Dahai Guo,
On 10/7/2015 3:08 AM, Dahai Guo wrote:
Thanks, Jeff. It is very helpful. some more questions :-):
1. There are many coll components, such as basic, tuned, self, cuda,
sm, and etc. Are they all selected at the MPI_Init time? or it just
select those satisfying some criteria,
Thanks for the reply Ralph. Best wishes for your surgery and recuperation.
Regards
--Nysal
On Wed, Oct 7, 2015 at 12:08 AM, Ralph Castain wrote:
> You hit upon a problem that Jeff and I were talking about yesterday. Both
> master and 2.x need to be fixed. The problem is a little more subtle as
You hit upon a problem that Jeff and I were talking about yesterday. Both
master and 2.x need to be fixed. The problem is a little more subtle as direct
launch under non-PMIx systems doesn’t have access to a non-blocking fence, and
thus there is no way to call progress while waiting for the barr
In v1.8 there is a RTE barrier in finalize. OMPI_LAZY_WAIT_FOR_COMPLETION
waits for the barrier to complete. Internally opal_progress() is invoked.
In the master branch we call PMIX fence instead. PMIX_WAIT_FOR_COMPLETION
seems to only call usleep. How will ompi progress outstanding operations ?
R
Thanks, Jeff. It is very helpful. some more questions :-):
1. There are many coll components, such as basic, tuned, self, cuda, sm, and
etc. Are they all selected at the MPI_Init time? or it just select those
satisfying some criteria, hardware, communicator size? or only some specific
ones a
On Oct 6, 2015, at 10:19 AM, Dahai Guo wrote:
>
> Thanks, Gilles. Some more questions:
>
> 1. how does Open MPI define the priorities of the different collective
> components? what criteria is based on?
The priorities are in the range of [0, 100] (100=highest). The priorities tend
to be fai
On Oct 3, 2015, at 10:12 PM, Dahai Guo wrote:
>
> Thanks, Jeff. I am trying to understand in detail how Open MPI works in the
> run time. What main functions does it call to select and initialize the coll
> components?
In general, each framework (e.g., coll, pml, ...etc.) is polled during
MPI
Thanks, Gilles. Some more questions:
1. how does Open MPI define the priorities of the different collective
components? what criteria is based on?
2. how does a MPI collective function (MPI_Barrier for example) choose the
exact algorithm it use? based on message size, and communicator size? any
at first, you can check the priorities of the various coll modules
with ompi_info
$ ompi_info --all | grep \"coll_ | grep priority
MCA coll: parameter "coll_basic_priority" (current
value: "10", data source: default, level: 9 dev/all, type: int)
MCA coll: parameter
