sPHENIX discussion of software

[pinkenburg <pinkenburg AT bnl.gov>]

Hi folks,

we have two interactive hosts for sphenix - sphenix01 and sphenix02. 
They are identical to the other interactive nodes but we can fine tune 
this (within limits - 100GB jobs won't run) .

The main reason for their existence is that it depends on the submission 
machine into which queue the jobs get submitted. Jobs submitted from 
these machines go into the sphenix condor queue which has its own 
allocation of condor slots.

There seems to be an issue with condor with large memory allocation 
which prevents sphenix jobs to start quickly. They do start but I have 
seen delays of 24 hours. Since this is a separate queue queue we might 
be able to do something about that on the queue level (at least there 
will be no low memory phenix jobs out-competing the sphenix jobs).

Please give this a shot, lets see how this goes, especially in terms of 
jobs starting with a reasonable waiting period

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

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