This issue has been addressed in the Open MPI trunk with r20114. This
fix will be moved to the v1.3 branch in the next few days and will be
in the eventual v1.3.0 release.
Thanks again for the heads up on this issue.
Cheers,
Josh
On Dec 4, 2008, at 7:50 AM, Josh Hursey wrote:
Matthias,
T
Hi List,
I have a MPI program which uses one sided communication with derived
datatypes (MPI_Type_create_indexed_block). I developed the code with
MPICH2 and unfortunately didn't thought about trying it out with
OpenMPI. Now that I'm "porting" the Application to OpenMPI I'm facing
some problems.
As announced on the insideHPC blog, I'll be interviewed by Leo Laporte
(remember TechTV and ZDNet?) and Randal Schwartz on the TWIT.tv FLOSS
Weekly podcast (Free Libre Open Source Software) about the Open MPI
project.
The interview is live at 2pm, US Pacific time, on Wednesday, December
3
FWIW, the README describes some of the Fortran issues:
- Open MPI will build bindings suitable for all common forms of
Fortran 77 compiler symbol mangling on platforms that support it
(e.g., Linux). On platforms that do not support weak symbols (e.g.,
OS X), Open MPI will build Fortran 77
On Dec 8, 2008, at 8:03 AM, Number Cruncher wrote:
I notice that bug ticket #954 https://svn.open-mpi.org/trac/ompi/ticket/954
has the very issue I'm encountering: I want to know when mpirun
fails because of a missing executable during some automated tests.
At the moment, nearly 2 years aft
Thanks I will pass that on to the person asking. And it answers my
own questions of observed behavior with IMB and other tools.
Brock Palen
www.umich.edu/~brockp
Center for Advanced Computing
bro...@umich.edu
(734)936-1985
On Dec 10, 2008, at 10:37 AM, George Bosilca wrote:
Brock,
Open
Bernard,
Could you give us more details on the version of Open MPI you are
using ? I suppose from your message this is one of the 1.2 series, but
mode details would be greatly helpful.
An utility called orte-clean can be used to try to remove all the
garbage left, should something go wron
Brock,
Open MPI has a very fancy collective implementation. While we don't
use multicast, we dynamically change the algorithm used for the
collective based on the number of nodes, the size and type of the
message, and eventually the operation involved if any. Moreover, we
detect hierarchi
Hi everybody
I want to use MPI_publish_name function to do communicaztion between two
independant code.
I saw on the web i must use the orted daemon with the following command:
orted --persistent --seed --scope public --universe foo
The communication success, but when i close the communicat
On Dec 9, 2008, at 9:49 PM, Brock Palen wrote:
Is OpenMPI using Multicast for things like MPI_Bcast() ?
Not today. There have been some experiments with it, but nothing yet.
Some simple tests of really old versions (1.0.2) showed binary tree
type performance, while latter versions gave me
On Dec 10, 2008, at 12:05 AM, Максим Чусовлянов wrote:
Hello! I appreciate your help. I received a library of my
algorithm's. But I don't know how to use it.
Can you explain what you mean / what you did?
E.g., did you build your algorithm code inside an OMPI tree and now
have an mca_coll_
Currently, the only way is to name your environment variable
OMPI_MCA_; mpirun will automatically scoop all of those up and
send them along.
It wouldn't be *terribly* difficult to add other mechanisms for
sending along environment variables, though. What do you need / want
to do?
On D
On Dec 10, 2008, at 4:17 AM, Gilbert Grosdidier wrote:
I think there is no cheap way to check this other than switching on
several
'btl' related mca debug flags.
Sadly, a feature for printing the actual connectivity map got cut from
the v1.3 feature list when we started trying to get v1.3
Dear OpenMPI developers,
i wold like to know how i can export by default some enviroment
variables, so without using" -x" option after mpirun
Is it possible to add some flags in openmpi-mca-params.conf ?
Thanks.
--
Ing. Gabriele Fatigati
Parallel programmer
CINECA Systems & Tecnologies Depart
Hello,
I suggest you use the '--mca btl sm,openib,self' option when launching your job
with the mpirun command. Or '--mca btl openib,self' if shared-memory is not
involved.
I think there is no cheap way to check this other than switching on several
'btl' related mca debug flags. Please ref
Now its working fine.
Thanks for the suggestion.
Some clarifications required:
I think its possible to mention: communication should happen thru only
IB, not Ethernet.
Not getting how to do it.
How to check whether IB is used or not?
Regards,
Sangamesh
On Sun, Dec 7, 2008 at 9:08 PM, Brian Dob
Hello! I appreciate your help. I received a library of my algorithm's. But I
don't know how to use it.
2008/12/6 Jeff Squyres
> On Dec 6, 2008, at 10:32 AM, Максим Чусовлянов wrote:
>
> Hello. Thanks for your response. I'm create a my module and component
>> into the coll. But when I execute
17 matches
Mail list logo
