Go to each machine and run laminfo
1] you should check that the same version of lam is installed as well as the same compilers
2] Make sure you can ssh without password to the machines
from the master to nodes and frome nodes to master
Do that and see what happens
Roberto Scipioni
ICYS Research Fellow
and ICYS Cluster Manager
ICYS, NIMS, Japan
> Hi LAM users and developer,
>
> I have lam-7.1.1-pgi compiled on my machine and its
> path is /usr/local/lam-7.1.1-pgi. The suffix pgi
> indicates to the compiler's name. I've successfully
> compiled my application using this version of LAM
> compiled on c,c+ and f77 compilers. However, it failed
> to run the application on multiple prosseces and when
> I attempted to run the application, I got the message
> showed below at the tail of rsl.out.* files :
>
> MPI_Reduce: invalid operation: Invalid argument (rank
> 0, MPI_COMM_WORLD)
> Rank (0, MPI_COMM_WORLD): Call stack within LAM:
> Rank (0, MPI_COMM_WORLD): - MPI_Reduce()
> Rank (0, MPI_COMM_WORLD): - main()
>
> q1. Should I rebuild the new release of LAM or not
> necessary?How?
>
> q2. How I can find out weather the LAM's version
> compiled on all of remote machines are identical or
> not?
> q3. Why has my application failed to finish the task?
>
> Your help will be highly appriciated.
> Thanks in advance
> Ibrahim
> Master student
> Australia
>
>
>
>
>
>
> ____________________________________________________________________________________Be a better friend, newshound, and
> know-it-all with Yahoo! Mobile. Try it now.
> http://mobile.yahoo.com/;_ylt=Ahu06i62sR8HDtDypao8Wcj9tAcJ
>
> _______________________________________________
> This list is archived at http://www.lam-mpi.org/MailArchives/lam/
>
|
