Hi,
first of all: we implemented a reasonable I/O mechanism to write and
read datasets from root files. VVF now runs on the complete Run1-4
statistics in about 8 (wall clock) minutes on my 1.8 GHz Athlon and
requires ~650MB of memory. We also added pull plots to the mES plots,
see
http://www.slac.stanford.edu/~bozzi/fitr14mES_Aeff.eps
Roberto is investigating the possibility to fit mES/sqrt(s), some
figures are here:
http://www.slac.stanford.edu/~bozzi/mesfitR1.eps (mES/sqrt(s) fit)
http://www.slac.stanford.edu/~bozzi/mesfitoldR1-rebin.eps (mES fit)
The same fits with 71 bins instead of 40 are at
http://www.slac.stanford.edu/~bozzi/mesfitR1-rebin.eps
http://www.slac.stanford.edu/~bozzi/mesfitoldR1-rebin.eps
We see a dependence of the chisq as a function of the number of bins...
to be further investigated.
Concezio.
|