LAM/MPI logo

LAM/MPI General User's Mailing List Archives

  |   Home   |   Download   |   Documentation   |   FAQ   |   all just in this list

From: Kemp, Eric M. (eric.kemp_at_[hidden])
Date: 2005-01-14 17:00:47


LAM/MPIer's:

As a followup: we have rolled back our version of LAM to 7.0.6, and this version works fine -- no problems with redirected stdin. This is an acceptable solution for us, but
it also suggests a bug in 7.1.1.

Eric Kemp
Northrop Grumman IT TASC

-----Original Message-----
From: lam-bounces_at_[hidden] [mailto:lam-bounces_at_[hidden]]
Sent: Tuesday, January 11, 2005 11:21 AM
To: General LAM/MPI mailing list
Subject: RE: LAM: Odd problem with running MPI jobs

On Tue, 11 Jan 2005, Kemp, Eric M. wrote:

> I found a thread ("Some GM-related problems") on the lam-devel
> mailing list back in mid-February 2004 that discusses the same
> behavior. Is this the same bug?

As being the reporter of that bug, I thought of the similarities when
I saw your message, but I wanted to see more details before jumping
in. As the discussion revealed, I had no problems with LAM 7.0.3 while
all the CVS versions (that later became 7.1.x) gave the same strange
behaviour; however, something like 2-3 months after that discussion
took place I did try again the CVS for an unrelated reason and I was
able to do stdin redirection; that was for sure before 7.1 was
released, so I concluded that the problem was found and fixed.

Due to the fact that a massive recompilation of MPI applications would
be required for a LAM version change, I stayed with 7.0.3 until now
and I can't therefore provide any data about 7.1.

> We are using Fedora Core 1 (I mistakenly said Core 3 earlier),
> kernel version 2.4.22-1.2197.nptlsmp.

I was also using FC1 at that time, both when I originally reported the
problem and later when the stdin redirection worked. I switched to
TaoLinux1 (RHEL3 clone) in the mean time as FC1 is no longer
supported, but I have not tested any LAM 7.1 since I made the switch.

I realize that I'm not bringing any new data, but I thought that a
summary might help...

-- 
Bogdan Costescu
IWR - Interdisziplinaeres Zentrum fuer Wissenschaftliches Rechnen
Universitaet Heidelberg, INF 368, D-69120 Heidelberg, GERMANY
Telephone: +49 6221 54 8869, Telefax: +49 6221 54 8868
E-mail: Bogdan.Costescu_at_[hidden]
_______________________________________________
This list is archived at http://www.lam-mpi.org/MailArchives/lam/