Tracking Efficiency Uncertainty
17.11.2020, The charges to the task force
- Conduct systematic studies on the TPC single track efficiency uncertainty on a few standard datasets, e.g. AuAu200, BES, pp200 or 500 GeV etc. It will be preferred that the uncertainty can be evaluated more differentially in pT, eta/y, luminosity etc when statistics allows.
- Investigate on any possible improvement in the TPC simulator to improve the efficiency uncertainty, especially for the BES-II datasets.
- Expect a report from the task force (to be released as a STAR note) The report should include the above studies, as well as a description on standard procedures to evaluate the tracking efficiency uncertainty for other datasets.
- Timeline: about 6 months - 1 year. If the full efforts are expected to take longer than this, we expect a preliminary report focusing on the findings from 1) to be released in about 6 months.
Weekly meetings: 10:00 AM Eastern Time (US and Canada)
https://bnl.zoomgov.com/j/1607035037?pwd=WDVqaEdwa0RtTEM5a1VvZ24yU2gwdz09
Meeting ID: 160 703 5037
Passcode: 099371
One tap mobile
+16692545252,,1607035037#,,,,,,0#,,099371# US (San Jose)
+16468287666,,1607035037#,,,,,,0#,,099371# US (New York)
Dial by your location
+1 669 254 5252 US (San Jose)
+1 646 828 7666 US (New York)
Meeting ID: 160 703 5037
Passcode: 099371
Find your local number: https://bnl.zoomgov.com/u/aoARs6rfR
Join by SIP
1607035037@sip.zoomgov.com
Join by H.323
161.199.138.10 (US West)
161.199.136.10 (US East)
Meeting ID: 160 703 5037
Passcode: 099371
Related documents, studies, and link to tool can be found HERE.
Tracking Efficiency Uncertainty meeting
2021-07-27 10:00
2021-07-27 11:00
America/New York
Tuesday, 27 July 2021
https://bnl.zoomgov.com/j/1607035037?pwd=WDVqaEdwa0RtTEM5a1VvZ24yU2gwdz09, at 14:00 (GMT), duration : 01:00
https://bnl.zoomgov.com/j/1607035037?pwd=WDVqaEdwa0RtTEM5a1VvZ24yU2gwdz09, at 14:00 (GMT), duration : 01:00
Tracking Efficiency Uncertainty meeting
2021-07-20 10:00
2021-07-20 11:00
America/New York
Tuesday, 20 July 2021
https://bnl.zoomgov.com/j/1607035037?pwd=WDVqaEdwa0RtTEM5a1VvZ24yU2gwdz09, at 14:00 (GMT), duration : 01:00
https://bnl.zoomgov.com/j/1607035037?pwd=WDVqaEdwa0RtTEM5a1VvZ24yU2gwdz09, at 14:00 (GMT), duration : 01:00
Tracking Efficiency Uncertainty meeting
2021-07-13 10:00
2021-07-13 11:00
America/New York
Tuesday, 13 July 2021
https://bnl.zoomgov.com/j/1607035037?pwd=WDVqaEdwa0RtTEM5a1VvZ24yU2gwdz09, at 14:00 (GMT), duration : 01:00
https://bnl.zoomgov.com/j/1607035037?pwd=WDVqaEdwa0RtTEM5a1VvZ24yU2gwdz09, at 14:00 (GMT), duration : 01:00
Tracking Efficiency Uncertainty meeting
2021-07-06 10:00
2021-07-06 11:00
America/New York
Tuesday, 6 July 2021
https://bnl.zoomgov.com/j/1607035037?pwd=WDVqaEdwa0RtTEM5a1VvZ24yU2gwdz09, at 14:00 (GMT), duration : 01:00
https://bnl.zoomgov.com/j/1607035037?pwd=WDVqaEdwa0RtTEM5a1VvZ24yU2gwdz09, at 14:00 (GMT), duration : 01:00
Tracking Efficiency Uncertainty meeting
2021-06-29 11:45
2021-06-29 12:00
America/New York
Tuesday, 29 June 2021
https://bnl.zoomgov.com/j/1607035037?pwd=WDVqaEdwa0RtTEM5a1VvZ24yU2gwdz09, at 15:45 (GMT), duration : 00:15
https://bnl.zoomgov.com/j/1607035037?pwd=WDVqaEdwa0RtTEM5a1VvZ24yU2gwdz09, at 15:45 (GMT), duration : 00:15
Tracking Efficiency Uncertainty meeting
2021-06-22 10:00
2021-06-22 11:00
America/New York
Tuesday, 22 June 2021
https://bnl.zoomgov.com/j/1607035037?pwd=WDVqaEdwa0RtTEM5a1VvZ24yU2gwdz09, at 14:00 (GMT), duration : 01:00
https://bnl.zoomgov.com/j/1607035037?pwd=WDVqaEdwa0RtTEM5a1VvZ24yU2gwdz09, at 14:00 (GMT), duration : 01:00
| Time | Talk | Presenter |
|---|---|---|
| 09:00 | single track efficiency from KFP ( 00:20 ) 1 file | Petr Chaloupka |
Tracking Efficiency Uncertainty meeting
2021-06-15 10:00
2021-06-15 11:00
America/New York
Tuesday, 15 June 2021
https://bnl.zoomgov.com/j/1607035037?pwd=WDVqaEdwa0RtTEM5a1VvZ24yU2gwdz09, at 14:00 (GMT), duration : 01:00
https://bnl.zoomgov.com/j/1607035037?pwd=WDVqaEdwa0RtTEM5a1VvZ24yU2gwdz09, at 14:00 (GMT), duration : 01:00
Tracking Efficiency Uncertainty meeting
2021-06-08 10:00
2021-06-08 11:00
America/New York
Tuesday, 8 June 2021
https://bnl.zoomgov.com/j/1607035037?pwd=WDVqaEdwa0RtTEM5a1VvZ24yU2gwdz09, at 14:00 (GMT), duration : 01:00
https://bnl.zoomgov.com/j/1607035037?pwd=WDVqaEdwa0RtTEM5a1VvZ24yU2gwdz09, at 14:00 (GMT), duration : 01:00
Note:
The observed difference beween SL12 and SL13 seems to hold even when the same daq files are ruconstructed manually under SL12d_embed.
I was suggested that the likely cause is some difference between vertex finder and related track cuts introduced into SL13 library.
The observed difference beween SL12 and SL13 seems to hold even when the same daq files are ruconstructed manually under SL12d_embed.
I was suggested that the likely cause is some difference between vertex finder and related track cuts introduced into SL13 library.
| Time | Talk | Presenter |
|---|---|---|
| 09:00 | Embedding in 2012 pp vs data ( 00:20 ) 1 file |
Tracking Efficiency Uncertainty meeting
2021-06-01 10:00
2021-06-01 11:00
America/New York
Tuesday, 1 June 2021
https://bnl.zoomgov.com/j/1607035037?pwd=WDVqaEdwa0RtTEM5a1VvZ24yU2gwdz09, at 14:00 (GMT), duration : 01:00
https://bnl.zoomgov.com/j/1607035037?pwd=WDVqaEdwa0RtTEM5a1VvZ24yU2gwdz09, at 14:00 (GMT), duration : 01:00
Tracking Efficiency Uncertainty meeting
2021-05-25 10:00
2021-05-25 11:00
America/New York
Tuesday, 25 May 2021
https://bnl.zoomgov.com/j/1607035037?pwd=WDVqaEdwa0RtTEM5a1VvZ24yU2gwdz09, at 14:00 (GMT), duration : 01:00
https://bnl.zoomgov.com/j/1607035037?pwd=WDVqaEdwa0RtTEM5a1VvZ24yU2gwdz09, at 14:00 (GMT), duration : 01:00
Minutes:
Update on GMT software: David is trying to reconstruct results in DEV2 library. The code runs according to the technical note. There seems to be a half of the final fit results. It was suggested o get back to the plots that are used for the fits and check if those are filled properly.
Update on pp analyses from Robert: There is an agreement for results when the code is run on MuDST and PicoDST. It is not clear why the same code produces different results on the data that Dima produced by himself previously. It was concluded that Robert should checkout daq files for the same runs that he uses and rerun the production as Dima did.
Update on GMT software: David is trying to reconstruct results in DEV2 library. The code runs according to the technical note. There seems to be a half of the final fit results. It was suggested o get back to the plots that are used for the fits and check if those are filled properly.
Update on pp analyses from Robert: There is an agreement for results when the code is run on MuDST and PicoDST. It is not clear why the same code produces different results on the data that Dima produced by himself previously. It was concluded that Robert should checkout daq files for the same runs that he uses and rerun the production as Dima did.
| Time | Talk | Presenter |
|---|---|---|
| 09:00 | Embedding in 2012 pp vs data ( 00:20 ) 1 file |
Tracking Efficiency Uncertainty meeting
2021-05-18 10:00
2021-05-18 11:00
America/New York
Tuesday, 18 May 2021
https://bnl.zoomgov.com/j/1607035037?pwd=WDVqaEdwa0RtTEM5a1VvZ24yU2gwdz09, at 14:00 (GMT), duration : 01:00
https://bnl.zoomgov.com/j/1607035037?pwd=WDVqaEdwa0RtTEM5a1VvZ24yU2gwdz09, at 14:00 (GMT), duration : 01:00
Tracking Efficiency Uncertainty meeting
2021-05-11 10:00
2021-05-11 11:00
America/New York
Tuesday, 11 May 2021
https://bnl.zoomgov.com/j/1607035037?pwd=WDVqaEdwa0RtTEM5a1VvZ24yU2gwdz09, at 14:00 (GMT), duration : 01:00
https://bnl.zoomgov.com/j/1607035037?pwd=WDVqaEdwa0RtTEM5a1VvZ24yU2gwdz09, at 14:00 (GMT), duration : 01:00
Minutes:
Update from Robert: The issue with different distribution of primary vertex was resolved thanks to trigger selection. There is remaining difference in the number of tracks.
It seems it could be because of different vertex selection, respectively summing over different vertices.
Update on MTG: The issue with maker being called multiple time was removed. There is some difference in output tree between dev and .DEV2. It was suggested
to first look if the results in .DEV reproduce the previous work done in the .DEV2 and then revisit the dev again.
Update from Robert: The issue with different distribution of primary vertex was resolved thanks to trigger selection. There is remaining difference in the number of tracks.
It seems it could be because of different vertex selection, respectively summing over different vertices.
Update on MTG: The issue with maker being called multiple time was removed. There is some difference in output tree between dev and .DEV2. It was suggested
to first look if the results in .DEV reproduce the previous work done in the .DEV2 and then revisit the dev again.
| Time | Talk | Presenter |
|---|---|---|
| 09:00 | Embedding in 2012 pp vs data ( 00:20 ) 1 file |
Tracking Efficiency Uncertainty meeting
2021-05-04 10:00
2021-05-04 11:00
America/New York
Tuesday, 4 May 2021
https://bnl.zoomgov.com/j/1607035037?pwd=WDVqaEdwa0RtTEM5a1VvZ24yU2gwdz09, at 14:00 (GMT), duration : 01:00
https://bnl.zoomgov.com/j/1607035037?pwd=WDVqaEdwa0RtTEM5a1VvZ24yU2gwdz09, at 14:00 (GMT), duration : 01:00
| Time | Talk | Presenter |
|---|---|---|
| 09:00 | Embedding in 2012 pp vs data ( 00:20 ) 1 file | Robert Licenik |
Tracking Efficiency Uncertainty meeting
2021-04-27 10:00
2021-04-27 11:00
America/New York
Tuesday, 27 April 2021
https://bnl.zoomgov.com/j/1607035037?pwd=WDVqaEdwa0RtTEM5a1VvZ24yU2gwdz09, at 14:00 (GMT), duration : 01:00
https://bnl.zoomgov.com/j/1607035037?pwd=WDVqaEdwa0RtTEM5a1VvZ24yU2gwdz09, at 14:00 (GMT), duration : 01:00
Minutes:
Petr - update on K->3pi. Based on the the track and hit visualisation it seems that the previously observed high efficiency for matching the mother kaon to the decay vertex in th inactive parts of the TPC is caused by traccker picking up additional hits which prolong the track beyond the decay vertex. The KFPF is then unable to match the track when trying to project it forward hence the decrease in the efficiency.
GMT- David has changed the order of commands in BFC which likely removed the previously seen problem.
Robert - The comparisons between his and DIma's embedding shows same results. Now the question is if there is a possible difference in real data. Robert was ask to run his code on the data that Dima previously reprocessed for his study.
Petr - update on K->3pi. Based on the the track and hit visualisation it seems that the previously observed high efficiency for matching the mother kaon to the decay vertex in th inactive parts of the TPC is caused by traccker picking up additional hits which prolong the track beyond the decay vertex. The KFPF is then unable to match the track when trying to project it forward hence the decrease in the efficiency.
GMT- David has changed the order of commands in BFC which likely removed the previously seen problem.
Robert - The comparisons between his and DIma's embedding shows same results. Now the question is if there is a possible difference in real data. Robert was ask to run his code on the data that Dima previously reprocessed for his study.
| Time | Talk | Presenter |
|---|---|---|
| 09:00 | Note on K->3pi with KFP ( 00:20 ) 1 file | |
| 09:20 | Embedding in 2012 pp vs data ( 00:20 ) 1 file |
Tracking Efficiency Uncertainty meeting
2021-04-20 10:00
2021-04-20 11:00
America/New York
Tuesday, 20 April 2021
https://bnl.zoomgov.com/j/1607035037?pwd=WDVqaEdwa0RtTEM5a1VvZ24yU2gwdz09, at 14:00 (GMT), duration : 01:00
https://bnl.zoomgov.com/j/1607035037?pwd=WDVqaEdwa0RtTEM5a1VvZ24yU2gwdz09, at 14:00 (GMT), duration : 01:00
Minutes:
GMT - David is looking into the number of clusters produced by the cluster maker. It seem that the maker is run multiple times in the chain. This could explain the difference between dev and .DEV2.
Robert has rerun the code from Dima including embedding, but sees a difference (see presentation). The only difference seema to be that Dima has run his own data production. However that does not seem as a plausible explanation. It was suggested for Robert to run his code on the embedding that Dima used previously to see if there is also the difference.
GMT - David is looking into the number of clusters produced by the cluster maker. It seem that the maker is run multiple times in the chain. This could explain the difference between dev and .DEV2.
Robert has rerun the code from Dima including embedding, but sees a difference (see presentation). The only difference seema to be that Dima has run his own data production. However that does not seem as a plausible explanation. It was suggested for Robert to run his code on the embedding that Dima used previously to see if there is also the difference.
| Time | Talk | Presenter |
|---|---|---|
| 09:00 | Embedding in 2012 pp vs data ( 00:20 ) 1 file |
Tracking Efficiency Uncertainty meeting
2021-04-13 10:00
2021-04-13 11:00
America/New York
Tuesday, 13 April 2021
https://bnl.zoomgov.com/j/1607035037?pwd=WDVqaEdwa0RtTEM5a1VvZ24yU2gwdz09, at 14:00 (GMT), duration : 01:00
https://bnl.zoomgov.com/j/1607035037?pwd=WDVqaEdwa0RtTEM5a1VvZ24yU2gwdz09, at 14:00 (GMT), duration : 01:00
Tracking Efficiency Uncertainty meeting
2021-04-06 10:00
2021-04-06 11:00
America/New York
Tuesday, 6 April 2021
https://bnl.zoomgov.com/j/1607035037?pwd=WDVqaEdwa0RtTEM5a1VvZ24yU2gwdz09, at 14:00 (GMT), duration : 01:00
https://bnl.zoomgov.com/j/1607035037?pwd=WDVqaEdwa0RtTEM5a1VvZ24yU2gwdz09, at 14:00 (GMT), duration : 01:00
Tracking Efficiency Uncertainty meeting
2021-03-30 10:00
2021-03-30 11:00
America/New York
Tuesday, 30 March 2021
https://bnl.zoomgov.com/j/1607035037?pwd=WDVqaEdwa0RtTEM5a1VvZ24yU2gwdz09, at 14:00 (GMT), duration : 01:00
https://bnl.zoomgov.com/j/1607035037?pwd=WDVqaEdwa0RtTEM5a1VvZ24yU2gwdz09, at 14:00 (GMT), duration : 01:00
Minutes:
David has presented results from running the GMT code:
The entries in the final ntuple are coming from tracks being matched to MTD hits. The question now is if the actual number of hits reconstructed in the detector is also different between the libraries.
David has presented results from running the GMT code:
There are three attached sets of slides. Each compares the TTrees generated with a bfc chain:
- https://drupal.star.bnl.gov/STAR/system/files/gmt_lib_check.pdf compares trees generated under dev & .DEV2, in which the .DEV2 run includes a bfc chain option for KFVertex
- https://drupal.star.bnl.gov/STAR/system/files/DEV2_v2_noKFVertex.pdf is the same as 1. above, but .DEV2 does not include the bfc chain option for KFVertex
- https://drupal.star.bnl.gov/STAR/system/files/DEV2_KFVertex_effect.pdf compares the .DEV2 with the KFVertex & .DEV2 without the KFVertex
The entries in the final ntuple are coming from tracks being matched to MTD hits. The question now is if the actual number of hits reconstructed in the detector is also different between the libraries.
Tracking Efficiency Uncertainty meeting
2021-03-23 10:00
2021-03-23 11:00
America/New York
Tuesday, 23 March 2021
https://bnl.zoomgov.com/j/1607035037?pwd=WDVqaEdwa0RtTEM5a1VvZ24yU2gwdz09, at 14:00 (GMT), duration : 01:00
https://bnl.zoomgov.com/j/1607035037?pwd=WDVqaEdwa0RtTEM5a1VvZ24yU2gwdz09, at 14:00 (GMT), duration : 01:00
Minutes:
1. Yuri has done a transfer of 7.7GeV data to disk space leased by PAC that we can use for the studies at /gpfs01/star/pwg_tasks/TF_TrkEff
2. MTD software is now running in both .DEV2 and dev and returns data trees. Only problem remains that in dev there is a spurious crash at the end after the data is saved. David is now
comparing the outputs.
3. Yang He has presented QA studies of embedding in isobar data https://drupal.star.bnl.gov/STAR/system/files/isobar_embeding_testrun_QA_March18.pdf.
There is a strange pt-dependence observed in efficiency. It was discussed and suggest to look into 2D eta-phi dependence to see if it is related to the bad TPC sector.
It was discussed the effect effect of missing TPC sector could show up through nHits cut.
It was also suggested to try simulation with realistic pt spectra since some effects could be due to reweighting.
1. Yuri has done a transfer of 7.7GeV data to disk space leased by PAC that we can use for the studies at /gpfs01/star/pwg_tasks/TF_TrkEff
2. MTD software is now running in both .DEV2 and dev and returns data trees. Only problem remains that in dev there is a spurious crash at the end after the data is saved. David is now
comparing the outputs.
3. Yang He has presented QA studies of embedding in isobar data https://drupal.star.bnl.gov/STAR/system/files/isobar_embeding_testrun_QA_March18.pdf.
There is a strange pt-dependence observed in efficiency. It was discussed and suggest to look into 2D eta-phi dependence to see if it is related to the bad TPC sector.
It was discussed the effect effect of missing TPC sector could show up through nHits cut.
It was also suggested to try simulation with realistic pt spectra since some effects could be due to reweighting.
| Time | Talk | Presenter |
|---|---|---|
| 09:00 | Isobar embedding QA ( 00:20 ) 0 files | Yang He |
Tracking Efficiency Uncertainty meeting
2021-03-16 10:00
2021-03-16 11:00
America/New York
Tuesday, 16 March 2021
https://bnl.zoomgov.com/j/1607035037?pwd=WDVqaEdwa0RtTEM5a1VvZ24yU2gwdz09, at 14:00 (GMT), duration : 01:00
https://bnl.zoomgov.com/j/1607035037?pwd=WDVqaEdwa0RtTEM5a1VvZ24yU2gwdz09, at 14:00 (GMT), duration : 01:00