The ABINIT Users Mailing List ( CLOSED )

["Allan, Douglas C Dr" <AllanDC@Corning.com>]

Abinit uses dynamic allocation of arrays that have natom as an index, and 
defines "mxnatom" as the maximum value of natom for the set of input values.  
There should be no limit on the number of atoms based on any array dimension 
related to natom, unless there is a bug ....

From your description below, it sounds like a bug.  The code is written to 
give a more explicit error message relating to array dimensions if you ever 
exceed one, although that should not happen because of dynamic allocation.  
Note that you can peruse the sources yourself and add additional debugging 
output to try to trace down the failure.  There is no internal limit (at 
least not a consciously coded one), but there may well be a bug.

Doug Allan

-----Original Message-----
From: Aaron Deskins [mailto:aarondesk@yahoo.com] 
Sent: Thursday, June 26, 2003 12:32 PM
To: forum@abinit.org
Subject: [abinit-forum] Abinit natom limit


Is there a limit to the number of atoms that Abinit
can calculate? I'm trying to perform a single point
energy calculation for a supercell containing 40 Pt
atoms. I've been able to get 20 Pt atoms to work, but
the program crashes above this. A 30-atom job requires
132.382 Mbytes of memory and approximately 90 MB of
disk space. I have more than enough of memory and disk
space available. The program crashes on two different
IBM SP clusters with the error message: 

1525-002 The I/O statement on the internal file cannot
be completed because the end of the internal file was
reached.  The program will stop.

This message is repeated a few times. 
I've also tried the job on a DEC Alpha machine, but it
also crashes, which leads me to believe that there is
a bug or internal limit with Abinit. 
Is this so? Any help is appreciated. 

Aaron 


__________________________________
Do you Yahoo!?
SBC Yahoo! DSL - Now only $29.95 per month! http://sbc.yahoo.com