July 2018 - mpi-forum - mailman3.cels.anl.gov

Reminder: Virtual Meeting - TODAY
by Martin Schulz 24 Jul '18

24 Jul '18

Hi all, Today at 11am US Eastern time (which should be 8am Pacific, 5pm Central Europe, midnight the following day in Japan) we will have our first virtual meeting for this period. The topic will be on preparing the MPI_T Events proposal for a reading in Barcelona and Marc-Andre will lead the discussion. The dial-in information is at https://www.mpi-forum.org/meetings/ <https://www.mpi-forum.org/meetings/> Thanks! Martin — Prof. Dr. Martin Schulz, Chair of Computer Architecture and Parallel Systems Department of Informatics, TU-Munich, Boltzmannstraße 3, D-85748 Garching Email: schulzm(a)in.tum.de

1 0

Virtual Meeting WebEx Info Posted
by Bland, Wesley 23 Jul '18

23 Jul '18

Hi folks, The virtual meeting webex information is now posted on the website. Unfortunately because of the musically inclined bots that infested our previous calls, the information is now behind a captcha. You'll be able to find links to everything on the meeting page. Also, for those of you who asked me about calendar invitations for these, you can get that on the website too. I've set up a Google calendar for the MPI Forum events. Click on the date for any of the virtual meetings to be taken to it. You should be able to subscribe to that in any calendar app or you could copy the events into your calendar if you like. Hope to see to everyone on the virtual call this Wednesday at 11 AM Eastern US. Thanks, Wesley

1 0

MPI Meetings Dates and Locations, Virtual Meetings, MPI 4.0 Timeline
by Martin Schulz 17 Jul '18

17 Jul '18

Hi all, My apologies that this comes a bit delayed - during the last MPI forum meeting we had quite some discussions on future planning topics. Here is summary - if you have any comments, please speak up as soon as possible. I will add this to the webpages in the coming days and will consider it final, unless someone speaks up. Thanks! Martin A) Virtual Meetings until the September Meeting We agreed on the following meeting schedule - all meetings will be held at 8am US West Coast time, which seems to work better for US, Europe and Japan (should be 5pm Central Europe, midnight in Japan). July 25: Pre-reading of MPI_T Events led by Marc-Andre August 8: Discussion on semantic terms led by Puri August 29: Pre-reading of MPI Process Failure ticket by Wesley B) Dates and Locations for 2019 Spring: March 4-7 in Chattanooga (thanks, Tony) Summer: May 28-31 in Chicago (thanks, Jithin) (note Tue-Fri due to Memorial Day) Fall: around EuroMPI (which is 9/9-13) in Zürich (thanks, Torsten) Winter: December 9-12 in Albuquerque (thanks Matthew) C) MPI 4.0 Timeline The general feeling in the forum was that it is a good time to start planning for a next release of the standard, since several large items are already in process of voting or close to it. The following is a proposed time line: - Publish Draft Standard at SC 2018 (as per our new rules) Includes all tickets passed in Barcelona - Publish Draft MPI 4.0 at SC 2019 start ratification process in December 2019 Final ratification meeting in March 2020 - Publish standard in time (latest) by ISC 2020 — Prof. Dr. Martin Schulz, Chair of Computer Architecture and Parallel Systems Department of Informatics, TU-Munich, Boltzmannstraße 3, D-85748 Garching Email: schulzm(a)in.tum.de

1 0

Re: [Mpi-forum] MPI - write to same binary file. Each process shows huge difference in time intervals to complete
by Catherine Jenifer Rajam Rajendran 10 Jul '18

10 Jul '18

Hi Jeff, Thank you so much for the update! I already posted in Stack Overflow - no help yet! I will check with our Cray machine support staff and I will shoot an email to MPICH too! Really, thanks for the response! Best, Catherine On Tue, Jul 10, 2018 at 2:21 PM, Jeff Hammond <jeff.science(a)gmail.com> wrote: > Hi Catherine, > > It sounds like this is an implementation issue. This email list is for > discussion of the MPI standard itself. While many of us have a great deal > of implementation expertise, you are likely to get the best response from > the user list associated with the implementation you are using: > > MPICH <discuss(a)mpich.org> > Open-MPI <users(a)lists.open-mpi.org> > MVAPICH2 <mvapich-discuss(a)cse.ohio-state.edu> > > If you are using Cray MPI, you'll need to contact Cray support, perhaps > via the staff that support your Cray machine locally. For Intel MPI, start > with https://software.intel.com/en-us/forums/intel- > clusters-and-hpc-technology/. I don't know about SGI or NEC support, > unfortunately. You may also have good luck with StackOverflow - there are > quite a few MPI experts there. > > I'll note that most of the implementations of MPI I/O are based on ROMIO, > which is part of MPICH, so you might want to start with the MPICH user list. > > Best, > > Jeff > > > On Tue, Jul 10, 2018 at 11:01 AM, Catherine Jenifer Rajam Rajendran < > catrajen(a)iu.edu> wrote: > >> Hi All, >> >> I am trying to write in the same binary file using MPI. I set the offset >> for each process in the beginning as per the rank. Then the following code >> snippet in C runs. All MPI process executes and computes the value and it >> writes to the exact offset as set. >> >> The problem I am facing is, say, out of 32 Process, one process is >> executed in 2 hours. Rest of the process keeps running for more than 24 >> hours, The thing is, it computes the values as expected but it takes so >> much time. It seems like a deadlock situation, each process waits for some >> resource. But, I am not sharing/communicating between the processes. I am >> just using MPI_File_write_at to write at a specific location in the binary >> file. >> >> I need to mention that each process computes huge amount of data so >> storing it temporarily seemed inappropriate. I want to write the output in >> single file as number of processes is increased depending on input data. >> Number of computations are evenly distributed to all process. So, why does >> process takes different time interval to finish its job?! >> >> for(i=1;i<=limit;i++) >> { >> for(j=i+1;j<=limit;j++) >> { >> if(my_rank == step%num_cpus) >> { >> Calc = Calculation(); >> buf[0] = (double)Calc; >> MPI_File_write_at(outFile, OUT_ofst, buf, 1, MPI_DOUBLE, >> &status); >> Calc = 0.0; >> OUT_ofst += num_cpus*MPI_File_write_at(sizeof(double)); >> count++; >> } >> step++; >> } >> } >> >> I am new to MPI and I guess people must have had similar issues while >> executing in MPI. Can anyone help me out please! I can provide more details >> if needed. >> >> Thanks, >> Catherine >> >> _______________________________________________ >> mpi-forum mailing list >> mpi-forum(a)lists.mpi-forum.org >> https://lists.mpi-forum.org/mailman/listinfo/mpi-forum >> >> > > > -- > Jeff Hammond > jeff.science(a)gmail.com > http://jeffhammond.github.io/ > > _______________________________________________ > mpi-forum mailing list > mpi-forum(a)lists.mpi-forum.org > https://lists.mpi-forum.org/mailman/listinfo/mpi-forum > >

1 0

Re: [Mpi-forum] MPI - write to same binary file. Each process shows huge difference in time intervals to complete
by Jeff Hammond 10 Jul '18

10 Jul '18

Hi Catherine, It sounds like this is an implementation issue. This email list is for discussion of the MPI standard itself. While many of us have a great deal of implementation expertise, you are likely to get the best response from the user list associated with the implementation you are using: MPICH <discuss(a)mpich.org> Open-MPI <users(a)lists.open-mpi.org> MVAPICH2 <mvapich-discuss(a)cse.ohio-state.edu> If you are using Cray MPI, you'll need to contact Cray support, perhaps via the staff that support your Cray machine locally. For Intel MPI, start with https://software.intel.com/en-us/forums/intel-clusters-and-hpc-technology/. I don't know about SGI or NEC support, unfortunately. You may also have good luck with StackOverflow - there are quite a few MPI experts there. I'll note that most of the implementations of MPI I/O are based on ROMIO, which is part of MPICH, so you might want to start with the MPICH user list. Best, Jeff On Tue, Jul 10, 2018 at 11:01 AM, Catherine Jenifer Rajam Rajendran < catrajen(a)iu.edu> wrote: > Hi All, > > I am trying to write in the same binary file using MPI. I set the offset > for each process in the beginning as per the rank. Then the following code > snippet in C runs. All MPI process executes and computes the value and it > writes to the exact offset as set. > > The problem I am facing is, say, out of 32 Process, one process is > executed in 2 hours. Rest of the process keeps running for more than 24 > hours, The thing is, it computes the values as expected but it takes so > much time. It seems like a deadlock situation, each process waits for some > resource. But, I am not sharing/communicating between the processes. I am > just using MPI_File_write_at to write at a specific location in the binary > file. > > I need to mention that each process computes huge amount of data so > storing it temporarily seemed inappropriate. I want to write the output in > single file as number of processes is increased depending on input data. > Number of computations are evenly distributed to all process. So, why does > process takes different time interval to finish its job?! > > for(i=1;i<=limit;i++) > { > for(j=i+1;j<=limit;j++) > { > if(my_rank == step%num_cpus) > { > Calc = Calculation(); > buf[0] = (double)Calc; > MPI_File_write_at(outFile, OUT_ofst, buf, 1, MPI_DOUBLE, > &status); > Calc = 0.0; > OUT_ofst += num_cpus*MPI_File_write_at(sizeof(double)); > count++; > } > step++; > } > } > > I am new to MPI and I guess people must have had similar issues while > executing in MPI. Can anyone help me out please! I can provide more details > if needed. > > Thanks, > Catherine > > _______________________________________________ > mpi-forum mailing list > mpi-forum(a)lists.mpi-forum.org > https://lists.mpi-forum.org/mailman/listinfo/mpi-forum > > -- Jeff Hammond jeff.science(a)gmail.com http://jeffhammond.github.io/

1 0

MPI - write to same binary file. Each process shows huge difference in time intervals to complete
by Catherine Jenifer Rajam Rajendran 10 Jul '18

10 Jul '18

Hi All, I am trying to write in the same binary file using MPI. I set the offset for each process in the beginning as per the rank. Then the following code snippet in C runs. All MPI process executes and computes the value and it writes to the exact offset as set. The problem I am facing is, say, out of 32 Process, one process is executed in 2 hours. Rest of the process keeps running for more than 24 hours, The thing is, it computes the values as expected but it takes so much time. It seems like a deadlock situation, each process waits for some resource. But, I am not sharing/communicating between the processes. I am just using MPI_File_write_at to write at a specific location in the binary file. I need to mention that each process computes huge amount of data so storing it temporarily seemed inappropriate. I want to write the output in single file as number of processes is increased depending on input data. Number of computations are evenly distributed to all process. So, why does process takes different time interval to finish its job?! for(i=1;i<=limit;i++) { for(j=i+1;j<=limit;j++) { if(my_rank == step%num_cpus) { Calc = Calculation(); buf[0] = (double)Calc; MPI_File_write_at(outFile, OUT_ofst, buf, 1, MPI_DOUBLE, &status); Calc = 0.0; OUT_ofst += num_cpus*MPI_File_write_at(sizeof(double)); count++; } step++; } } I am new to MPI and I guess people must have had similar issues while executing in MPI. Can anyone help me out please! I can provide more details if needed. Thanks, Catherine

1 0