sPHENIX discussion of software

[pinkenburg <pinkenburg AT bnl.gov>]

Hi folks,

I did send a mail about being mindful of your condor jobs yesterday:

https://lists.bnl.gov/pipermail/sphenix-software-l/2023-July/001215.html

and I did go over the example condor job file in our wiki in the 
software meeting, because this is about to cause real pain if this 
continues. You have a lot of resources at your hands - with this 
capability comes responsibility to make efficient use of them.

From what I can tell there are some submission scripts floating around 
which just do the wrong thing.  Even if you got some script handed to 
you - this does not absolve you from your responsibility to make sure it 
does the right thing.

This also means to make sure your memory request is accurate. If you 
request 12GB but your job only needs 6GB, you are preventing other user 
jobs from running because condor will not over-subscribe the memory of 
our nodes and therefore will leave some cores idling. This is not too 
important for a few jobs but for a large number of jobs this easily 
forces significant parts of our farm to idle.

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

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