If you are running fewer processes on your nodes than they have
processors, then you can improve performance by adding
-mca mpi_paffinity_alone 1
to your cmd line. This will bind your processes to individual cores,
which helps with latency. If your program involves collectives, then
you ca
Hi all,
I am new to OpenMPI, and have an urgent run-time question. I
have openmpi-1.3.2 compiled with Intel Fortran compiler v.11 simply by
./configure --prefix= F77=ifort FC=ifort
then I set my LD_LIBRARY_PATH to include and
and compile my Fortran program properly. No compilation error.
I run
On Thu, Jun 25, 2009 at 10:29:39AM -0700, D'Auria, Raffaella wrote:
>
>Dear All,
>I have been encountering a fatal type "error polling LP CQ with status
>RETRY EXCEEDED ERROR status number 12" whenever I try to run a simple
>MPI code (see below) that performs an AlltoAll call.
>
Man, was this a PITA to chase down. Finally found it, though. Fixed on
trunk as of r21549
Thanks!
Ralph
So something else is wrong.
On Jun 25, 2009, at 3:19 PM, Mostyn Lewis wrote:
Just local machine - direct from the command line wth a script like
the one below. So, no launch mechanism.
F
Not sure I understand the problem. Your server will have the ability
to recv messages from multiple sources on the port it picks. Is there
some reason why you want to specify the port for each process?
Note that you can still do so - you just have to do it -before-
calling MPI_Init. MPI job
Jeff Layton wrote:
> Jeff Squyres wrote:
>> On Jun 1, 2009, at 2:04 PM, Jeff Layton wrote:
>>
>>> error: executing task of job 3084 failed: execution daemon on host
>>> "compute-2-2.local" didn't accept task
>>>
>>
>> This looks like an error message from the resource manager/scheduler
>> -- not fr
On Fri, Jun 26, 2009 at 9:57 AM, Ralph Castain wrote:
> How are you trying to use this port? Are you expecting that to be the port
> used for MPI communications?
i am thinking to use MPI_Send and MPI_Recv on this connection. One node
will run in server mode to server data to others, only one proce
How are you trying to use this port? Are you expecting that to be the
port used for MPI communications?
We create the port string based on the actual TCP/IP contact info for
this process on our out-of-band comm channel. This has nothing to do
with whatever port might be used for MPI communi
On Jun 26, 2009, at 9:45 AM, Dave Love wrote:
Scott Atchley writes:
I believe the answer is yes as long as all NICs are in the same
fabric
(they usually are).
Thanks. Do you mean it won't if, in this case, the two NICs are on
separate switches?
Dave,
George's answer supersedes mine. Y
George Bosilca writes:
> It is not the BTL who open the second endpoint, it is the MTL. It's a
> very long story, but unfortunately right now the two components (MTL
> and BTL) each open an endpoint. Once the upper level complete the
> selection of the component for the run, one of the endpoints
Scott Atchley writes:
> I believe the answer is yes as long as all NICs are in the same fabric
> (they usually are).
Thanks. Do you mean it won't if, in this case, the two NICs are on
separate switches?
Hi,
i would like to know if it is possible to make a client/server with
MPI where the application A (client) would connect to application B
(server) using a specific TCP/IP port ? I have tried this code but it
opens a port using some mysterious protocol (which looks like TIPC
btw), what is going on
12 matches
Mail list logo
