PMDWide nanoDst

1
PMD-Wide nano-Dst Charge Team Plan Implementation
Status To-do
2
Charge framed in last collaboration meeting
(Aug 15th, 2005)
A pmd-wide nano-dst format starting from STAR
Mu-dst has to be designed and implemented.This
framework will then be used to generate nano-dst
from all available Nano-dst files. Features to
be implemented are Star Global Event
information Fully calibrated PMD Clusters (info
as in mudst) Hits associated to the cluster.
Discrimination will not be part of this
exercise Working version of the job MUST be
completed by next collaboration meeting and
large scale production will start based on the
feedback available.
3
TEAM
Subhasis will co-ordinate the effort and will
take help from Interested persons.
A mail was send to pmd mailing list for
expression of interest, Many (mostly from
students) responded and then some of the
Supervisors wanted their names to be included so
that they can monitor the progress of their
work.
List included YPV, Tapan, Bedanga, Zubayer,
Pawan, MMA, Monika, Navneet, Rashmi, Sudhir,
Sunil, Neeraj, Surjeet, Baba, Anju, Sanjeev,
Badyal, Nandi, Prabhat
4
PLAN
Detailed Evolved, Step by step approach..

• Get Calibration Constants and fill into the DB
• Make proper category of runs
• Develop code for finding isolated cells from
• Mu-dst file.
• (iii) Work on DataBase table for PMD Calibrations
• (iv) Store them in DB with proper Time-Stamp

5
3 Teams formed to look at different
Dataset Monika/Navneet/Prabhat 200 GeV
CuCu Sunil/Satish/Surjeet 200 GeV
AuAu Neeraj/Za 62.4 GeV CuCu
They Made categories, will be discussed by them
It went through some level of discussions in
mailing list
Based on Rashmis code for Calibration from
Mudst, Pawan/Bedanga made a first version of
code for isolated cell search by Sept 1st.
Main Features It is a macro Reads Mudst
Cleans Data for ADC spectra Does Isolated
Cell search WriteOut tree for isolated cell
RunNo, evt No, Cell Row, Col, ADC Lots
of histgram for various levels.
6
Subsequent developments and formation of Beta
version Mapping Utilities were initiated for
every Run_no. Small fixes in the code.
Date,Time, Chain_no, Channel_No included in
tree. 3 teams started generating respective
iso.root files.
One Macro developed based on hadd.C where ONLY
hists are added, not tree. These added hists
are inspected before going to next step.
Reading Macro developed for reading iso.root
files. (Pawan/Monika) Added features Fit
(Landau) for each isolated cell. Ntuple made
with sm, row, col, chain, channel, date,
time, mean, mean1,
Content, par3, error3, chisq
7
Ntuple files read, Based on selection criteria,
Gains found for each cell with following params
Denominator for a cell comes from mean of
mean for a SM-Chain combinations. After
some discussion we decided to take MEAN
They are then push into STAR DB with proper time
stamp.
In summary, Pmd_beta_prod.C -? Write tree
for isolated cell after cleanup hadd_pmd.C -gt
adding hists for iso.root files. iso_read.C -gt
read ntuples from iso.root files, fits and write
ntuple for isolated
cells with fit info. load_pmdCalib.C ? Loads
gain value into DB
8
PMD Calibration DataBase table(s)
New Table
Old table
9
Tables being used NOW
Gain Table
Summary Table
10
Step II towards Nano-Dst
Design Read MuDst.root files Read Pmd Gain
from DB and apply on PMD hits

• PmdClustering on calibrated PMD Hits

• Design PmdNanoEvent and PmdNanoCluster

• Write them in tree and in a file (Event.root)

11
Some Technical Issues We are reading Mudst
info, mpdifying them (e.g. gain application) ?
But not rewriting, so modified info must be kept
in the chain and written to PMD nano Event. 3
Makers to be in a Chain.. StMuDstMaker
StDBMaker StMuAnalysisMaker
StPmdClusterMaker StMu2Maker
12
StMuAnalysisMaker Reads PMD hits from
Mudst.root, reads gain from DB (timestamped)
apply Gain modify ADC in StMuPmdHit
13
StPmdClusterMaker Reads MuDst.root Perform
Clustering Writeout StMuPmdCluster
Several Issues and problems sorted out
Clustering should work for calibrated hits CVS
StPmdCluster works only for StEvent Needs to have
option for Phataks/ TKNs clustering included
Rashmis new clusterMaker
14
Design of PMDNanoEvent and PMDNanoCluster
A mail went out to PMD mailing list for
suggestions, several options (wish list)
discussed
PMD Cluster SM, Eta Phi ADC Sigma Ncell X,
Y Chain, channel Isocell Hits to Cluster
PMD Hits Sm, row, col ADC
Global info RunNo, Event No, TrigId
Vertex Date Time RefMult ZDC CTB RP

Tracks px,py,pz DCA NFits
15
Taking root tutorial example to start with for
writing Event, implementation has been made and
Event.root file written All these done during
october 15th to nov 15th (during my stay at
BNL)
It was proposed that students working on the
project should come to VECC so that detailed
debugging and production can start. Monika/Sunil
/Neeraj/Surjeet arrived VECC around last week of
Nov.
16
Several issues resolved (mainly debugging)
year flag problem with MapUtil Memory Leak
problem det0/1 convention problem
PMDEvent.root files written by all 3 groups
using scheduler (Max 20(?) mudst.root files,
files transferred to GRID machine, Reading Macro
developed (Monika) for those PmdNanoevent.root
files Individual analysis codes being
developed Let the show begin!!!
17
Acceptance Correction
It was part of debugging, but can be used for
analysis.. Cells with NONZERO gains are read
from DB Eta/Phi Calculated Eta/Phi
distribution obtained Looped over all possible
cells Eta/Phi distribution obtained Ratio
gives acceptance Used heavily to look at
SM-wise Eta distribution.
18
Known problems ClusterMaker for StEvent to be
tested Hits to be attached to the clusters
TArrays to be read out need to look at memory
leak for further tuning.
19
Status and Plan Once Clearence obtained,
production will start.. Proposal First team
stays up to 15th Jan to have one dN/dEta plots
for respective dataset. New team (Satish/
Prabhat/Navneet) has arrived and plan to train
them, so that they start production. Vikas
should find space in GRID machine (buy new
disks) to accommodate them. Everyone should
start reading the files from respective places.