pp-pileup in MDC4

1
pp-pileup in MDC4

• Goals
• MDC4 statistics
• vertex finder performance (ppLMV)
• leading particle recon efficiency

• Jan Balewski, IUCF
• Star Analysis Meeting
• May 13, 2001 , BNL

2
MDC4 overview

• Small problems of MDC4
• crash after 400 eve ? fixed
• SVT crash on empty eve ? fixed
• root crash after 15,000 eve ? fixed
• root version RCF vs. PDSF ? recompiled
• NFS problem corrupted DST ? rerun

• 100,000 pp events processed
• PYTHI 6.1, minbias, nondiffractive
• GSTAR with y2001 (actual) geometry
• 8 pileup events added
• root4star --gt ppDST
• spin sorted ?-distributions for rLP

• Resources MDC41 hour of STAR
• CPU 50 sec/event
• ? total 1400 hours CPU
• ? 1000 jobs with 100 events
• ? 2 days on 100 CPU (real time)

Thanks to Yuri, Lidia, Helen and all other for
help
3
MDC4 pp event topology in 2001
L41030 cm-2 s-1, ?NON DIF 28 mb Collision
rate 120 KHz 60 bunches/ring ? 213 nsec 1
coll/50 bunches 1 coll/10 ?sec 40 ?sec TPC
drift ? 8 coll/ TPC frame pileup ? few 100 tracks
in TPC diamond total width of 15 cm
PYTHIA minbias trigger event, generated at ?s200
GeV
Leading Particle
Track color encodes its transverse momentum
(GeV/c)
Y
104
Total 105 events ( 1 hour of STAR)
X
minbias trigger
12 primary tracks, 30 TPC tracks
Leading Particle pT (GeV/c)
4
MDC4 Vertex finder (ppLMV)
Only TPC tracks used
with 8 pileups /event (without pileup in green)
RMS6.7 cm generated ?7.5 cm
Z (cm)
RMS4.7 mm (3.2 mm) efficiency96
(90) signal/noise 5/2 (68/1)
RMS4.3 mm (3.2 mm)
Vertex ?Z (cm)
Vertex ?X (cm)
5
MDC4 Efficiency of ppLMV vs. LP pT
with 8 pileups /event
Only TPC tracks used
Leading Particle pT lt1 GeV/c RMS5.1
mm signal/noise 5/2
pT 2,3 GeV/c RMS3.3 mm signal/noise 9/1
pT 1,2 GeV/c RMS3.9 mm signal/noise 6/1
Vertex ?Z (cm)
Vertex ?Z (cm)
6
MDC4 Reconstructed LP vs. ?
With 8 pileups per event (no SVT)
No pileup (no SVT)
GeV/c
GeV/c
problem
GeV/c
GeV/c
? of recon LP (deg)
? of recon LP (deg)