[Hpc-forum] Only for Gaussian09 users: revised subg09 (5.0)

[Ferenc Bartha]

Dear G09 Users,



A new revision (5.0) of subg09 has been installed instead of the previous 
4.0 script.



1) Why to change?

The good old subg09 (version 3.0 is still available as subg09_old) can be 
used but not recommended. The main reason is that default and hard memory 
and time limits have been introduced on the NIIF computers (e.g. soft 2.5G 
memory/slot, 1 week runtime) which you could not overwrite with subg09 and 
your job was killed. The new revision gives you a full control on these 
complexes.

2) What was the problem, with the recently published 4.0 version?

Although, now you could control these variables, there were several jobs, 
which died because of requesting more memory than reserved for the job for 
larger calculations. It could have been solved with the -r option within the 
4.0 version, however, it is better to have safer defaults and not to ask 
users to use advanced options.

3) Is revision 5.0 well tested and stable?

No. I cannot test it with all its new advanced features. Please let me know, 
if something went wrong!

You can always make a copy (maybe renamed) in your ~/bin directory and 
modify the script according to your taste. Also in this case, please let me 
know, if some modifications are worth to do in the public version.

In the new version a notification about version number is sent to the user. 
This is now 5.0, if I have to make minor modifications; I overwrite the 
number, keep an eye on it. I shall write e-mail only in the case of a major 
revision.



Good luck, Feri
------------:
Dr. Ferenc BARTHA, tel: 62/54-6821, E-mail: barthaf at sol.cc.u-szeged.hu
SKYPE: ferenc.bartha, WWW: http://www.staff.u-szeged.hu/~barthaf/
SZTE DNT - High Performance Computing Group, 6725 Szeged, Szikra u. 2.
SZTE, Department of Medical Chemistry, 6720 Szeged, Dóm tér 8.