sPHENIX discussion of software

[pinkenburg <pinkenburg AT bnl.gov>]

Hi folks,

besides the perennial reminder to write the userlog to /tmp (in a form 
which makes it unique - means prepend your username), please write the 
output and error file to gpfs (e.g. /sphenix/user). They are small and 
get continuously written to - something lustre (/sphenix/tg/tg01) just 
isn't good at.  Sometimes jobs seem to die because it takes too long to 
open a log in lustre.

Also keep in mind that the phenix interactive nodes (rcas2061-rcas2068) 
submit condor jobs to the shared pool which has more competition and 
inferior resources than sPHENIX, especially if the jobs need more than 
the 2GB memory which is the standard for the shared pool.

Thanks,
Chris

--
*************************************************************

Christopher H. Pinkenburg       ;    pinkenburg AT bnl.gov
                                ;    http://www.phenix.bnl.gov/~pinkenbu

Brookhaven National Laboratory  ;    phone: (631) 344-5692
Physics Department Bldg 510 C   ;    fax:   (631) 344-3253
Upton, NY 11973-5000

*************************************************************