[Root_list] Fwd: Different results in different machines

[Mark M. Ito]

Message from Carlos Ayerbe-Gayoso <gayoso at jlab.org> .


-------- Original Message --------
Subject: 	Different results in different machines
Date: 	Wed, 07 Aug 2013 16:13:18 -0400
From: 	Carlos Ayerbe-Gayoso <gayoso at jlab.org>
Reply-To: 	gayoso at jlab.org
To: 	marki at jlab.org




*************************************************************
Dear ROOTers at Jlab

in the last weeks I'm being migrating my Geant4 code from my local
machine to the ifarm. The G4 code make use of root in order to storage
the data.

It happens that what seems normal in my local machine (running ROOT
5.32/04) turns to be weird at farm (in principle with ROOT 5.34/05). The
problem comes from some values which are stored in the rootfile.

The simulation stores the total energy measured by a calorimeter and
this value (variable) is what is stored. With an output to screen, the
value seems to be normal, but in the rootfile, the variable becomes
strange, high or lower orders of magnitude, nonsense value, etc. It
happen with some leafs in different branches.

I tried to compile the code in the farm with a different version of ROOT
but nothing changes (I checked the dependencies of the compiled program
with the older version)

In order to check the problem, I set the value of the problematic
variable to a fix number (I tried different, 0, 10, 100) and the curious
is the output in these cases is 0 (no other value).

In my computer there is no problem, and I compiled several times to
cross-check, but I think I reach the limit of my knowledge in ROOT (or
compiling in general)

I really appreciate any suggestion/advices/tips I could try. I can
extend my explanation if it is necessary.

Thank you

Carlos Ayerbe



-------------- next part --------------
An HTML attachment was scrubbed...
URL: https://mailman.jlab.org/pipermail/root_list/attachments/20130807/817bae2e/attachment.html