Tue Feb 25
=========
THelix result with HFT etc. Problem with HFT... hits
=========================================

PURE DEV, THelixDEV
====================

In File(s) 0.log
nFiles =1 nEvents = 5 <nEvents>=5
nGlob =       5823 <nGlob> =   1164.60 nGlobMax= 1567
nGgoo =       4293 <nGgoo> =    858.60 nGgooMax= 1181 Ggoo/Gtot=0.737
nPrim =       1658 <nPrim> =    331.60 nPrimMax=  486 Ptot/Gtot=0.285
nPgoo =       1363 <nPgoo> =    272.60 nPgooMax=  399 Pgoo/Ggoo=0.317

Tpc Hits =       231101 used 132145 (57.18%) perTrack 22.694
 Sst Hits =             176 used     35 (19.89%) perTrack 0.006
 Pxl Hits =         40932 used   171 ( 0.42%) perTrack 0.029
 Ist Hits =               219 used     89 (40.64%) perTrack 0.015

Tpc Gits =      231101 used 117679 (50.92%) perTrack 27.412
Sst Gits =             176 used          35 (19.89%) perTrack 0.008
Pxl Gits =         40932 used        171 ( 0.42%) perTrack 0.040
Ist Gits =               219 used          89 (40.64%) perTrack 0.021

THelix.DEV + other DEV
======================
nFiles =1 nEvents = 5 <nEvents>=5
nGlob =       5811 <nGlob> =   1162.20 nGlobMax= 1572
nGgoo =       4271 <nGgoo> =    854.20 nGgooMax= 1181 Ggoo/Gtot=0.735
nPrim =       1686 <nPrim> =    337.20 nPrimMax=  441 Ptot/Gtot=0.290
nPgoo =       1378 <nPgoo> =    275.60 nPgooMax=  363 Pgoo/Ggoo=0.323

Tpc Hits =      231101 used 131672 (56.98%) perTrack 22.659
Sst Hits =             176 used            5 ( 2.84%) perTrack 0.001
Pxl Hits =         40932 used          17 ( 0.04%) perTrack 0.003

Tpc Gits =      231101 used 117295 (50.75%) perTrack 27.463
Sst Gits =             176 used            5 ( 2.84%) perTrack 0.001
Pxl Gits =         40932 used          17 ( 0.04%) perTrack 0.004

Thu Feb 20
=========

Thu Sep 12 2019
==============

Thu Aug 8 2019
=============

First track
--------------

Then the same track again:

Thu May 16 2019
==============
It was found that there are double hits in simulation.
file hijing_988_fieldon_ideal_sdt20160517_5evts.fzd

So you see twins splited by Z. I checked fz file, no twins. So they are created in  StTpcHitMaker.
I tested in DEV and SL18h. They are there.
Then I got event file:
/star/rcf/simu/auau200/hijing_382/b0_3/central/y2007/gheisha_on/event/rcf1296_02_100evts.event.root
It was created starreco 352373037 Oct  7  2013

You see, that not accepted hits actually are better than accepted ones. These additional hits really destroy the quality.

======================================================================================================================
Thu Apr11 2019
============

file rcf1296_02_100evts.event.root
------------------------------------------------

fil = "hijing_988_fieldon_ideal_sdt20160517_5evts.event.root";
---------------------------------------------------------------------------------------

Thu Feb14 2019
=============
After long attempts was fixed matrix of derivatives & direction changes for TRungeKutta
In result we have:

opt = "P2016a,Sti,btof,mtd,pxlHit,istHit,sstHit,picoWrite,PicoVtxVpd,BEmcChkStat,QAalltrigs,CorrX,OSpaceZ2,OGridLeak3D,-hitfilt";
fil = "/star/rcf/test/daq/2016/135/st_physics_17135012_raw_2000025.daq";
=====================================================================
dotDEV Sti
Feb14 2019
=====================================================================
I <nGlob> =   1181.33
 <nGgoo> =    881.00
 <nPrim> =     317.33
 <nPgoo> =    265.33

Tpc Hits =    140917 used     80820(57.35%) perTrack 22.805
Sst Hits =       109 used        17(15.60%) perTrack 0.005
Pxl Hits =     24357 used        90( 0.37%) perTrack 0.025
Ist Hits =        97 used        42(43.30%) perTrack 0.012

Tpc Gits =    140917 used     72100(51.16%) perTrack 27.280
Sst Gits =       109 used        17(15.60%) perTrack 0.006
Pxl Gits =     24357 used        90( 0.37%) perTrack 0.034
Ist Gits =        97 used        42(43.30%) perTrack 0.016

=====================================================================
DEV Sti
Feb14 2019
=====================================================================

 <nGlob> =   1182.33
 <nGgoo> =    882.67
 <nPrim> =    318.67
<nPgoo> =    262.33

Tpc Hits =    140917 used     80860(57.38%) perTrack 22.797
Sst Hits =       109 used         2( 1.83%) perTrack 0.001
Pxl Hits =     24357 used         5( 0.02%) perTrack 0.001

Tpc Gits =    140917 used     72236(51.26%) perTrack 27.279
Sst Gits =       109 used         2( 1.83%) perTrack 0.001
Pxl Gits =     24357 used         5( 0.02%) perTrack 0.002

============================================================================

FIXED derivatives + backward() TkDir + cut factor removed

.DEV/StvKutta  Tue Feb 13 2019

============================================================================
<nGlob> =   1545.00
<nGgoo> =    722.67
<nPrim> =    448.33
<nPgoo> =    287.00
Tpc Hits =    140917 used     71994(51.09%) perTrack 15.533
Ist Hits =        97 used        61(62.89%) perTrack 0.013
Pxl Hits =     24357 used       138( 0.57%) perTrack 0.030

Tpc Gits =    140917 used     61044(43.32%) perTrack 28.157
Ist Gits =        97 used        48(49.48%) perTrack 0.022
Pxl Gits =     24357 used       103( 0.42%) perTrack 0.048

	gGlob	gPrim	TpcHits/tk	HftHits/tk	N Hft hits
Sti  DEV	883	262	27.28	0.003	7
Sti  dotDEV	881	265	27.28	0.056	149
StvKutta	723	287	28.16	0.070	151

So StvKutta  sometimes worse , sometimes better than Sti.
So now it is comparable. Sure, additional tuning is needed.

Thu Aug23 2018
=============

                                Seed image
                               ==========

  

                        Track image, see missed hits
                        =============================
   

    Now seed image for StvEval:
   ======================

        And track image for StvEval
       ======================

Thu Aug16 2018
=============

p { margin-bottom: 0.1in; line-height: 120%; }

1. Chi2 always bigger than Sti & StvEval

2. Parameter differences usually worse than Sti & StvEval,

BUT for Q/Pt is better than StvEval and close to Sti

Fig31-34

 

3. Difference fpt tan(Lamda) the same for all (Fig 36)

 

4. DcaZ all very close(Fig 42)

 

5. Efficiency Sti 5% better StvEval 5% better StvKutta (fig 168)

From 200 jobs 10 events in each 33 failed.

========================================================================================================================

Thu Nov30 2017
============
   Test made on file /star/rcf/test/daq/2016/135/st_physics_17135012_raw_2000025.daq
   Only 3 events. More statisics is needed

StvKutta is working but some results worse than StvOld, some better. For sure better than Sti

The StvOld Xi2(Pt)

Now the same StvKutta: More flat

StvOld Xi2(Lamda)

And the same Xi2(Lamda) in StvKutta:

    Now the same logic but nHits instead of Xi2:

   StvOld nHits(Pt);

  
      StvKutta nHits(Pt)

      Now StvOld nHits(lamda)

   And the last one StvKutta nHits(Lamda)

	nGlob	nG>15	nPrim	nP>15	nHits	HitUsage
Sti	1194	537	298	239	22	56%
StvOld	1526	862	558	370	19	61%
StvKutta	1844	725	489	261	13	51%

                                                      Summary

 1. StvOld Xi2 increasing with Pt but StvKutta is flat.

 2. Both StvOld and SkvKutta  no dependency Xi2 from lambda.

 3. StvOld nHits(Pt) decreasing with Pt but in StvKutta flat
 4. nHits(Lambda) more flat for StvKutta

5. Looking on the table above results are very unclear.

But amount of tracks 30% still less in StvKutta