February 2012 - mpi-forum - mailman3.cels.anl.gov

Re: [Mpi-forum] MPI_WAIT/MPI_TEST and generalized, I/O, and NBC requests
by Rolf Rabenseifner 13 Feb '12

13 Feb '12

> Non-blocking I/O requests weren't covered by the text in the 2.2 > standard, so it's not just generalized requests that were missed. Non-blocking I/O requests did not exist in MPI-2.2. They may come with MPI-3.0. Yes, your text is perfect. Do you file the ticket? Please use official MPI-2.2 page and line numbers, as in nearly all other tickets. I would say, it is enough when you copy the paragraphes mentioned in may email into the ticket. As Change-log, I would recommend: Section 3.7 on page 48.[[BR]] It is clarified that MPI_Wait and MPI_Test set the request handle to MPI_REQUEST_NULL if a generalized request is completed. Best regards Rolf ----- Original Message ----- > From: "Fab Tillier" <ftillier(a)microsoft.com> > To: "Main MPI Forum mailing list" <mpi-forum(a)lists.mpi-forum.org> > Sent: Monday, February 13, 2012 12:45:13 AM > Subject: Re: [Mpi-forum] MPI_WAIT/MPI_TEST and generalized, I/O, and NBC requests > > Non-blocking I/O requests weren't covered by the text in the 2.2 > standard, so it's not just generalized requests that were missed. > Other than that, you have it correct Rolf. I would word it slightly > differently, though, something along the lines of: > > "If the operation associated with this request was a persistent > communication operation, the persistent communication request is > marked as inactive. Other nonblocking operations are deallocated and > the request handle is set to MPI_REQUEST_NULL." > > Basically, mention the special case of persistent requests first, so > as to avoid the negative (non-persistent requests). > > -Fab > > Rolf Rabenseifner wrote on Sun, 12 Feb 2012 at 05:46:17 > > > When I understand correctly, then we have two problems: > > > > 1. MPI-2.2 has the "bug" that MPI-2.2 > > page 53 line 47 - page 54 line 3, and page 54 lines 40-45 > > do not mention the generalize requests although it was intended > > that MPI_Wait or MPI_Test of a generalized request acts > > as it would have been a isend or irecv request. > > 2. Nobody has checked for nonblocking collectives and > > nonblocking I/O that this text must now also include those > > routines. > > When I also understand correctly, then we currently say > > > > "If the communication object associated with this request > > was created by a nonblocking send or receive all, > > then ..." > > but we wanted to say > > > > "If the communication object associated with this request > > is not a persistent communication request (see Section 3.9 on > > page 69), > > then ..." > > Yes? > > > > This change would solve both problems. Yes? > > > > Best regards > > Rolf > > > > > > ----- Original Message ----- > >> From: "Bronis R. de Supinski" <bronis(a)llnl.gov> To: "Fab Tillier" > >> <ftillier(a)microsoft.com> Cc: "Main MPI Forum mailing list" > >> <mpi-forum(a)lists.mpi-forum.org> Sent: Saturday, February 11, 2012 > >> 7:57:30 PM Subject: Re: [Mpi-forum] MPI_WAIT/MPI_TEST and > >> generalized, > >> I/O, and NBC requests I will leave it to you to decide. I am sure I > >> will be very pressed for time... > >> > >> On Sat, 11 Feb 2012, Fab Tillier wrote: > >> > >>> I think it would make more sense to fix the text for the WAIT and > >>> TEST routines, rather than spread the information around. The > >>> generalized request section might not need changes if we do this. > >>> > >>> I don't know if I'll have time to put the text together before the > >>> meeting, though. > >>> > >>> -Fab > >>> > >>> Bronis R. de Supinski wrote on Sat, 11 Feb 2012 at 10:13:48 > >>> > >>>> > >>>> I am fine with the decision to make it the chapter committee > >>>> responsibility. I suppose that means I need to draft text > >>>> for the generalized requests? > >>>> > >>>> On Sat, 11 Feb 2012, William Gropp wrote: > >>>> > >>>>> It looks like this is a result of adding new request types > >>>>> without > >>>>> revisiting > >>>>> the original text. Under generalized requests, there is text > >>>>> that > >>>>> seems to > >>>>> imply that the request is set to MPI_REQUEST_NULL on completion, > >>>>> but it > >>>>> isn't explicit and should be. > >>>>> > >>>>> This is a chapter committee correction. > >>>>> > >>>>> Bill > >>>>> > >>>>> On Feb 10, 2012, at 6:04 PM, Fab Tillier wrote: > >>>>> > >>>>> Hi Folks, > >>>>> > >>>>> Looking at the text for MPI_WAIT (page 54, line 15) and MPI_TEST > >>>>> (page 55, > >>>>> line 7), it seems to imply that the request handle is only set > >>>>> to > >>>>> MPI_REQUEST_NULL for non-blocking send and receive requests. Are > >>>>> generalized, I/O, and NBC requests not completed the same way? > >>>>> Are > >>>>> users > >>>>> required to free such requests explicitly after they complete > >>>>> using > >>>>> MPI_REQUEST_FREE? > >>>>> > >>>>> I would have expected that all request handles except those to > >>>>> persistent > >>>>> requests are set to MPI_REQUEST_NULL, and persistent request > >>>>> just > >>>>> rather > >>>>> marked inactive. If that is indeed the intent, shouldn't we > >>>>> update > >>>>> the text to > >>>>> reflect this? Is this a ticket-0 level change? > >>>>> > >>>>> Thanks, > >>>>> -Fab > >>>>> > >>>>> > >>>>> _______________________________________________ mpi-forum > >>>>> mailing > >>>>> list mpi-forum(a)lists.mpi-forum.org<mailto:[email protected] > >>>>> forum.org> > >>>>> http://lists.mpi-forum.org/mailman/listinfo.cgi/mpi-forum > >>>>> > >>>>> William Gropp > >>>>> Director, Parallel Computing Institute > >>>>> Deputy Director for Research > >>>>> Institute for Advanced Computing Applications and Technologies > >>>>> Paul and Cynthia Saylor Professor of Computer Science > >>>>> University of Illinois Urbana-Champaign > >>>>> > >>>>> > >>>>> > >>>>> > >>>> _______________________________________________ > >>>> 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 > > > > > _______________________________________________ > mpi-forum mailing list > mpi-forum(a)lists.mpi-forum.org > http://lists.mpi-forum.org/mailman/listinfo.cgi/mpi-forum -- Dr. Rolf Rabenseifner . . . . . . . . . .. email rabenseifner(a)hlrs.de High Performance Computing Center (HLRS) . phone ++49(0)711/685-65530 University of Stuttgart . . . . . . . . .. fax ++49(0)711 / 685-65832 Head of Dpmt Parallel Computing . . . www.hlrs.de/people/rabenseifner Nobelstr. 19, D-70550 Stuttgart, Germany . (Office: Allmandring 30)

1 0

Re: [Mpi-forum] MPI_WAIT/MPI_TEST and generalized, I/O, and NBC requests
by Fab Tillier 12 Feb '12

12 Feb '12

Non-blocking I/O requests weren't covered by the text in the 2.2 standard, so it's not just generalized requests that were missed. Other than that, you have it correct Rolf. I would word it slightly differently, though, something along the lines of: "If the operation associated with this request was a persistent communication operation, the persistent communication request is marked as inactive. Other nonblocking operations are deallocated and the request handle is set to MPI_REQUEST_NULL." Basically, mention the special case of persistent requests first, so as to avoid the negative (non-persistent requests). -Fab Rolf Rabenseifner wrote on Sun, 12 Feb 2012 at 05:46:17 > When I understand correctly, then we have two problems: > > 1. MPI-2.2 has the "bug" that MPI-2.2 > page 53 line 47 - page 54 line 3, and page 54 lines 40-45 > do not mention the generalize requests although it was intended > that MPI_Wait or MPI_Test of a generalized request acts > as it would have been a isend or irecv request. > 2. Nobody has checked for nonblocking collectives and > nonblocking I/O that this text must now also include those routines. > When I also understand correctly, then we currently say > > "If the communication object associated with this request > was created by a nonblocking send or receive all, > then ..." > but we wanted to say > > "If the communication object associated with this request > is not a persistent communication request (see Section 3.9 on page 69), > then ..." > Yes? > > This change would solve both problems. Yes? > > Best regards > Rolf > > > ----- Original Message ----- >> From: "Bronis R. de Supinski" <bronis(a)llnl.gov> To: "Fab Tillier" >> <ftillier(a)microsoft.com> Cc: "Main MPI Forum mailing list" >> <mpi-forum(a)lists.mpi-forum.org> Sent: Saturday, February 11, 2012 >> 7:57:30 PM Subject: Re: [Mpi-forum] MPI_WAIT/MPI_TEST and generalized, >> I/O, and NBC requests I will leave it to you to decide. I am sure I >> will be very pressed for time... >> >> On Sat, 11 Feb 2012, Fab Tillier wrote: >> >>> I think it would make more sense to fix the text for the WAIT and >>> TEST routines, rather than spread the information around. The >>> generalized request section might not need changes if we do this. >>> >>> I don't know if I'll have time to put the text together before the >>> meeting, though. >>> >>> -Fab >>> >>> Bronis R. de Supinski wrote on Sat, 11 Feb 2012 at 10:13:48 >>> >>>> >>>> I am fine with the decision to make it the chapter committee >>>> responsibility. I suppose that means I need to draft text >>>> for the generalized requests? >>>> >>>> On Sat, 11 Feb 2012, William Gropp wrote: >>>> >>>>> It looks like this is a result of adding new request types without >>>>> revisiting >>>>> the original text. Under generalized requests, there is text that >>>>> seems to >>>>> imply that the request is set to MPI_REQUEST_NULL on completion, >>>>> but it >>>>> isn't explicit and should be. >>>>> >>>>> This is a chapter committee correction. >>>>> >>>>> Bill >>>>> >>>>> On Feb 10, 2012, at 6:04 PM, Fab Tillier wrote: >>>>> >>>>> Hi Folks, >>>>> >>>>> Looking at the text for MPI_WAIT (page 54, line 15) and MPI_TEST >>>>> (page 55, >>>>> line 7), it seems to imply that the request handle is only set to >>>>> MPI_REQUEST_NULL for non-blocking send and receive requests. Are >>>>> generalized, I/O, and NBC requests not completed the same way? Are >>>>> users >>>>> required to free such requests explicitly after they complete >>>>> using >>>>> MPI_REQUEST_FREE? >>>>> >>>>> I would have expected that all request handles except those to >>>>> persistent >>>>> requests are set to MPI_REQUEST_NULL, and persistent request just >>>>> rather >>>>> marked inactive. If that is indeed the intent, shouldn't we update >>>>> the text to >>>>> reflect this? Is this a ticket-0 level change? >>>>> >>>>> Thanks, >>>>> -Fab >>>>> >>>>> >>>>> _______________________________________________ mpi-forum mailing >>>>> list mpi-forum(a)lists.mpi-forum.org<mailto:[email protected] >>>>> forum.org> http://lists.mpi-forum.org/mailman/listinfo.cgi/mpi-forum >>>>> >>>>> William Gropp >>>>> Director, Parallel Computing Institute >>>>> Deputy Director for Research >>>>> Institute for Advanced Computing Applications and Technologies >>>>> Paul and Cynthia Saylor Professor of Computer Science >>>>> University of Illinois Urbana-Champaign >>>>> >>>>> >>>>> >>>>> >>>> _______________________________________________ >>>> 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 >

1 0

Re: [Mpi-forum] March MPI Forum meeting
by Graham, Richard L. 12 Feb '12

12 Feb '12

Thanks Rolf, I updated most of the requests, but did not move things around like you asked. Also, ticket 264 is being delayed by one meeting. All - please check and make sure that I am not missing something else. Thanks again, Rich On Jan 31, 2012, at 6:08 PM, Rolf Rabenseifner wrote: > Hi Rich, > > please can you do a few changes/corrections... to the agenda: > (Nothing spectacular - only small but needed things) > > - Monday 1:00pm - 6:00pm Plenary Session: Fault Tolerance working group - Formal reading - Ticket #276 > > you should add #291 (it is related to FT an otherwise it will > have no chance because it is not announced for formal reading) > > - At the end of Monday, please add > > At any free time Plenary Session: Formal reading > Use of [] in output arrays ticket #125 > Consistent use of [] for input arrays ticket #126 > "same arguments" to MPI_Reduce ticket #218 > > - Tuesday 9:00-9:30am: please extend to 45 minutes (I add 3 further tickets) > and add the following ticket numbers: > #185 #207 #215 #254 #255 #262 #263 #267 #268 #312 (from Rolf's slide 9) > #162 #182 #219 (from Rolf's slide 11) > > - Tuesday 9:30-10:00am moves to 9:45-10:15 > - Tuesday 10-10:15 is a 1st (!) vote and should go to the votes session > after the break. It is a trivial correction to NBC and will not have > any discussion. > > - Tuesday 10:20-11:30am Votes session > 2nd (!! not 1st) votes are #299 and #229 > In continues with 1st votes: > - starting with #272 (moved from 10:00 slot) > - and adding at the end: > MPI_Icomm_dup ticket #168 > Please move #294 (Goodell/Rabenseifner) at the end of 12:30-4:30pm slot. > Please add the following additional vote: > "One Single vote together *) for tickets > #185 #207 #215 #254 #255 #262 #263 #267 #268 #312 #162 #182 #219 > *) If somebody complains now or in the meeting, then single voting > can be done for specific tickets or voting can be done on the next day." > > - Tuesday 12:30-4:30pm: > - Please add after #305 a new ticket-line > Scalable irregular collectives (Hoefler) ticket #264 > - and at the end #294 from above: > MPI_UNWEIGHTED should not be NULL (Goodell/Rabenseifner) ticket #294 > > - Tuesday 4:45-6:30pm: > - 158: (Traeff/Rabenseifner) > - 187: (Ritzdorf/Rabenseifner) > > - Tuesday 7-9pm: > - MPI_IN_PLACE in Gather (Hoefler) ticket #159 > has wrong ticket number: it is #163 > > I also updated my slides based on your agenda (together with these changes). > It is attached. > > Best regards > Rolf > > > ----- Original Message ----- >> From: "Richard L. Graham" <rlgraham(a)ornl.gov> >> To: "Main MPI Forum mailing list" <mpi-forum(a)lists.mpi-forum.org> >> Sent: Monday, January 30, 2012 2:35:57 AM >> Subject: [Mpi-forum] March MPI Forum meeting >> Registration is up for the meeting, as is the agenda. Please check out >> the agenda - Rolf has done a lot of work to figure out the open 3.0 >> tickets. For those that have already passe a reading, there is no >> issue - voting will progress. On the other ones, I listed the ticket >> owner for all of these, to make sure that folks are aware that their >> tickets are up for discussion. If folks do not want to proceed with >> these, please let me know. Those that are on the hook are Koziel, >> Gropp, Hoefler, Dinan, Tillier, Snir, Squyres, and Moody - please >> confirm that you are OK with these tickets. Also, Traff and Ritzdorf >> are owners of two tickets, so someone else will need to move these >> along. >> >> Rich >> _______________________________________________ >> mpi-forum mailing list >> mpi-forum(a)lists.mpi-forum.org >> http://lists.mpi-forum.org/mailman/listinfo.cgi/mpi-forum > > -- > Dr. Rolf Rabenseifner . . . . . . . . . .. email rabenseifner(a)hlrs.de > High Performance Computing Center (HLRS) . phone ++49(0)711/685-65530 > University of Stuttgart . . . . . . . . .. fax ++49(0)711 / 685-65832 > Head of Dpmt Parallel Computing . . . www.hlrs.de/people/rabenseifner > Nobelstr. 19, D-70550 Stuttgart, Germany . (Office: Allmandring 30)<MPI_Forum_Overview_MPI-3.0_Mar2012_action-items.ppt>_______________________________________________ > 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] MPI_WAIT/MPI_TEST and generalized, I/O, and NBC requests
by William Gropp 12 Feb '12

12 Feb '12

Yes, I believe this change would match the intent, and would work with, for example, the RMA requests that haven't been mentioned yet :) Bill On Feb 12, 2012, at 7:46 AM, Rolf Rabenseifner wrote: > When I understand correctly, then we have two problems: > > 1. MPI-2.2 has the "bug" that MPI-2.2 > page 53 line 47 - page 54 line 3, and page 54 lines 40-45 > do not mention the generalize requests although it was intended > that MPI_Wait or MPI_Test of a generalized request acts > as it would have been a isend or irecv request. > > 2. Nobody has checked for nonblocking collectives and > nonblocking I/O that this text must now also include those routines. > > When I also understand correctly, then we currently say > > "If the communication object associated with this request > was created by a nonblocking send or receive all, > then ..." > > but we wanted to say > > "If the communication object associated with this request > is not a persistent communication request (see Section 3.9 on page 69), > then ..." > > Yes? > > This change would solve both problems. Yes? > > Best regards > Rolf > > > ----- Original Message ----- >> From: "Bronis R. de Supinski" <bronis(a)llnl.gov> >> To: "Fab Tillier" <ftillier(a)microsoft.com> >> Cc: "Main MPI Forum mailing list" <mpi-forum(a)lists.mpi-forum.org> >> Sent: Saturday, February 11, 2012 7:57:30 PM >> Subject: Re: [Mpi-forum] MPI_WAIT/MPI_TEST and generalized, I/O, and NBC requests >> I will leave it to you to decide. I am sure I >> will be very pressed for time... >> >> On Sat, 11 Feb 2012, Fab Tillier wrote: >> >>> I think it would make more sense to fix the text for the WAIT and >>> TEST routines, rather than spread the information around. The >>> generalized request section might not need changes if we do this. >>> >>> I don't know if I'll have time to put the text together before the >>> meeting, though. >>> >>> -Fab >>> >>> Bronis R. de Supinski wrote on Sat, 11 Feb 2012 at 10:13:48 >>> >>>> >>>> I am fine with the decision to make it the chapter committee >>>> responsibility. I suppose that means I need to draft text >>>> for the generalized requests? >>>> >>>> On Sat, 11 Feb 2012, William Gropp wrote: >>>> >>>>> It looks like this is a result of adding new request types without >>>>> revisiting >>>>> the original text. Under generalized requests, there is text that >>>>> seems to >>>>> imply that the request is set to MPI_REQUEST_NULL on completion, >>>>> but it >>>>> isn't explicit and should be. >>>>> >>>>> This is a chapter committee correction. >>>>> >>>>> Bill >>>>> >>>>> On Feb 10, 2012, at 6:04 PM, Fab Tillier wrote: >>>>> >>>>> Hi Folks, >>>>> >>>>> Looking at the text for MPI_WAIT (page 54, line 15) and MPI_TEST >>>>> (page 55, >>>>> line 7), it seems to imply that the request handle is only set to >>>>> MPI_REQUEST_NULL for non-blocking send and receive requests. Are >>>>> generalized, I/O, and NBC requests not completed the same way? Are >>>>> users >>>>> required to free such requests explicitly after they complete >>>>> using >>>>> MPI_REQUEST_FREE? >>>>> >>>>> I would have expected that all request handles except those to >>>>> persistent >>>>> requests are set to MPI_REQUEST_NULL, and persistent request just >>>>> rather >>>>> marked inactive. If that is indeed the intent, shouldn't we update >>>>> the text to >>>>> reflect this? Is this a ticket-0 level change? >>>>> >>>>> Thanks, >>>>> -Fab >>>>> >>>>> >>>>> _______________________________________________ >>>>> mpi-forum mailing list >>>>> mpi-forum(a)lists.mpi-forum.org<mailto:[email protected]> >>>>> http://lists.mpi-forum.org/mailman/listinfo.cgi/mpi-forum >>>>> >>>>> William Gropp >>>>> Director, Parallel Computing Institute >>>>> Deputy Director for Research >>>>> Institute for Advanced Computing Applications and Technologies >>>>> Paul and Cynthia Saylor Professor of Computer Science >>>>> University of Illinois Urbana-Champaign >>>>> >>>>> >>>>> >>>>> >>>> _______________________________________________ >>>> 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 > > -- > Dr. Rolf Rabenseifner . . . . . . . . . .. email rabenseifner(a)hlrs.de > High Performance Computing Center (HLRS) . phone ++49(0)711/685-65530 > University of Stuttgart . . . . . . . . .. fax ++49(0)711 / 685-65832 > Head of Dpmt Parallel Computing . . . www.hlrs.de/people/rabenseifner > Nobelstr. 19, D-70550 Stuttgart, Germany . (Office: Allmandring 30) > _______________________________________________ > mpi-forum mailing list > mpi-forum(a)lists.mpi-forum.org > http://lists.mpi-forum.org/mailman/listinfo.cgi/mpi-forum William Gropp Director, Parallel Computing Institute Deputy Director for Research Institute for Advanced Computing Applications and Technologies Paul and Cynthia Saylor Professor of Computer Science University of Illinois Urbana-Champaign

1 0

Re: [Mpi-forum] MPI_WAIT/MPI_TEST and generalized, I/O, and NBC requests
by Rolf Rabenseifner 12 Feb '12

12 Feb '12

When I understand correctly, then we have two problems: 1. MPI-2.2 has the "bug" that MPI-2.2 page 53 line 47 - page 54 line 3, and page 54 lines 40-45 do not mention the generalize requests although it was intended that MPI_Wait or MPI_Test of a generalized request acts as it would have been a isend or irecv request. 2. Nobody has checked for nonblocking collectives and nonblocking I/O that this text must now also include those routines. When I also understand correctly, then we currently say "If the communication object associated with this request was created by a nonblocking send or receive all, then ..." but we wanted to say "If the communication object associated with this request is not a persistent communication request (see Section 3.9 on page 69), then ..." Yes? This change would solve both problems. Yes? Best regards Rolf ----- Original Message ----- > From: "Bronis R. de Supinski" <bronis(a)llnl.gov> > To: "Fab Tillier" <ftillier(a)microsoft.com> > Cc: "Main MPI Forum mailing list" <mpi-forum(a)lists.mpi-forum.org> > Sent: Saturday, February 11, 2012 7:57:30 PM > Subject: Re: [Mpi-forum] MPI_WAIT/MPI_TEST and generalized, I/O, and NBC requests > I will leave it to you to decide. I am sure I > will be very pressed for time... > > On Sat, 11 Feb 2012, Fab Tillier wrote: > > > I think it would make more sense to fix the text for the WAIT and > > TEST routines, rather than spread the information around. The > > generalized request section might not need changes if we do this. > > > > I don't know if I'll have time to put the text together before the > > meeting, though. > > > > -Fab > > > > Bronis R. de Supinski wrote on Sat, 11 Feb 2012 at 10:13:48 > > > >> > >> I am fine with the decision to make it the chapter committee > >> responsibility. I suppose that means I need to draft text > >> for the generalized requests? > >> > >> On Sat, 11 Feb 2012, William Gropp wrote: > >> > >>> It looks like this is a result of adding new request types without > >>> revisiting > >>> the original text. Under generalized requests, there is text that > >>> seems to > >>> imply that the request is set to MPI_REQUEST_NULL on completion, > >>> but it > >>> isn't explicit and should be. > >>> > >>> This is a chapter committee correction. > >>> > >>> Bill > >>> > >>> On Feb 10, 2012, at 6:04 PM, Fab Tillier wrote: > >>> > >>> Hi Folks, > >>> > >>> Looking at the text for MPI_WAIT (page 54, line 15) and MPI_TEST > >>> (page 55, > >>> line 7), it seems to imply that the request handle is only set to > >>> MPI_REQUEST_NULL for non-blocking send and receive requests. Are > >>> generalized, I/O, and NBC requests not completed the same way? Are > >>> users > >>> required to free such requests explicitly after they complete > >>> using > >>> MPI_REQUEST_FREE? > >>> > >>> I would have expected that all request handles except those to > >>> persistent > >>> requests are set to MPI_REQUEST_NULL, and persistent request just > >>> rather > >>> marked inactive. If that is indeed the intent, shouldn't we update > >>> the text to > >>> reflect this? Is this a ticket-0 level change? > >>> > >>> Thanks, > >>> -Fab > >>> > >>> > >>> _______________________________________________ > >>> mpi-forum mailing list > >>> mpi-forum(a)lists.mpi-forum.org<mailto:[email protected]> > >>> http://lists.mpi-forum.org/mailman/listinfo.cgi/mpi-forum > >>> > >>> William Gropp > >>> Director, Parallel Computing Institute > >>> Deputy Director for Research > >>> Institute for Advanced Computing Applications and Technologies > >>> Paul and Cynthia Saylor Professor of Computer Science > >>> University of Illinois Urbana-Champaign > >>> > >>> > >>> > >>> > >> _______________________________________________ > >> 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 -- Dr. Rolf Rabenseifner . . . . . . . . . .. email rabenseifner(a)hlrs.de High Performance Computing Center (HLRS) . phone ++49(0)711/685-65530 University of Stuttgart . . . . . . . . .. fax ++49(0)711 / 685-65832 Head of Dpmt Parallel Computing . . . www.hlrs.de/people/rabenseifner Nobelstr. 19, D-70550 Stuttgart, Germany . (Office: Allmandring 30)

1 0

Re: [Mpi-forum] MPI_WAIT/MPI_TEST and generalized, I/O, and NBC requests
by Bronis R. de Supinski 11 Feb '12

11 Feb '12

I will leave it to you to decide. I am sure I will be very pressed for time... On Sat, 11 Feb 2012, Fab Tillier wrote: > I think it would make more sense to fix the text for the WAIT and TEST routines, rather than spread the information around. The generalized request section might not need changes if we do this. > > I don't know if I'll have time to put the text together before the meeting, though. > > -Fab > > Bronis R. de Supinski wrote on Sat, 11 Feb 2012 at 10:13:48 > >> >> I am fine with the decision to make it the chapter committee >> responsibility. I suppose that means I need to draft text >> for the generalized requests? >> >> On Sat, 11 Feb 2012, William Gropp wrote: >> >>> It looks like this is a result of adding new request types without revisiting >>> the original text. Under generalized requests, there is text that seems to >>> imply that the request is set to MPI_REQUEST_NULL on completion, but it >>> isn't explicit and should be. >>> >>> This is a chapter committee correction. >>> >>> Bill >>> >>> On Feb 10, 2012, at 6:04 PM, Fab Tillier wrote: >>> >>> Hi Folks, >>> >>> Looking at the text for MPI_WAIT (page 54, line 15) and MPI_TEST (page 55, >>> line 7), it seems to imply that the request handle is only set to >>> MPI_REQUEST_NULL for non-blocking send and receive requests. Are >>> generalized, I/O, and NBC requests not completed the same way? Are users >>> required to free such requests explicitly after they complete using >>> MPI_REQUEST_FREE? >>> >>> I would have expected that all request handles except those to persistent >>> requests are set to MPI_REQUEST_NULL, and persistent request just rather >>> marked inactive. If that is indeed the intent, shouldn't we update the text to >>> reflect this? Is this a ticket-0 level change? >>> >>> Thanks, >>> -Fab >>> >>> >>> _______________________________________________ >>> mpi-forum mailing list >>> mpi-forum(a)lists.mpi-forum.org<mailto:[email protected]> >>> http://lists.mpi-forum.org/mailman/listinfo.cgi/mpi-forum >>> >>> William Gropp >>> Director, Parallel Computing Institute >>> Deputy Director for Research >>> Institute for Advanced Computing Applications and Technologies >>> Paul and Cynthia Saylor Professor of Computer Science >>> University of Illinois Urbana-Champaign >>> >>> >>> >>> >> _______________________________________________ >> 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] MPI_WAIT/MPI_TEST and generalized, I/O, and NBC requests
by Fab Tillier 11 Feb '12

11 Feb '12

I think it would make more sense to fix the text for the WAIT and TEST routines, rather than spread the information around. The generalized request section might not need changes if we do this. I don't know if I'll have time to put the text together before the meeting, though. -Fab Bronis R. de Supinski wrote on Sat, 11 Feb 2012 at 10:13:48 > > I am fine with the decision to make it the chapter committee > responsibility. I suppose that means I need to draft text > for the generalized requests? > > On Sat, 11 Feb 2012, William Gropp wrote: > >> It looks like this is a result of adding new request types without revisiting >> the original text. Under generalized requests, there is text that seems to >> imply that the request is set to MPI_REQUEST_NULL on completion, but it >> isn't explicit and should be. >> >> This is a chapter committee correction. >> >> Bill >> >> On Feb 10, 2012, at 6:04 PM, Fab Tillier wrote: >> >> Hi Folks, >> >> Looking at the text for MPI_WAIT (page 54, line 15) and MPI_TEST (page 55, >> line 7), it seems to imply that the request handle is only set to >> MPI_REQUEST_NULL for non-blocking send and receive requests. Are >> generalized, I/O, and NBC requests not completed the same way? Are users >> required to free such requests explicitly after they complete using >> MPI_REQUEST_FREE? >> >> I would have expected that all request handles except those to persistent >> requests are set to MPI_REQUEST_NULL, and persistent request just rather >> marked inactive. If that is indeed the intent, shouldn't we update the text to >> reflect this? Is this a ticket-0 level change? >> >> Thanks, >> -Fab >> >> >> _______________________________________________ >> mpi-forum mailing list >> mpi-forum(a)lists.mpi-forum.org<mailto:[email protected]> >> http://lists.mpi-forum.org/mailman/listinfo.cgi/mpi-forum >> >> William Gropp >> Director, Parallel Computing Institute >> Deputy Director for Research >> Institute for Advanced Computing Applications and Technologies >> Paul and Cynthia Saylor Professor of Computer Science >> University of Illinois Urbana-Champaign >> >> >> >> > _______________________________________________ > 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] MPI_WAIT/MPI_TEST and generalized, I/O, and NBC requests
by Bronis R. de Supinski 11 Feb '12

11 Feb '12

I am fine with the decision to make it the chapter committee responsibility. I suppose that means I need to draft text for the generalized requests? On Sat, 11 Feb 2012, William Gropp wrote: > It looks like this is a result of adding new request types without revisiting the original text. Under generalized requests, there is text that seems to imply that the request is set to MPI_REQUEST_NULL on completion, but it isn't explicit and should be. > > This is a chapter committee correction. > > Bill > > On Feb 10, 2012, at 6:04 PM, Fab Tillier wrote: > > Hi Folks, > > Looking at the text for MPI_WAIT (page 54, line 15) and MPI_TEST (page 55, line 7), it seems to imply that the request handle is only set to MPI_REQUEST_NULL for non-blocking send and receive requests. Are generalized, I/O, and NBC requests not completed the same way? Are users required to free such requests explicitly after they complete using MPI_REQUEST_FREE? > > I would have expected that all request handles except those to persistent requests are set to MPI_REQUEST_NULL, and persistent request just rather marked inactive. If that is indeed the intent, shouldn't we update the text to reflect this? Is this a ticket-0 level change? > > Thanks, > -Fab > > > _______________________________________________ > mpi-forum mailing list > mpi-forum(a)lists.mpi-forum.org<mailto:[email protected]> > http://lists.mpi-forum.org/mailman/listinfo.cgi/mpi-forum > > William Gropp > Director, Parallel Computing Institute > Deputy Director for Research > Institute for Advanced Computing Applications and Technologies > Paul and Cynthia Saylor Professor of Computer Science > University of Illinois Urbana-Champaign > > > >

1 0

Re: [Mpi-forum] MPI_WAIT/MPI_TEST and generalized, I/O, and NBC requests
by William Gropp 11 Feb '12

11 Feb '12

It looks like this is a result of adding new request types without revisiting the original text. Under generalized requests, there is text that seems to imply that the request is set to MPI_REQUEST_NULL on completion, but it isn't explicit and should be. This is a chapter committee correction. Bill On Feb 10, 2012, at 6:04 PM, Fab Tillier wrote: > Hi Folks, > > Looking at the text for MPI_WAIT (page 54, line 15) and MPI_TEST (page 55, line 7), it seems to imply that the request handle is only set to MPI_REQUEST_NULL for non-blocking send and receive requests. Are generalized, I/O, and NBC requests not completed the same way? Are users required to free such requests explicitly after they complete using MPI_REQUEST_FREE? > > I would have expected that all request handles except those to persistent requests are set to MPI_REQUEST_NULL, and persistent request just rather marked inactive. If that is indeed the intent, shouldn't we update the text to reflect this? Is this a ticket-0 level change? > > Thanks, > -Fab > > > _______________________________________________ > mpi-forum mailing list > mpi-forum(a)lists.mpi-forum.org > http://lists.mpi-forum.org/mailman/listinfo.cgi/mpi-forum William Gropp Director, Parallel Computing Institute Deputy Director for Research Institute for Advanced Computing Applications and Technologies Paul and Cynthia Saylor Professor of Computer Science University of Illinois Urbana-Champaign

1 0

Re: [Mpi-forum] MPI_WAIT/MPI_TEST and generalized, I/O, and NBC requests
by Torsten Hoefler 11 Feb '12

11 Feb '12

Fab, > Looking at the text for MPI_WAIT (page 54, line 15) and MPI_TEST (page > 55, line 7), it seems to imply that the request handle is only set to > MPI_REQUEST_NULL for non-blocking send and receive requests. Are > generalized, I/O, and NBC requests not completed the same way? Are > users required to free such requests explicitly after they complete > using MPI_REQUEST_FREE? Yes, this is a minor bug. It should probably say "communication function" instead of "send or receive call". > I would have expected that all request handles except those to > persistent requests are set to MPI_REQUEST_NULL, and persistent > request just rather marked inactive. If that is indeed the intent, > shouldn't we update the text to reflect this? Yes we should. > Is this a ticket-0 level change? I think it's an 0.5 change and should be handled by the p2p chapter committee and pushed into MPI-3.0. We can also still go through the ticket process but that may not be the best investment of our time. All the Best, Torsten -- bash$ :(){ :|:&};: --------------------- http://www.unixer.de/ ----- Torsten Hoefler | Performance Modeling and Simulation Lead Blue Waters Directorate | University of Illinois (UIUC) 1205 W Clark Street | Urbana, IL, 61801 NCSA Building | +01 (217) 244-7736

1 0