pedantic; but if you feel strongly about it, I don't. The wording isn't incorrect, this is just word smithing. Strong words "always" and "finite" imply something about when a thing will occur, when in fact the semantic is not any stronger than "eventually". "Eventually" is used for semantics like this in the RMA chapter. ~Jim.

1 0

Re: [mpiwg-ft] Proposal Feedback
by Aurélien Bouteiller 21 Feb '14

21 Feb '14

Le 21 févr. 2014 à 17:54, Jim Dinan <james.dinan(a)gmail.com> a écrit : > Hi All, > > In reviewing the proposal, a couple questions/comments came up: > > (1) Are the semantics for the MPI probe operations (probe, iprobe, mprobe) well defined? Including when MPI_ANY_SOURCE is used? > Yes, they are exactly the same as for normal recv operations. We had a WG phone call issue on mprobe alone, sometime last september, to verify that it was not problematic. > (2) For comm_shrink, what happens when the remote group in an intercommunicator becomes empty? Is it valid to get back an intercommunicator with an empty remote group? That’s a pretty bizarre case. It is to be noted that the semantic for the shrink is well defined: the outcome is the same as if the user had called split with colors set to MPI_UNDEFINED at all remote group processes. From the definition of SPLIT stated page 247, the well defined outcome is that MPI_COMM_NULL is returned when a color is used on only one side of the intercomm. > > (3) In 15.2 (pg. 594:21), can we replace the sentence "always complete in a finite amount of time" with the simpler "eventually complete”? I believe the original formulation carries the argument more visibly by being pedantic; but if you feel strongly about it, I don't. > ~Jim. > _______________________________________________ > mpiwg-ft mailing list > mpiwg-ft(a)lists.mpi-forum.org > http://lists.mpi-forum.org/mailman/listinfo.cgi/mpiwg-ft

1 0

Proposal Feedback
by Jim Dinan 21 Feb '14

21 Feb '14

Hi All, In reviewing the proposal, a couple questions/comments came up: (1) Are the semantics for the MPI probe operations (probe, iprobe, mprobe) well defined? Including when MPI_ANY_SOURCE is used? (2) For comm_shrink, what happens when the remote group in an intercommunicator becomes empty? Is it valid to get back an intercommunicator with an empty remote group? (3) In 15.2 (pg. 594:21), can we replace the sentence "always complete in a finite amount of time" with the simpler "eventually complete"? ~Jim.

1 0

FTWG Con Call 2014-02-18
by Bland, Wesley B. 18 Feb '14

18 Feb '14

Dear WG members, This is a reminder that we have our regular phone meeting today at 3:00 PM Eastern. This week we’ll be continuing our run-up for the reading in March. Date: February 18 Time: 3 PM Eastern Dial-in information: 712-432-0360 Code: 623998# Agenda: * Wrap-up of external discussions * Final touches before reading Next Meetings: * March 4, 2014 (face-to-face)

1 0

Re: [mpiwg-ft] Proof reading the draft
by Aurélien Bouteiller 11 Feb '14

11 Feb '14

Thanks Thomas, I’ll send in the changes. Aurelien Le 11 févr. 2014 à 18:40, Thomas Naughton <naughtont(a)ornl.gov> a écrit : > Hello, > > I read through the current proposal and fixed a few minor typos/wording > items. I don't have write access to the repo so attached is a diff of the > changes from my local copy. > > Also, in Example 17.3: > - I wasn't exactly sure what side effect the comment > before the MPI_Comm_failure_ack() was describing. > It might be good to improve that comment for clarity. > - Should the "T" variable passed to agree be re-initialized > inside the do-while loop? > > Otherwise, it looks good to me. > > Thanks, > --tjn > > _________________________________________________________________________ > Thomas Naughton naughtont(a)ornl.gov > Research Associate (865) 576-4184 > > > On Fri, 7 Feb 2014, Aurélien Bouteiller wrote: > >> Dear WG members, >> >> I would like to engage you into provisioning a little bit of proof reading time on monday/tuesday. >> >> We have reached semantic freeze immediately after the december meeting, and we are getting very close to freeze wording too at this point. The current plan is to have all WG edits done by tuesday evening, so that we accept only orthographic edits and comments from external reviewers until the final closure date in a week. >> >> Aurelien >> >> PS: if you do not have access yet to the working repository, kick me an email with your bitbucket ID. It is not planned to update the publicly visible draft before tuesday evening. >> >> >> _______________________________________________ >> mpiwg-ft mailing list >> mpiwg-ft(a)lists.mpi-forum.org >> http://lists.mpi-forum.org/mailman/listinfo.cgi/mpiwg-ft > <chap_ft_tjn_11feb2014.patch>_______________________________________________ > mpiwg-ft mailing list > mpiwg-ft(a)lists.mpi-forum.org > http://lists.mpi-forum.org/mailman/listinfo.cgi/mpiwg-ft

1 0

Re: [mpiwg-ft] Proof reading the draft
by Thomas Naughton 11 Feb '14

11 Feb '14

Hello, I read through the current proposal and fixed a few minor typos/wording items. I don't have write access to the repo so attached is a diff of the changes from my local copy. Also, in Example 17.3: - I wasn't exactly sure what side effect the comment before the MPI_Comm_failure_ack() was describing. It might be good to improve that comment for clarity. - Should the "T" variable passed to agree be re-initialized inside the do-while loop? Otherwise, it looks good to me. Thanks, --tjn _________________________________________________________________________ Thomas Naughton naughtont(a)ornl.gov Research Associate (865) 576-4184 On Fri, 7 Feb 2014, Aurï¿½lien Bouteiller wrote: > Dear WG members, > > I would like to engage you into provisioning a little bit of proof reading time on monday/tuesday. > > We have reached semantic freeze immediately after the december meeting, and we are getting very close to freeze wording too at this point. The current plan is to have all WG edits done by tuesday evening, so that we accept only orthographic edits and comments from external reviewers until the final closure date in a week. > > Aurelien > > PS: if you do not have access yet to the working repository, kick me an email with your bitbucket ID. It is not planned to update the publicly visible draft before tuesday evening. > > > _______________________________________________ > mpiwg-ft mailing list > mpiwg-ft(a)lists.mpi-forum.org > http://lists.mpi-forum.org/mailman/listinfo.cgi/mpiwg-ft >

1 0

Re: [mpiwg-ft] New slide material on ULFM
by Wesley Bland 09 Feb '14

09 Feb '14

Aurelien, Thanks for these slides. Can we get them in pptx (or whatever editable) format so we can merge these with our own slides? Thanks, Wesley

1 0

Proof reading the draft
by Aurélien Bouteiller 07 Feb '14

07 Feb '14

Dear WG members, I would like to engage you into provisioning a little bit of proof reading time on monday/tuesday. We have reached semantic freeze immediately after the december meeting, and we are getting very close to freeze wording too at this point. The current plan is to have all WG edits done by tuesday evening, so that we accept only orthographic edits and comments from external reviewers until the final closure date in a week. Aurelien PS: if you do not have access yet to the working repository, kick me an email with your bitbucket ID. It is not planned to update the publicly visible draft before tuesday evening.

1 0

New slide material on ULFM
by Aurélien Bouteiller 07 Feb '14

07 Feb '14

I presented the following deck of slides this morning. The content could be used to upgrade the WG deck of slides. Aurelien

1 0