Sorry -- I've been offline since Friday morning, so I am slow to reply on this thread.
To be totally clear: I am seeing --enable-heterogeneous fail on homogeneous clusters. I was seeing timeouts and segv's in Cisco's MTT last week, IIRC. So I disabled the --enable-heterogeneous builds.
I only have access to Intel/x86-based servers for Cisco's MTT, so I can only test this one case.
If we want to keep the heterogeneous code:
1. George's point of doing a bisect to find the problem would probably be a good first step. I unfortunately do not have the cycles to do this. Does someone else?
2. Someone really needs to commit to doing regular periodic testing of actual heterogeneous test cases (as I think I mentioned in a prior email, a minimum of once a week would probably be good). I think Gilles mention running in big endian, little endian, and mixed big-little endian cases -- that would cover the entire range, and would be great.
On Apr 28, 2014, at 9:05 AM, Ralph Castain <rhc_at_[hidden]> wrote:
> No, it looks like something has broken it since I last tested. Sorry about the confusion.
> On Apr 27, 2014, at 10:55 PM, Gilles Gouaillardet <gilles.gouaillardet_at_[hidden]> wrote:
>> I might have misunderstood Jeff's comment :
>>> The broken part(s) is(are) likely somewhere in the datatype and/or PML code (my guess). Keep in mind that my only testing of this feature is in *homogeneous* mode -- i.e., I compile with --enable-heterogeneous and then run tests on homogeneous machines. Meaning: it's not only broken for actual heterogeneity, it's also broken in the "unity"/homogeneous case.
>> Unfortunatly, a trivial send/recv can hang in this case (--enable-heterogeneous and homogenous cluster of little endian procs).
>> i opened #4568 https://svn.open-mpi.org/trac/ompi/ticket/4568 in order to track this issue
>> (uninitialized data can cause a hang with this config)
>> trunk is affected, v1.8 is very likely affected too
>> On 2014/04/28 12:22, Ralph Castain wrote:
>>> I think you misunderstood his comment. It works fine on a homogeneous cluster, even with --enable-hetero. I've run it that way on my cluster.
>>> On Apr 27, 2014, at 7:50 PM, Gilles Gouaillardet
>>>> According to Jeff's comment, OpenMPI compiled with
>>>> --enable-heterogeneous is broken even in an homogeneous cluster.
>>>> as a first step, MTT could be ran with OpenMPI compiled with
>>>> --enable-heterogenous and running on an homogeneous cluster
>>>> (ideally on both little and big endian) in order to identify and fix the
>>>> /* this build is currently disabled in the MTT config of the
>>>> cisco-community cluster */
>> devel mailing list
>> Subscription: http://www.open-mpi.org/mailman/listinfo.cgi/devel
>> Link to this post: http://www.open-mpi.org/community/lists/devel/2014/04/14624.php
> devel mailing list
> Subscription: http://www.open-mpi.org/mailman/listinfo.cgi/devel
> Link to this post: http://www.open-mpi.org/community/lists/devel/2014/04/14625.php
For corporate legal information go to: http://www.cisco.com/web/about/doing_business/legal/cri/