August 2017 - mpiwg-sessions - mailman3.cels.anl.gov

Re: [mpiwg-sessions] Multiple MPI implementations in a single process - Application composition
by Vallee, Geoffroy R. 09 Aug '17

09 Aug '17

My point was absolutely not to say that it cannot be done one way or another (I actually use Hobbes enclaves and DTK to do this right now). I have been asked to mail this use case to the mailing list since the point was raised during the last MPI forum meeting that MPI sessions could be used for application composition, and I did so. >You can probably do this today with wi4mpi ( >https://github.com/cea-hpc/wi4mpi) Jeff On Wed, Aug 9, 2017 at 7:58 AM, Vallee, Geoffroy R. <valleegr at ornl.gov> wrote: > Hi, >> >> To follow up on a discussion we had a few weeks ago, here is a brief >> description of a use case where multiple implementations of MPI could be >> used by a single process. This is more an illustration of the idea rather >> than a detailed description of a model. >> >> Let’s assume we are composing a multi-physics applications where each >> kernel, that has been developed separately, is an MPI application and >> requires a specific implementation of MPI (e.g., it requires a customized >> flavor of Trilinos, ultimately creating a dependency to OpenMPI). Let’s >> also assume that a composition library is based MPI sessions and rely on a >> specific implementation of MPI. It is then possible to compose the >> multi-physics application by coupling the two MPI codes with the >> composition library, the two MPI implementations being potentially >> different, in which case a single process is, in fact, using two different >> MPI implementations. >> >> Let me know if you have any question. >> >> Thanks, >> _______________________________________________ >> mpiwg-sessions mailing list >> mpiwg-sessions at lists.mpi-forum.org >> https://lists.mpi-forum.org/mailman/listinfo/mpiwg-sessions >-- Jeff Hammond >jeff.science at gmail.com >http://jeffhammond.github.io/

1 0

Re: [mpiwg-sessions] Multiple MPI implementations in a single process - Application composition
by Jeff Hammond 09 Aug '17

09 Aug '17

You can probably do this today with wi4mpi ( https://github.com/cea-hpc/wi4mpi) Jeff On Wed, Aug 9, 2017 at 7:58 AM, Vallee, Geoffroy R. <valleegr(a)ornl.gov> wrote: > Hi, > > To follow up on a discussion we had a few weeks ago, here is a brief > description of a use case where multiple implementations of MPI could be > used by a single process. This is more an illustration of the idea rather > than a detailed description of a model. > > Let’s assume we are composing a multi-physics applications where each > kernel, that has been developed separately, is an MPI application and > requires a specific implementation of MPI (e.g., it requires a customized > flavor of Trilinos, ultimately creating a dependency to OpenMPI). Let’s > also assume that a composition library is based MPI sessions and rely on a > specific implementation of MPI. It is then possible to compose the > multi-physics application by coupling the two MPI codes with the > composition library, the two MPI implementations being potentially > different, in which case a single process is, in fact, using two different > MPI implementations. > > Let me know if you have any question. > > Thanks, > _______________________________________________ > mpiwg-sessions mailing list > mpiwg-sessions(a)lists.mpi-forum.org > https://lists.mpi-forum.org/mailman/listinfo/mpiwg-sessions -- Jeff Hammond jeff.science(a)gmail.com http://jeffhammond.github.io/

1 0

Multiple MPI implementations in a single process - Application composition
by Vallee, Geoffroy R. 09 Aug '17

09 Aug '17

Hi, To follow up on a discussion we had a few weeks ago, here is a brief description of a use case where multiple implementations of MPI could be used by a single process. This is more an illustration of the idea rather than a detailed description of a model. Let’s assume we are composing a multi-physics applications where each kernel, that has been developed separately, is an MPI application and requires a specific implementation of MPI (e.g., it requires a customized flavor of Trilinos, ultimately creating a dependency to OpenMPI). Let’s also assume that a composition library is based MPI sessions and rely on a specific implementation of MPI. It is then possible to compose the multi-physics application by coupling the two MPI codes with the composition library, the two MPI implementations being potentially different, in which case a single process is, in fact, using two different MPI implementations. Let me know if you have any question. Thanks,

1 0

Next meeting: Monday 7th August 2017 1pm Eastern US *today*
by Dan Holmes 07 Aug '17

07 Aug '17

Hi all, The next Sessions WG meeting is today at 1pm Eastern US time via webex. Connection details are now available on a Captcha-protected web page linked from the WG wiki page. The connection details can be obtained from here: http://mpi-forum.epcc.ed.ac.uk/mpiwg-sessions.php <http://mpi-forum.epcc.ed.ac.uk/mpiwg-sessions.php> The Sessions WG web page is here: https://github.com/mpiwg-sessions/sessions-issues/wiki <https://github.com/mpiwg-sessions/sessions-issues/wiki> Minutes from the last meeting (including homework/actions!) can be found at: https://github.com/mpiwg-sessions/sessions-issues/wiki/2017-07-24-webex <https://github.com/mpiwg-sessions/sessions-issues/wiki/2017-07-24-webex> The new cheatsheet produced by Howard is at: https://github.com/mpiwg-sessions/sessions-issues/wiki/sessions_cheat_sheet <https://github.com/mpiwg-sessions/sessions-issues/wiki/sessions_cheat_sheet> Cheers, Dan. The University of Edinburgh is a charitable body, registered in Scotland, with registration number SC005336.

1 0

Re: [mpiwg-sessions] a sessions cheat sheet wiki
by Dan Holmes 03 Aug '17

03 Aug '17

Hi Howard, Thanks, that is really useful and clear. I added a spoiler regarding the MPI_INIT/MPI_FINALIZE question. Cheers, Dan. > On 2 Aug 2017, at 21:14, Pritchard Jr., Howard <howardp(a)lanl.gov> wrote: > > Hi Folks > > I find the power point slide deck we had been using for discussing some of the Sessions > ideas a bit cumbersome. As a step toward something lighter weight, I came up with a > “Sessions cheat sheet” wiki that summarizes part of the API extensions being proposed. > The process spawn and implicit init/finalize are not covered. > > Anyway, the wiki is at https://github.com/mpiwg-sessions/sessions-issues/wiki/sessions_cheat_sheet <https://github.com/mpiwg-sessions/sessions-issues/wiki/sessions_cheat_sheet>. > > Please feel free to adjust/correct if I got something wrong from the slides. > > Next week I’ll try to sync up the "Current Issues and problems” wiki with the slide deck. > > Howard > > -- > Howard Pritchard > HPC-DES > Los Alamos National Laboratory > > _______________________________________________ > mpiwg-sessions mailing list > mpiwg-sessions(a)lists.mpi-forum.org > https://lists.mpi-forum.org/mailman/listinfo/mpiwg-sessions The University of Edinburgh is a charitable body, registered in Scotland, with registration number SC005336.

1 0

a sessions cheat sheet wiki
by Pritchard Jr., Howard 02 Aug '17

02 Aug '17

Hi Folks I find the power point slide deck we had been using for discussing some of the Sessions ideas a bit cumbersome. As a step toward something lighter weight, I came up with a “Sessions cheat sheet” wiki that summarizes part of the API extensions being proposed. The process spawn and implicit init/finalize are not covered. Anyway, the wiki is at https://github.com/mpiwg-sessions/sessions-issues/wiki/sessions_cheat_sheet. Please feel free to adjust/correct if I got something wrong from the slides. Next week I’ll try to sync up the "Current Issues and problems” wiki with the slide deck. Howard -- Howard Pritchard HPC-DES Los Alamos National Laboratory

1 0