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

Re: [mpiwg-sessions] Slides for MPI BoF on Wed at SC17
by HOLMES Daniel 14 Nov '17

14 Nov '17

Hi Howard, I have 10 minutes - it’ll be a whistle-stop tour - fast enough to prevent detailed questions :) I’ve changed the title of the 5th slide. Good suggestion, thanks. Cheers, Dan. > On 14 Nov 2017, at 08:44, Pritchard Jr., Howard <howardp(a)lanl.gov> wrote: > > Hi Dan, > > Thanks for putting these together. How much time do you have? There > seem to be a lot > of concepts here. > > For slide 5, should there be something about the fact that we may be > dropping this and going to something more endpointy? May title it > “Previously Proposed…”? > > Thanks, > > Howard > > -- > Howard Pritchard > > B Schedule > HPC-ENV > Los Alamos National Laboratory > > > > > > On 11/13/17, 6:22 PM, "mpiwg-sessions on behalf of HOLMES Daniel" > <mpiwg-sessions-bounces(a)lists.mpi-forum.org on behalf of > d.holmes(a)epcc.ed.ac.uk> wrote: > >> Hi all, >> >> I finally put together some slides for the MPI BoF on Wednesday (see >> attached). >> >> I’ll be sending these to Martin for inclusion in his single slide-deck >> for the BoF *tomorrow afternoon*. >> >> If you have comments/suggestions/changes, please let me know by lunch >> tomorrow. >> >> Cheers, >> Dan. >> >> >> -- >> The University of Edinburgh is a charitable body, registered in >> Scotland, with registration number SC005336. >> > > _______________________________________________ > 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

Re: [mpiwg-sessions] Slides for MPI BoF on Wed at SC17
by Pritchard Jr., Howard 14 Nov '17

14 Nov '17

Hi Dan, Thanks for putting these together. How much time do you have? There seem to be a lot of concepts here. For slide 5, should there be something about the fact that we may be dropping this and going to something more endpointy? May title it “Previously Proposed…”? Thanks, Howard -- Howard Pritchard B Schedule HPC-ENV Los Alamos National Laboratory On 11/13/17, 6:22 PM, "mpiwg-sessions on behalf of HOLMES Daniel" <mpiwg-sessions-bounces(a)lists.mpi-forum.org on behalf of d.holmes(a)epcc.ed.ac.uk> wrote: >Hi all, > >I finally put together some slides for the MPI BoF on Wednesday (see >attached). > >I’ll be sending these to Martin for inclusion in his single slide-deck >for the BoF *tomorrow afternoon*. > >If you have comments/suggestions/changes, please let me know by lunch >tomorrow. > >Cheers, >Dan. > > >-- >The University of Edinburgh is a charitable body, registered in >Scotland, with registration number SC005336. >

1 0

Slides for MPI BoF on Wed at SC17
by HOLMES Daniel 13 Nov '17

13 Nov '17

Hi all, I finally put together some slides for the MPI BoF on Wednesday (see attached). I’ll be sending these to Martin for inclusion in his single slide-deck for the BoF *tomorrow afternoon*. If you have comments/suggestions/changes, please let me know by lunch tomorrow. Cheers, Dan. -- The University of Edinburgh is a charitable body, registered in Scotland, with registration number SC005336.

1 0

Re: [mpiwg-sessions] notes for today's meeting 11/6/17
by Martin Schulz 07 Nov '17

07 Nov '17

Hi Howard, all, Thanks for the notes. As for the BoF question - I agree, I see no reason why not. In this case, I would put you after the Hybrid WG, so the concept of Endpoints is known. I would also suggest to coordinate with Pavan, so he is not surprised :) Martin > On 6. Nov 2017, at 19:05, Pritchard Jr., Howard <howardp(a)lanl.gov> wrote: > > Hi Folks, > > Notes for today’s meeting are at > > https://github.com/mpiwg-sessions/sessions-issues/wiki/2017-11-06-webex <https://github.com/mpiwg-sessions/sessions-issues/wiki/2017-11-06-webex> > > Martin, there’s a bit about the BOF too. > > Howard > > -- > Howard Pritchard > B Schedule > HPC-ENV > Los Alamos National Laboratory > > _______________________________________________ > mpiwg-sessions mailing list > mpiwg-sessions(a)lists.mpi-forum.org > https://lists.mpi-forum.org/mailman/listinfo/mpiwg-sessions

1 0

notes for today's meeting 11/6/17
by Pritchard Jr., Howard 06 Nov '17

06 Nov '17

Hi Folks, Notes for today’s meeting are at https://github.com/mpiwg-sessions/sessions-issues/wiki/2017-11-06-webex Martin, there’s a bit about the BOF too. Howard -- Howard Pritchard B Schedule HPC-ENV Los Alamos National Laboratory

1 0

Next meeting - today - midday Eastern US time
by HOLMES Daniel 06 Nov '17

06 Nov '17

Hi all, The next meeting for the Sessions WG is today at midday Eastern US time (in 15 minutes). The connection details for the call can be found here: http://mpi-forum.epcc.ed.ac.uk/mpiwg-sessions.php The wiki for the working group, including documentation about Sessions and minutes of previous meetings is here: https://github.com/mpiwg-sessions/sessions-issues/wiki All: I’d like to have something concrete and interesting to say about the progress/state of this WG at the MPI BoF during SC17. That gives us 1 more meeting - today’s! Cheers, Dan. The University of Edinburgh is a charitable body, registered in Scotland, with registration number SC005336.

1 0

Next meeting - today - midday Eastern US time
by HOLMES Daniel 30 Oct '17

30 Oct '17

Hi all, The next meeting for the Sessions WG is today at midday Eastern US time. The connection details for the call can be found here: http://mpi-forum.epcc.ed.ac.uk/mpiwg-sessions.php The wiki for the working group, including documentation about Sessions and minutes of previous meetings is here: https://github.com/mpiwg-sessions/sessions-issues/wiki All: I’d like to have something concrete and interesting to say about the progress/state of this WG at the MPI BoF during SC17. That gives us 2 more meetings including today’s! Cheers, Dan. The University of Edinburgh is a charitable body, registered in Scotland, with registration number SC005336.

1 0

Next meeting - today - midday Eastern US time
by HOLMES Daniel 23 Oct '17

23 Oct '17

Hi all, The next meeting for the Sessions WG is today at midday Eastern US time. The connection details for the call can be found here: http://mpi-forum.epcc.ed.ac.uk/mpiwg-sessions.php The wiki for the working group, including documentation about Sessions and minutes of previous meetings is here: https://github.com/mpiwg-sessions/sessions-issues/wiki Ralph: thanks for your presentation last week. This week I’m intending to return to our previous discussions around sessions+endpoints and global vs local operation scope. The WG will look at the PMIx changes triggered by last week’s conversation (probably after SC17) and will bear in mind what PMIx can implement for us but we should focus on what value-add MPI can bring to users of MPI. All: I’d like to have something concrete and interesting to say about the progress/state of this WG at the MPI BoF during SC17. That gives us 3 more meetings including today’s! Cheers, Dan. The University of Edinburgh is a charitable body, registered in Scotland, with registration number SC005336.

1 0

Re: [mpiwg-sessions] notes from today's call
by rhc＠open-mpi.org 20 Oct '17

20 Oct '17

I have updated both the RFC (including adding example code on how to make multi-cluster connect work) and the prototype code per the below comments. Comments welcome! Ralph > On Oct 19, 2017, at 10:43 AM, rhc(a)open-mpi.org wrote: > > In thinking about this more, I realized that these new attributes won’t solve the problem raised by Martin. Instead, they were focused on giving the application the ability to define the nspace name - doesn’t help with the issue of specifying procs from different clusters in the PMIx_Connect call. > > After circling around again with the RM folks, a couple of things were resolved: > > 1. They really cannot support user-defined nspaces as the nspace directly correlates to their assignment of a “jobid” to the operation. Trying to build correlation tables to map a user definition to the RM’s identifier would be overly burdensome. > > 2. They are willing to standardize on prepending the cluster ID string to the nspace when referencing remote clusters. We tentatively agreed on using a colon ‘:’ as the delimiter. So when referencing a proc in nspace “bar” on cluster “foo”, you would provide an nspace of “foo:bar”. If no cluster ID is provided, then all parties will assume the nspace refers to the local cluster. > > 3. For the issue of having multiple, parallel PMIx_Connect operations spanning identical procs, we agreed to define an attribute PMIX_CONNECT_ID (string) whereby the application can provide its own unique “tag” for that operation. When provided, the RM and PMIx libraries will use this tag to separate out the operations. Note that all procs participating in the connect operation must provide the same tag. > > I will update the RFC accordingly, and provide a couple of new macros in pmix_common.h to make insertion and parsing of the cluster ID to/from the nspace easier. > > HTH > Ralph > >> On Oct 18, 2017, at 8:51 PM, rhc(a)open-mpi.org <mailto:[email protected]> wrote: >> >> Hello all >> >> I followed up on my AR from the meeting to check with the RMs on how they handle unique identifiers for procs on different clusters. As we had surmised, they provide a string name for each cluster, and they agreed that adding that to the PMIx nspace would be a reasonable path forward. >> >> I have accordingly updated the PMIx RFC (https://github.com/pmix/RFCs/pull/3 <https://github.com/pmix/RFCs/pull/3>) to include three new attributes: >> >> * PMIX_CONNECT_ID_MODIFIER_PREPEND: modify the nspace returned by the host by prepending the given modifier to the nspace string. This allows the application to "tag" the connected group in a recognizable fashion. >> >> * PMIX_CONNECT_ID_MODIFIER_APPEND: modify the nspace returned by the host by appending the given modifier to the nspace string. This allows the application to "tag" the connected group in a recognizable fashion. >> >> * PMIX_CONNECT_ID_REQUEST: request that the given identifier be used as the assigned nspace for the connected group. The "required" flag in the directive can be used to indicate that this identifier is required (i.e., the host RM must use it for the group, returning an error if it is already in use) as opposed to requested (i.e., the host RM can substitute its own unique identifier if the specified one is already in use). >> >> I also added the PMIX_CLUSTER_ID attribute to the list of information to be provided by the RM at process start - you can see the list here: >> https://github.com/pmix/pmix/wiki/2.8-Pmix-Server-Data-Requirements <https://github.com/pmix/pmix/wiki/2.8-Pmix-Server-Data-Requirements> >> >> Ralph >> >>> On Oct 16, 2017, at 11:48 AM, Pritchard Jr., Howard <howardp(a)lanl.gov <mailto:[email protected]>> wrote: >>> >>> Hi Folks, >>> >>> Notes from today’s call are on the wiki: >>> >>> https://github.com/mpiwg-sessions/sessions-issues/wiki/2017-10-16-webex <https://github.com/mpiwg-sessions/sessions-issues/wiki/2017-10-16-webex> >>> >>> Howard >>> >>> -- >>> Howard Pritchard >>> B Schedule >>> HPC-ENV >>> Los Alamos National Laboratory >>> >>> _______________________________________________ >>> mpiwg-sessions mailing list >>> mpiwg-sessions(a)lists.mpi-forum.org <mailto:[email protected]> >>> https://lists.mpi-forum.org/mailman/listinfo/mpiwg-sessions <https://lists.mpi-forum.org/mailman/listinfo/mpiwg-sessions> >> _______________________________________________ >> mpiwg-sessions mailing list >> mpiwg-sessions(a)lists.mpi-forum.org <mailto:[email protected]> >> https://lists.mpi-forum.org/mailman/listinfo/mpiwg-sessions >

1 0

Re: [mpiwg-sessions] notes from today's call
by rhc＠open-mpi.org 19 Oct '17

19 Oct '17

In thinking about this more, I realized that these new attributes won’t solve the problem raised by Martin. Instead, they were focused on giving the application the ability to define the nspace name - doesn’t help with the issue of specifying procs from different clusters in the PMIx_Connect call. After circling around again with the RM folks, a couple of things were resolved: 1. They really cannot support user-defined nspaces as the nspace directly correlates to their assignment of a “jobid” to the operation. Trying to build correlation tables to map a user definition to the RM’s identifier would be overly burdensome. 2. They are willing to standardize on prepending the cluster ID string to the nspace when referencing remote clusters. We tentatively agreed on using a colon ‘:’ as the delimiter. So when referencing a proc in nspace “bar” on cluster “foo”, you would provide an nspace of “foo:bar”. If no cluster ID is provided, then all parties will assume the nspace refers to the local cluster. 3. For the issue of having multiple, parallel PMIx_Connect operations spanning identical procs, we agreed to define an attribute PMIX_CONNECT_ID (string) whereby the application can provide its own unique “tag” for that operation. When provided, the RM and PMIx libraries will use this tag to separate out the operations. Note that all procs participating in the connect operation must provide the same tag. I will update the RFC accordingly, and provide a couple of new macros in pmix_common.h to make insertion and parsing of the cluster ID to/from the nspace easier. HTH Ralph > On Oct 18, 2017, at 8:51 PM, rhc(a)open-mpi.org wrote: > > Hello all > > I followed up on my AR from the meeting to check with the RMs on how they handle unique identifiers for procs on different clusters. As we had surmised, they provide a string name for each cluster, and they agreed that adding that to the PMIx nspace would be a reasonable path forward. > > I have accordingly updated the PMIx RFC (https://github.com/pmix/RFCs/pull/3 <https://github.com/pmix/RFCs/pull/3>) to include three new attributes: > > * PMIX_CONNECT_ID_MODIFIER_PREPEND: modify the nspace returned by the host by prepending the given modifier to the nspace string. This allows the application to "tag" the connected group in a recognizable fashion. > > * PMIX_CONNECT_ID_MODIFIER_APPEND: modify the nspace returned by the host by appending the given modifier to the nspace string. This allows the application to "tag" the connected group in a recognizable fashion. > > * PMIX_CONNECT_ID_REQUEST: request that the given identifier be used as the assigned nspace for the connected group. The "required" flag in the directive can be used to indicate that this identifier is required (i.e., the host RM must use it for the group, returning an error if it is already in use) as opposed to requested (i.e., the host RM can substitute its own unique identifier if the specified one is already in use). > > I also added the PMIX_CLUSTER_ID attribute to the list of information to be provided by the RM at process start - you can see the list here: > https://github.com/pmix/pmix/wiki/2.8-Pmix-Server-Data-Requirements <https://github.com/pmix/pmix/wiki/2.8-Pmix-Server-Data-Requirements> > > Ralph > >> On Oct 16, 2017, at 11:48 AM, Pritchard Jr., Howard <howardp(a)lanl.gov <mailto:[email protected]>> wrote: >> >> Hi Folks, >> >> Notes from today’s call are on the wiki: >> >> https://github.com/mpiwg-sessions/sessions-issues/wiki/2017-10-16-webex <https://github.com/mpiwg-sessions/sessions-issues/wiki/2017-10-16-webex> >> >> Howard >> >> -- >> Howard Pritchard >> B Schedule >> HPC-ENV >> Los Alamos National Laboratory >> >> _______________________________________________ >> mpiwg-sessions mailing list >> mpiwg-sessions(a)lists.mpi-forum.org <mailto:[email protected]> >> https://lists.mpi-forum.org/mailman/listinfo/mpiwg-sessions > > _______________________________________________ > mpiwg-sessions mailing list > mpiwg-sessions(a)lists.mpi-forum.org > https://lists.mpi-forum.org/mailman/listinfo/mpiwg-sessions

1 0