August 2021 - mpiwg-tools - mailman3.cels.anl.gov

QMPI Slides
by Wesley Bland 31 Aug '21

31 Aug '21

Hi all, I finished up some slides for the QMPI plenary. They’re attached here. Marc-Andre, I’m not sure if you wanted me to present these or if you’d like to take it from here since you’re the "big man in charge”. Totally up to you. Either way, if anyone has any comments/edits, let me know and I can update these. Thanks, Wes

1 0

Re: [mpiwg-tools] QMPI at September Meeting
by Wesley Bland 30 Aug '21

30 Aug '21

Yep. Sounds good to me. > On Aug 30, 2021, at 2:12 PM, Marc-André Hermanns via mpiwg-tools <mpiwg-tools(a)lists.mpi-forum.org> wrote: > > Hi Wes, > > yes, some plenary time would be nice for this. > > Last WG meeting Martin, Bill, and I also discussed two new tickets 519 > & 520 that we also wanted to have some plenary feedback on (to check > whether this is worth continuing). > > Should I send out an announcement of those three topics? > > Cheers, > Marc-Andre > > On 30.08.21 21:06, Wesley Bland via mpiwg-tools wrote: >> I was just looking at my todo list and noticed that I made a note to create a QMPI slide deck for the MPI Forum meeting next week. I’d forgotten about this, but putting it together shouldn’t be too hard, but nothing was announced. Since this isn’t a reading, that’s not a problem, but I wanted to confirm that we still was to do this. Marc-Andre, do you want to announce the plenary? >> >> Thanks, >> Wes >> _______________________________________________ >> mpiwg-tools mailing list >> mpiwg-tools(a)lists.mpi-forum.org >> https://lists.mpi-forum.org/mailman/listinfo/mpiwg-tools >> > > -- > Dr. rer. nat. Marc-André Hermanns > > IT Center > Group: High Performance Computing > Division: Computational Science and Engineering > RWTH Aachen University > Seffenter Weg 23 > 52074 Aachen > Phone: +49 241 80-24381 > hermanns(a)itc.rwth-aachen.de > www.itc.rwth-aachen.de > > Social Media Kanäle des IT Centers: > https://blog.rwth-aachen.de/itc/ > https://www.facebook.com/itcenterrwth > https://www.linkedin.com/company/itcenterrwth > https://twitter.com/ITCenterRWTH > https://www.youtube.com/channel/UCKKDJJukeRwO0LP-ac8x8rQ > > _______________________________________________ > mpiwg-tools mailing list > mpiwg-tools(a)lists.mpi-forum.org > https://lists.mpi-forum.org/mailman/listinfo/mpiwg-tools

1 0

Re: [mpiwg-tools] QMPI at September Meeting
by Marc-André Hermanns 30 Aug '21

30 Aug '21

Hi Wes, yes, some plenary time would be nice for this. Last WG meeting Martin, Bill, and I also discussed two new tickets 519 & 520 that we also wanted to have some plenary feedback on (to check whether this is worth continuing). Should I send out an announcement of those three topics? Cheers, Marc-Andre On 30.08.21 21:06, Wesley Bland via mpiwg-tools wrote: > I was just looking at my todo list and noticed that I made a note to create a QMPI slide deck for the MPI Forum meeting next week. I’d forgotten about this, but putting it together shouldn’t be too hard, but nothing was announced. Since this isn’t a reading, that’s not a problem, but I wanted to confirm that we still was to do this. Marc-Andre, do you want to announce the plenary? > > Thanks, > Wes > _______________________________________________ > mpiwg-tools mailing list > mpiwg-tools(a)lists.mpi-forum.org > https://lists.mpi-forum.org/mailman/listinfo/mpiwg-tools > -- Dr. rer. nat. Marc-André Hermanns IT Center Group: High Performance Computing Division: Computational Science and Engineering RWTH Aachen University Seffenter Weg 23 52074 Aachen Phone: +49 241 80-24381 hermanns(a)itc.rwth-aachen.de www.itc.rwth-aachen.de Social Media Kanäle des IT Centers: https://blog.rwth-aachen.de/itc/ https://www.facebook.com/itcenterrwth https://www.linkedin.com/company/itcenterrwth https://twitter.com/ITCenterRWTH https://www.youtube.com/channel/UCKKDJJukeRwO0LP-ac8x8rQ

1 0

QMPI at September Meeting
by Wesley Bland 30 Aug '21

30 Aug '21

I was just looking at my todo list and noticed that I made a note to create a QMPI slide deck for the MPI Forum meeting next week. I’d forgotten about this, but putting it together shouldn’t be too hard, but nothing was announced. Since this isn’t a reading, that’s not a problem, but I wanted to confirm that we still was to do this. Marc-Andre, do you want to announce the plenary? Thanks, Wes

1 0

Meeting reminder: August 26, 2021 (today)
by Marc-André Hermanns 26 Aug '21

26 Aug '21

Hi all, This is a reminder of our regularly scheduled call today: August 26, 2021 at 8 am Pacific / 11 am Eastern / 5pm CEST. Meeting Details: https://github.com/mpiwg-tools/mpi-standard/wiki/Meeting-Information Agenda: ------- - Preparation/Discussion of MPI_Request_get_status_(any|some|all) - Adding MPI_LONG_LONG_INT or MPI_UINT64_INT for reductions (also for Accumulate operations) Cheers, Marc-Andre -- Dr. rer. nat. Marc-André Hermanns IT Center Group: High Performance Computing Division: Computational Science and Engineering RWTH Aachen University Seffenter Weg 23 52074 Aachen Phone: +49 241 80-24381 hermanns(a)itc.rwth-aachen.de www.itc.rwth-aachen.de Social Media Kanäle des IT Centers: https://blog.rwth-aachen.de/itc/ https://www.facebook.com/itcenterrwth https://www.linkedin.com/company/itcenterrwth https://twitter.com/ITCenterRWTH https://www.youtube.com/channel/UCKKDJJukeRwO0LP-ac8x8rQ

1 0

Re: [mpiwg-tools] Tool question about handling requests in completion calls
by Joachim Protze 20 Aug '21

20 Aug '21

Hi all, Marc-Andre pointed me to MPI_Request_get_status during yesterday's call. The plan for next week's call is to talk about any/some/all versions of this call. Here is how I would use the function in test calls: int MPI_Test(MPI_Request* req, int* flag, MPI_Status* status){ int ret = PMPI_Request_get_status(*req, flag, status); if(*flag){ // tool code to handle the successful test PMPI_Test(req, flag, MPI_STATUS_IGNORE); } return ret; } int MPI_Testsome(int incount, MPI_Request req[], int* outcount, int* indices, MPI_Status* statuses[]){ int ret = PMPI_Request_get_status_some(incount, req, outcount, indices, statuses), flag; for(int i=0; i < *outcount; i++){ // tool code to handle the successful test on req[indices[i]] PMPI_Test(req[indices[i]], &flag, MPI_STATUS_IGNORE); } return ret; } int MPI_Testany(int count, MPI_Request req[], int* index, int* flag, MPI_Status* status){ int ret = PMPI_Request_get_status_any(count, req, index, flag, statuses); if(flag){ // tool code to handle the successful test on req[index] PMPI_Test(req[index], &flag, MPI_STATUS_IGNORE); } return ret; } int MPI_Testall(int count, MPI_Request req[], int* flag, MPI_Status* statuses[]){ int ret = PMPI_Request_get_status_all(count, req, flag, statuses); if(flag){ // tool code to handle the successful test on all req PMPI_Testall(req, &flag, MPI_STATUSES_IGNORE); } return ret; } However, I'm not convinced that I'll like to use a similar approach for MPI_Waitall: int MPI_Waitall(int count, MPI_Request req[], MPI_Status* statuses[]){ int flag; do{ int ret = PMPI_Request_get_status_all(incount, req, &flag, statuses); } while(!flag); // tool code to handle the successful test on all req PMPI_Testall(req, &flag, MPI_STATUSES_IGNORE); return ret; } Best Joachim Am 12.08.21 um 16:23 schrieb Joachim Protze via mpiwg-tools: > Hi all, > > working on a tool, I'm wondering whether someone has a better solution > than copying the whole array before calling the PMPI function. > > In the tool, we want to observe the lifetime of requests and > semantically bind information to the request handle (e.g. > std::map<MPI_Request, tool_info*>). > When we observe completion of the request, the request handle passed by > the application is already reset to MPI_REQUEST_NULL and cannot be used > for the lookup. So, we need to copy the whole array of requests in order > to lookup the information for the request after the PMPI call. > > For a call like MPI_Testany, this means to copy /count/ values to > possibly use one of the values. > > Is there any chance to improve the situation? Would it be possible to > bind the request handle to the status object for reference? > It is clear that the request handle is invalid at that moment, but could > still be used for some lookup. > > Alternatively, it would help to bind an info(?) object to the request > handle, which moves to the status on completion. > > Best > Joachim > > > > _______________________________________________ > mpiwg-tools mailing list > mpiwg-tools(a)lists.mpi-forum.org > https://lists.mpi-forum.org/mailman/listinfo/mpiwg-tools > -- Dipl.-Inf. Joachim Protze IT Center Group: High Performance Computing Division: Computational Science and Engineering RWTH Aachen University Seffenter Weg 23 D 52074 Aachen (Germany) Tel: +49 241 80- 24765 Fax: +49 241 80-624765 protze(a)itc.rwth-aachen.de www.itc.rwth-aachen.de

1 0

Meeting reminder/poll: August 19, 2021 (today)
by Marc-André Hermanns 19 Aug '21

19 Aug '21

Hi all, This is a reminder of our regularly scheduled call today: August 19, 2021 at 8 am Pacific / 11 am Eastern / 5pm CEST. Meeting Details: https://github.com/mpiwg-tools/mpi-standard/wiki/Meeting-Information Agenda: ------- - QMPI Cheers, Marc-Andre -- Dr. rer. nat. Marc-André Hermanns IT Center Group: High Performance Computing Division: Computational Science and Engineering RWTH Aachen University Seffenter Weg 23 52074 Aachen Phone: +49 241 80-24381 hermanns(a)itc.rwth-aachen.de www.itc.rwth-aachen.de Social Media Kanäle des IT Centers: https://blog.rwth-aachen.de/itc/ https://www.facebook.com/itcenterrwth https://www.linkedin.com/company/itcenterrwth https://twitter.com/ITCenterRWTH https://www.youtube.com/channel/UCKKDJJukeRwO0LP-ac8x8rQ

1 0

Tool question about handling requests in completion calls
by Joachim Protze 12 Aug '21

12 Aug '21

Hi all, working on a tool, I'm wondering whether someone has a better solution than copying the whole array before calling the PMPI function. In the tool, we want to observe the lifetime of requests and semantically bind information to the request handle (e.g. std::map<MPI_Request, tool_info*>). When we observe completion of the request, the request handle passed by the application is already reset to MPI_REQUEST_NULL and cannot be used for the lookup. So, we need to copy the whole array of requests in order to lookup the information for the request after the PMPI call. For a call like MPI_Testany, this means to copy /count/ values to possibly use one of the values. Is there any chance to improve the situation? Would it be possible to bind the request handle to the status object for reference? It is clear that the request handle is invalid at that moment, but could still be used for some lookup. Alternatively, it would help to bind an info(?) object to the request handle, which moves to the status on completion. Best Joachim -- Dipl.-Inf. Joachim Protze IT Center Group: High Performance Computing Division: Computational Science and Engineering RWTH Aachen University Seffenter Weg 23 D 52074 Aachen (Germany) Tel: +49 241 80- 24765 Fax: +49 241 80-624765 protze(a)itc.rwth-aachen.de www.itc.rwth-aachen.de

1 0

Re: [mpiwg-tools] Meeting reminder August 12, 2021 (today)
by Joachim Protze 12 Aug '21

12 Aug '21

Ok, we have at least 3 participants so let's meet and see how far we get with reading the QMPI PR. Since first John and then I'm on vacation the next weeks could be used for QMPI depending on the availability of people. - Joachim Am 10.08.21 um 17:19 schrieb Joachim Protze via mpiwg-tools: > Hi all, > > This is a reminder of our regularly scheduled call on Thursday: > > August 12, 2021 at 8 am Pacific / 11 am Eastern / 5pm CEST. > > I realize that people might be on vacation and we might end up > cancelling the call again. I'd like to do a head count of available > people to see whether we will have the call, so please let me know by > tomorrow whether you will be available. > > > Meeting Details: > https://github.com/mpiwg-tools/mpi-standard/wiki/Meeting-Information > > Agenda: > ------- > - QMPI > > > Cheers, > Joachim > > > _______________________________________________ > mpiwg-tools mailing list > mpiwg-tools(a)lists.mpi-forum.org > https://lists.mpi-forum.org/mailman/listinfo/mpiwg-tools > -- Dipl.-Inf. Joachim Protze IT Center Group: High Performance Computing Division: Computational Science and Engineering RWTH Aachen University Seffenter Weg 23 D 52074 Aachen (Germany) Tel: +49 241 80- 24765 Fax: +49 241 80-624765 protze(a)itc.rwth-aachen.de www.itc.rwth-aachen.de

1 0

Re: [mpiwg-tools] Meeting reminder/poll: August 12, 2021 (in 2 days)
by Michael Knobloch 11 Aug '21

11 Aug '21

I probably won't be able to make it- -Michael On 11.08.21 09:10, Martin Schulz via mpiwg-tools wrote: > I am around as well and should be able to make it, if nothing else short-term pops up. > > Martin > > -- Michael Knobloch Juelich Supercomputing Centre Institute for Advanced Simulation Forschungszentrum Juelich GmbH 52425 Juelich, Germany Phone: +49-2461-61-3546 Fax: +49-2461-61-6656 E-mail: m.knobloch(a)fz-juelich.de WWW: http://www.fz-juelich.de/jsc/

1 0