Is there any chance you can upgrade to Open MPI?
On Apr 26, 2009, at 11:59 PM, Ahmad A. J. Agung wrote:
> Dear experts,
>
> I try to install LAM in SGI-MPT machine. 4 core in master node and
> 32 core in execution nodes, respectively Intel(R) Xeon(R) CPU X5272
> @ 3.40GHz. Resulting in the following problem:
>
> If I only execute aplication only for running in master node, it
> works, ie:
> agung_at_ptmlxclus:~/bin/lam-7.1.4/examples/hello> mpirun -np 4 hello
> Hello, world! I am 0 of 4
> Hello, world! I am 1 of 4
> Hello, world! I am 2 of 4
> Hello, world! I am 3 of 4
> agung_at_ptmlxclus:
>
> But if I execute by including any execution nodes, the process
> freeze, ie:
> agung_at_ptmlxclus:~/bin/lam-7.1.4/examples/hello> mpirun -np 8 hello
> <freeze>
> (and after pressing Ctrl-C 3 times, out the following:)
> ********************* WARNING ***********************
> This is a vulnerable region. Exiting the application
> now may lead to improper cleanup of temporary objects
> To exit the application, press Ctrl-C again
> ********************* WARNING ************************
> agung_at_ptmlxclus:
>
> freezing also occurs in tping (success in the first time, but freeze
> after trying parallel execution), lamhalt and lamexec command
> (lamexec only show the name of master node). I tried to google and
> find the same problem in
>
> http://www.lam-mpi.org/MailArchives/lam/2007/05/13152.php
>
> But the problem written there wasn't resolved yet. Could anyone help
> me to solve this please?
>
> Thanks in advance,
> Agung
>
>
>
>
>
>
>
> _______________________________________________
> This list is archived at http://www.lam-mpi.org/MailArchives/lam/
--
Jeff Squyres
Cisco Systems
|
