sPHENIX is a new detector at RHIC.

[Anthony Frawley <afrawley AT fsu.edu>]

Hi Justin,

The tracking is using a lot of memory right now. Haiwang is working on understanding it. The tracking alone uses 15-20+ GB of memory at present.

If you don't need tracking, I would certainly turn it off until we get it fixed.

Tony

---

From: Frantz, Justin <frantz AT ohio.edu>
Sent: Monday, July 10, 2017 5:04 PM
To: Anthony Frawley; Chris Pinkenburg; sphenix-l AT lists.bnl.gov
Subject: RE: [Sphenix-tracking-l] highmem queue changes

 

Hi all,

 

When working on the EMC clustering I  was experiencing something similar since before the collab mtg. Since then we’ve been avoiding work needing high statistics, and hoping I could eventually just solve the issue by turning off something, like the tracking since I don’t need it.  Will that or something similar (please suggest?) work to just lower the mem usage (and thus benefit from more available machines)?

 

-Justin

----------------------------------------------
Justin Frantz, Ph. D. 

RHIC/AGS User Executive Committee Outgoing Chair

Brookhaven National Lab

Associate Professor
Ohio University Dept. Of Physics and Astronomy  
frantz AT ohio.edu  
646-228-2539 

 

From: sPHENIX-tracking-l [mailto:sphenix-tracking-l-bounces AT lists.bnl.gov] On Behalf Of Anthony Frawley
Sent: Monday, July 10, 2017 4:22 PM
To: Chris Pinkenburg; sphenix-tracking-l AT lists.bnl.gov
Subject: Re: [Sphenix-tracking-l] highmem queue changes

 

Hi Chris,

 

I am running some central Hijing jobs now, and it looks like quite a few are failing. Do we have to request the new queue specifically?

 

I think 30 GB should likely be enough.

 

Thanks

Tony

 

---

From: sPHENIX-tracking-l <sphenix-tracking-l-bounces AT lists.bnl.gov> on behalf of Chris Pinkenburg <pinkenburg AT bnl.gov>
Sent: Monday, July 10, 2017 11:33 AM
To: sphenix-tracking-l AT lists.bnl.gov
Subject: [Sphenix-tracking-l] highmem queue changes

 

Hi folks,

we just adjusted the highmem queue parameters. It now lets jobs grow to
30GB before evicting them (if someone could give me some idea if this is
enough I would appreciate it). On the downside the number of condor
slots is drastically reduced since only a subset of our machines can run
something like this (one job on 64+GB memory machines, 2 on 128+GB machines)

Chris

--

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

Christopher H. Pinkenburg      ; pinkenburg AT bnl.gov

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

http://www.phenix.bnl.gov/~pinkenbu/

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

_______________________________________________
sPHENIX-tracking-l mailing list
sPHENIX-tracking-l AT lists.bnl.gov
https://lists.bnl.gov/mailman/listinfo/sphenix-tracking-l