[Hps-analysis] Too few Pass6 Moller events in MC

Fri Mar 11 22:02:27 EST 2016

It looks like that fixed the problem.

-Brad

________________________________
From: hps-software at SLAC.STANFORD.EDU <hps-software at SLAC.STANFORD.EDU> on behalf of McCormick, Jeremy I. <jeremym at slac.stanford.edu>
Sent: Friday, March 11, 2016 5:54 PM
To: Maurik Holtrop
Cc: hps-software
Subject: Re: Too few Pass6 Moller events in MC

It should be just waiting on a new lcsim build to deploy....

Sent using OWA for iPhone
________________________________
From: Maurik Holtrop <maurik at physics.unh.edu>
Sent: Friday, March 11, 2016 2:50:16 PM
To: McCormick, Jeremy I.
Cc: Bradley Yale; hps-software
Subject: Re: Too few Pass6 Moller events in MC

Hello Jeremy,

Seems you broke the trunk. Then new JobManager class extends a private method in lcsim JobControlManager called setupDrivers(). Seems you cannot do that.

Best,
Maurik

On Mar 11, 2016, at 4:55 PM, McCormick, Jeremy I. <jeremym at slac.stanford.edu<mailto:jeremym at slac.stanford.edu>> wrote:

Hi,

I deleted some code related to conditions setup (ConditionsDriver stuff which is used all over the place in MC) when I did my “big merge” a few weeks ago.  Apologies for that.  Think it is fixed now.

Old behavior should be restored now if you run from trunk e.g.

./scripts/job-manager.sh -DoutputFile=output -Drun=5227 -Ddetector=HPS-EngRun2015-Nominal-v3 -i./input.slcio -n10 -r /org/hps/steering/recon/EngineeringRun2015FullReconMC.lcsim

Unlikely this would affect data at all, as the ConditionsDriver is not used in the recon jobs.

You should use 3.7-SNAPSHOT for MC data processing I guess (until the next release).

—Jeremy

On Mar 11, 2016, at 11:24 AM, Omar Moreno <omoreno1 at ucsc.edu<mailto:omoreno1 at ucsc.edu>> wrote:

Hi Jeremy,

Did you ever test your changes to the conditions DB using Monte Carlo?

--Omar Moreno

On Fri, Mar 11, 2016 at 11:22 AM, Bradley T Yale <btu29 at wildcats.unh.edu<mailto:btu29 at wildcats.unh.edu>> wrote:

For clarity, it looks like tpass6 actually used 'hps-distribution-3.6-20160229.015546-43-bin.jar' not the '3.6 release', but the same run number problem happens for that jar as well.

________________________________
From: hps-software at SLAC.STANFORD.EDU<mailto:hps-software at SLAC.STANFORD.EDU> <hps-software at SLAC.STANFORD.EDU<mailto:hps-software at SLAC.STANFORD.EDU>> on behalf of Bradley T Yale <btu29 at wildcats.unh.edu<mailto:btu29 at wildcats.unh.edu>>
Sent: Friday, March 11, 2016 1:24 PM
To: hps-software at SLAC.STANFORD.EDU<mailto:hps-software at SLAC.STANFORD.EDU>; hps-analysis at jlab.org<mailto:hps-analysis at jlab.org>; Sebouh Paul
Subject: Re: Too few Pass6 Moller events in MC

Investigating further, this may be the problem:

Apparently, the most recent jar (3.6 release) always sets the run number to '0', even when specifying it in the command with -Drun=5772.

All jars that were checked before the problematic 3.6.1 appear to assign the correct run number of 5772.

It should be checked if this also affects data reconstructed with 3.6, or only Monte Carlo. In any case, this would definitely be a thing to fix.

-Brad

________________________________
From: hps-software at SLAC.STANFORD.EDU<mailto:hps-software at SLAC.STANFORD.EDU> <hps-software at SLAC.STANFORD.EDU<mailto:hps-software at SLAC.STANFORD.EDU>> on behalf of Bradley T Yale <btu29 at wildcats.unh.edu<mailto:btu29 at wildcats.unh.edu>>
Sent: Thursday, March 10, 2016 7:06 PM
To: hps-software at SLAC.STANFORD.EDU<mailto:hps-software at SLAC.STANFORD.EDU>; hps-analysis at jlab.org<mailto:hps-analysis at jlab.org>; Sebouh Paul
Subject: Too few Pass6 Moller events in MC

I know this was supposed to be the "final pass",

but it looks like there are far too few Moller events being readout in Pass 6 Monte Carlo.

Looking further into this, it is because the >3-hit clusters in Pass 6 are getting decimated BEFORE the trigger for some reason (see attached plots), and so they no longer pass the cluster hit count threshold. These were made using exactly the same filtered SLIC events, and the same steering file with the same apparent thresholds in the GTP clusterer and trigger.

The difference must be elsewhere in the drivers themselves.

These plots compare

hps-distribution-3.5-20151218.205540-15-bin.jar

with

hps-distribution-3.6-bin.jar

both using the 'EngineeringRun2015TrigSingles1_Pass2.lcsim' steering file.

A more recent 3.5 jar (1/25/2016) gives a similar number of events as Pass 4, and so the key difference must be with 3.6.

-Brad

________________________________

Use REPLY-ALL to reply to list

To unsubscribe from the HPS-SOFTWARE list, click the following link:
https://listserv.slac.stanford.edu/cgi-bin/wa?SUBED1=HPS-SOFTWARE&A=1

________________________________

Use REPLY-ALL to reply to list

To unsubscribe from the HPS-SOFTWARE list, click the following link:
https://listserv.slac.stanford.edu/cgi-bin/wa?SUBED1=HPS-SOFTWARE&A=1

________________________________

Use REPLY-ALL to reply to list

To unsubscribe from the HPS-SOFTWARE list, click the following link:
https://listserv.slac.stanford.edu/cgi-bin/wa?SUBED1=HPS-SOFTWARE&A=1

________________________________

Use REPLY-ALL to reply to list

To unsubscribe from the HPS-SOFTWARE list, click the following link:
https://listserv.slac.stanford.edu/cgi-bin/wa?SUBED1=HPS-SOFTWARE&A=1

________________________________

Use REPLY-ALL to reply to list

To unsubscribe from the HPS-SOFTWARE list, click the following link:
https://listserv.slac.stanford.edu/cgi-bin/wa?SUBED1=HPS-SOFTWARE&A=1
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <https://mailman.jlab.org/pipermail/hps-analysis/attachments/20160312/c08a6627/attachment-0001.html>