HFT embedding, status and plans.

Speaker : HFT embedding, status and plans.


Talk time : 11:30, Duration : 00:20

 Please comment.

1. Verification. A couple months ago Xin checked HFT misalignment 
using Cosmics sample. He has found a problem with SSD.
The problem has been fixed, new sample has been generated. 
No new problem has been reported. We are still waiting for verification.
But for now I assume that the HFT misalignment is done properly.

2. Slow simulators. The digit (pixel) level embedding requires 
Slow Simulators.
  a. Pxl => StPxlSlowSim does not exist. DIGMAPSSim has been tuned for STAR data 
by Jakub Kvapil. But there is no code of StPxlSlowSim to use it. 
It will require significant efforts (~1 month) to do this.
  b. Ist => StIstSlowSimMaker does exist. It has to be tested and tuned.
  c. Sst => StSsdSimulationMaker does exist. It has to be tested and tuned for Sst.

3. Fast simulators for embedding. The hit level embedding can be done with assumptions:
   a. Detector efficiencies are close to 100%.
   b. Occupancies  are very low (<1%), and
   c. All inefficiencies are coming from list of rejected pixels / strips.
Besides the rejection of hot/dead pixels and strips we do have this embedding right now.
Adding the rejection it will take a few days.
  I propose to take the approach to use hit level embedding for now and introduce pixel level 
embedding for each system as soon as the slow simulators will be ready. This will allow to understand
systematics coming from hit level embedding.

4. For embedding and efficiency study I propose 
  a. to use 2016 AuAu200 days 125-134 data sample, ~ 1.5 Mevents.
  b. to use HLT cluster (~500 cpus and ~30 TB)

5. Data samples
  a. MuDst processed with the standard chain
/net/l404/data/fisyak/reco/2016/AuAu200_adc
  b. The sample of D0 with  
dN/dmT^2 = exp(-mT/T) pT generation with T = 0.457 GeV/c
-1 < eta < 1
0 < phi < 6.28319
-6 < zVer< 6
and 
1 < pT < 10 GeV/c
force to decay D0->K0s pi+ pi for Ideal and Real geometries
/net/l404/data/fisyak/reco/2016/D0K0s2pimTsqIdeal
/net/l404/data/fisyak/reco/2016/D0K0s2pimTsqReal
  c. D0->K-pi+
/net/l404/data/fisyak/reco/2016/D0KpimTsqIdeal
/net/l404/data/fisyak/reco/2016/D0KpimTsqReal
  d. Lc -> 3pi Lambda, 2 < pT < 10 GeV/c
/net/l404/data/fisyak/reco/2016/Lc3pimTsqIdeal
/net/l404/data/fisyak/reco/2016/Lc3pimTsqReal
  e. Lc -> K proton pi 2 < pT < 10 GeV/c
/net/l404/data/fisyak/reco/2016/LcKppimTsqIdeal
/net/l404/data/fisyak/reco/2016/LcKppimTsqReal

6. After adding the rejection of bad/hot strips and pixel 
I am planning to reproduce these samples for single particles and
to do embedding these decays into the above AuAu200_adc sample.

7. Maksym Zyzak and Iouri Vasiliev (CMB/GSI) will continue turning cuts 
and estimation efficiencies for both Lc samples.

8. Petr Chaloupka and Pavol Federic (Nuclear Physics Institute, The Czech Academy of Sciences), 
I hope, they will take care about D0->K0s2pi channel. They are coming to GSI next week to learn
how to use Maksym and Iouri method.

9. The idea is to use KFParticle get a reasonable signal from data with known efficiencies.
The good check point is joint CBM and STAR meeting in China, middle of September.