Hits on global tracks, from processing the first 30 events of one AuAu200 file in DEV on Nov. 4, 2010: 

root4star -b -q -l
'bfc.C(30,"P2010a,btof,BEmcChkStat,pmdReco,Corr4,OSpaceZ2,OGridLeak3D","/star/rcf/test/daq/2010/029/st_physics_11029020_raw_1030002.daq")'

The only cut is selecting global tracks with flag > 0, then looping over hits on those tracks.

Counts versus sector:

 Counts versus sector and z [cm]:

Counts versus y and x for z < -50 (limiting only to east sectors), where sector 20 is in the "4 o'clock" position, with a dead anode region in the middle:

*** After modifying StTpcHitMaker to exclude hits on "dead" padrows.... ***

Counts versus y and x for z < -50 (limiting only to east sectors), where sector 20 is now completely empty:

Ratio of after fix to before fix showing that there is a reduction of hits on tracks for hits close to sector 20:

Ratio of after fix to before fix, again showing reduction of hits on tracks which are in the neighborhood of sector 20 (sector 4 is on the west half of the TPC at the same azimuth, so tracks near the central membrane which use hits in sectors 3,4,5 may use hits from sector 20 also; note that there appears to be a general increase in usage of hits from these opposite side sectors):

Reconstruction time impact:

I also examined the amount of time spent in the TPC hit-making code and found that the new checks on live/dead padrows had no noticeable impact. In both cases, the "tpcChain" took 13.8% of the total bfc chain, while the hit maker ("tpc_hits") took ~37% of the tpcChain.

-Gene