Short writeup for pBin
Usage:
pBin -i<filename> -R<topology> [-n<numevents>] [-s<scalefactor>] logfile1 [logfile2] [etc]
Description
pBin is a perl script written by Dennis Weygand which serves as a front end
for two programs: predict and SelectWeight. These programs are currently
available only on the SGI platform (although they could probably be compiled
for the RS6000s without difficulty). predict reads amplitudes from a PWA
logfile and then runs through an E852 itape file (which must contain
GROUP_ESR_NPARTICLE) and assigns a weighting to each event based on the
aforementioned amplitudes. The output of predict is an itape file (the number
of events being selected by the user) whose events are identical to the input
events, with the sole addition of the weight described above. SelectWeight then
runs through the output of predict and rejects events based on comparison of
their assigned weight with a pseudorandom number. The output of SelectWeight
is then a set of events which should reflect the "prediction" of the PWA fit.
The interpretation of these events depends on the type of events input. For
example, if one used raw Monte Carlo as input, the output would reflect
acceptance-corrected data. If one used accepted Monte Carlo as input, the
output would reflect uncorrected data. The purpose of pBin is to automate
running of the two programs, and to allow one to loop through a series
of logfiles with a single command. pBin will output one itape file per logfile
input, with the names being the same as the respective logfiles with
the addition of ".dat" at the end.
Options
-i<filename>
Name of E852 itape file to be used as input.
-R<topology>
String representing the event topology (e.g. etapi-).
-n<numevents>
optional -- Maximum number of events desired in the
final output (i.e. the output of SelectWeight). If this option is not used, pBin will attempt
to select the proper number of events as predicted by the fit (i.e. the amplitudes from the
logfile).
-s<scalefactor>
optional -- This is a multiplicative scale factor
used to determine how many times more events to request from predict than the final number
desired from SelectWeight. A larger number gives SelectWeight a larger pool of events to
choose from, increasing the chance of achieving the desired number of output events. predict is
fairly slow, however, so making this number too large can be cumbersome. Default is 5.