Open MPI logo

Open MPI Development Mailing List Archives

  |   Home   |   Support   |   FAQ   |   all Development mailing list

Subject: Re: [OMPI devel] [OMPI svn] svn:open-mpi r17366
From: George Bosilca (bosilca_at_[hidden])
Date: 2008-02-04 13:39:28


I remember the discussion. However, the outcome is not that clear to
me anymore :). Moreover, I know that tuned collective do not rely on
the user supplied buffer on the intermediary nodes, which pinpoint to
the fact that we ignore it.

   george.

On Feb 4, 2008, at 1:29 PM, Edgar Gabriel wrote:

> well, but the point is, that the requirement I listed is explicitly
> mentioned in the chapter about MPI_Reduce. So I would claim, that it
> trumps (at least) the second citation. (BTW: I am pretty sure that I
> remember having this discussion early in Open MPI days already :-) )
>
> But maybe that's something for 2.1?
> Thanks
> edgar
>
>
> George Bosilca wrote:
>> Well, well, well ... There is really no way to make everybody
>> happy ...
>>
>> When I did the previous commit (where I added the check), Tim Prins
>> pointed out the following 2 paragraphs from the MPI standard.
>>
>>>> MPI-1.1 s4.9.1 (p113, line 28) says the recvbuf is "significant
>>>> only
>>>> at root".
>>>>
>>>> MPI-1.1 s2.2 (p7, line 7-9) and MPI-2 s2.3 (p7 line 1-3) have the
>>>> identical text:
>>>> "Another frequent situation arises when an argument value is needed
>>>> only by a subset of the processes. When an argument is not
>>>> significant at a process then an arbitrary value can be passed as
>>>> argument."
>>
>> As you can see on the second paragraph, the standard claims that an
>> arbitrary value can be passed.
>>
>> george.
>>
>> On Feb 4, 2008, at 8:17 AM, Edgar Gabriel wrote:
>>
>>> George,
>>>
>>> I hate to say this, but I don't think that this is correct. Page
>>> 114 in
>>> MPI 1 says (in the section about MPI_Reduce)
>>>
>>> "Thus, all processes provide input buffers and output buffers of the
>>> same length, with elements of the same time".
>>>
>>> The content is only significant at the root, but all processes
>>> have to
>>> provide the buffer according to my understanding...
>>>
>>> Thanks
>>> Edgar
>>>
>>> bosilca_at_[hidden] wrote:
>>>> Author: bosilca
>>>> Date: 2008-02-03 20:44:41 EST (Sun, 03 Feb 2008)
>>>> New Revision: 17366
>>>> URL: https://svn.open-mpi.org/trac/ompi/changeset/17366
>>>>
>>>> Log:
>>>> As the receive buffer is only significant at root, limit the
>>>> check only where it makes sense.
>>>>
>>>> Text files modified:
>>>> trunk/ompi/mpi/c/reduce.c | 4 +---
>>>> 1 files changed, 1 insertions(+), 3 deletions(-)
>>>>
>>>> Modified: trunk/ompi/mpi/c/reduce.c
>>>> =
>>>> =
>>>> =
>>>> =
>>>> =
>>>> =
>>>> =
>>>> =
>>>> =
>>>> =
>>>> =
>>>> ===================================================================
>>>>
>>>> --- trunk/ompi/mpi/c/reduce.c (original)
>>>> +++ trunk/ompi/mpi/c/reduce.c 2008-02-03 20:44:41 EST (Sun, 03
>>>> Feb
>>>> 2008)
>>>> @@ -59,9 +59,7 @@
>>>> free(msg);
>>>> return ret;
>>>> } else if ((ompi_comm_rank(comm) != root && MPI_IN_PLACE ==
>>>> sendbuf) ||
>>>> - (ompi_comm_rank(comm) == root && MPI_IN_PLACE
>>>> ==
>>>> recvbuf)) {
>>>> - err = MPI_ERR_ARG;
>>>> - } else if( sendbuf == recvbuf ) {
>>>> + (ompi_comm_rank(comm) == root && ((MPI_IN_PLACE
>>>> == recvbuf) || (sendbuf == recvbuf)))) {
>>>> err = MPI_ERR_ARG;
>>>> } else {
>>>> OMPI_CHECK_DATATYPE_FOR_SEND(err, datatype, count);
>>>> _______________________________________________
>>>> svn mailing list
>>>> svn_at_[hidden]
>>>> http://www.open-mpi.org/mailman/listinfo.cgi/svn
>>>
>>> --
>>> Edgar Gabriel
>>> Assistant Professor
>>> Parallel Software Technologies Lab http://pstl.cs.uh.edu
>>> Department of Computer Science University of Houston
>>> Philip G. Hoffman Hall, Room 524 Houston, TX-77204, USA
>>> Tel: +1 (713) 743-3857 Fax: +1 (713) 743-3335
>>> _______________________________________________
>>> devel mailing list
>>> devel_at_[hidden]
>>> http://www.open-mpi.org/mailman/listinfo.cgi/devel
>>
>>
>> ------------------------------------------------------------------------
>>
>> _______________________________________________
>> devel mailing list
>> devel_at_[hidden]
>> http://www.open-mpi.org/mailman/listinfo.cgi/devel
>
> --
> Edgar Gabriel
> Assistant Professor
> Parallel Software Technologies Lab http://pstl.cs.uh.edu
> Department of Computer Science University of Houston
> Philip G. Hoffman Hall, Room 524 Houston, TX-77204, USA
> Tel: +1 (713) 743-3857 Fax: +1 (713) 743-3335
> _______________________________________________
> devel mailing list
> devel_at_[hidden]
> http://www.open-mpi.org/mailman/listinfo.cgi/devel



  • application/pkcs7-signature attachment: smime.p7s