[SCore-users-jp] Re: [SCore-users] Re: SCore-users digest, Vol 1 #256 - 4 msgs

[Atsushi HORI]

Hi,

>Thankyou for your reply. I have talked to our key MPI-IO user and he
>insists that his code was running perfectly without errors with the
>previous version of SCore/MPICH/Redhat 7.2. And the MPI-IO users are
>mathemeticians, so they're pretty handy with numbers ;-)

Here I would like to clarify the facts.

FACT: The mathmaticians program was working well, before Streamline 
upgraded SCore.

Obviously the problem comes from the "upgrade." Asssuming that the 
other programs are runing well on the upgraded system, I believe we 
can focus on MPICH/SCore part, especially MPI-IO.

The difference between original MPICH and MPICH/SCore is only the 
part related to communication, and we did not change (even touch) 
MPI-IO part at all.

I am not sure which version you were using, but actually the 
MPICH/SCore in SCore 5.4 is based on MPICH 1.2.4, and in SCore 5.2 
based on MPICH 1.2.0. So, if you were using SCore 5.0 or older, then 
I think the problem comes from the MPICH 1.2.4, not MPICH 1.2.0.

>Assuming that they know their code is giving correct results, is there
>anyway they can bypass these error messages that they are getting now?
>If they can do that, then they can verify the results for themselves.
>If it's not possible to do that, is there any other way around this issue
>that you can think of? (which allows them to use a shared filesystem)

Fortunately, MPICH 1.2.0 is still supported by SCore 5.4 and you (or 
Streamline) can find in the SCore 5.4 distribution. And you can test 
if the problem is still found with MPICH 1.2.0 or not.

----
Atsushi HORI
Swimmy Software, Inc.

_______________________________________________
SCore-users mailing list
SCore-users ＠ pccluster.org
http://www.pccluster.org/mailman/listinfo/score-users