sPHENIX tracking discussion

[sourav Tarafdar <sourav.pheonix AT gmail.com>]

Hello again,

Sorry for not being enough clear in my last email. Yesterday my jobs were over mid central Hijing events. Just today while pushing through condor over most central Hijing events I am also having the issue of memory getting over 20 GB for quite a number of jobs. 

Best regards,

-Sourav

On Jul 3, 2017, at 10:53 AM, sourav Tarafdar <sourav.pheonix AT gmail.com> wrote:

Hi Christof and Tony,

Since yesterday I already did three iterations of Hijing embedded 100 muons per event. I didn’t really face any memory related issue. I submitted 1000 condor jobs in each iterations with 2 Hijing events per job. I have been registering only embedded muons. The phase space of the muons are 1. < pT < 50 GeV, 2 units of eta , and 2*pi in phi. 

So far I also didn’t notice the issue in clustering pointed by Sookhyun. I see about 0.94 clusters per true track-layer in pure muon events where muons have the same phase space as quoted by me for Hijing embedded muons.

Best regards,

-Sourav

On Jul 3, 2017, at 10:31 AM, Christof Roland <christof.roland AT cern.ch> wrote:

Hi Tony, 

thanks a lot for your response. Good to see that some events actually do finish. 

I'll try a diffent test event ans see if I get some result there. 

I am not sure I can reproduce the clustering problem. 

With the current setup I get about 0.94 hits per particle-layer crossing

for muon events. Did Sookhyun run the latest greatest macro?

Cheers

   Christof 

On 3. Jul 2017, at 17:07, Anthony Frawley <afrawley AT fsu.edu> wrote:

Hi Christof,

Over the weekend I have been running 2000 jobs at a time with central Hijing+ 100 pions + 1 upsilon. About 1400-1500 of them do not reach the 20 GB limit and finish within ~ 1.5 hours. Of the rest, some crash (~50), ~250 are put on hold when they get too big, and ~ 200 are running but are just taking too long. So after ~ 1500 jobs finish I "condor_rm" the rest and start another 2000. 

It is not very satisfactory, but I need to get these done so I can make performance plots that are needed now. We need to solve the clustering issue pointed out by Sookhyun, since that is likely the problem.

Tony

---

From: sPHENIX-tracking-l <sphenix-tracking-l-bounces AT lists.bnl.gov> on behalf of Christof Roland <christof.roland AT cern.ch>
Sent: Monday, July 3, 2017 10:25 AM
To: sphenix-tracking-l AT lists.bnl.gov
Subject: [Sphenix-tracking-l] Did anyone manange to run Hijing lately?

 

Hi Everybody, 

even with the more recent change to the TPC macro I am still having trouble running Hijing events. 
The jobs blow up to 20GB and then just hang. These are jobs that just do run simulation up to cluster 
level and then write to disk.

Did anyone mange to run any hijing events to the end lately?

Also in the log file root throws an error:
Error in <TBufferFile::WriteByteCount>: bytecount too large (more than 1073741822)
Has anyone seen these?

Thanks for your input

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

sPHENIX-tracking-l Info Page - lists.bnl.gov

lists.bnl.gov

To see the collection of prior postings to the list, visit the sPHENIX-tracking-l Archives. Using sPHENIX-tracking-l: To post a message to all the ...

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

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

Attachment: smime.p7s
Description: S/MIME cryptographic signature