sigh, another email !
I strongly suggest that each shift reports the measured
asymmetry, the target "rate" and ligit pressure (optionally target
temperature, deadtimes) using the rate program
this will gnuplot the target elastic rate as seen from events selected
with the tofs only as a function of run number (output goes to rate.ps,
and ligit.ps, see attachements).
usage is :
rate [n [m] ] _ use the n last charge files
optionally: use only up to file n-m
The target asymmetry is another thing that should be reported at every
shift. It takes only one command!
root show_ep_asym.C ### ###-###
of course, you will only be able to use runs that finished crunching and
have been filter (and of course the charge...). It does not take an
expert to understand this macro. Using filtered ntuples makes this macro
all that much more faster.
Bottom line is that analysis is the responsability of people on shift....
---------------
Finally it is very important that we pay attention to the runlist. To
check if a file has been finished crunching use
root lr.C ###
lr->Print() will give you "all the events". A non-closed file is usually
made a zombie. A file under-process may not be updated (modification
time) even for a period of a few hours... so "ls -l" gives you NO
information. Of course the lr-## file will appears to be on disk right
after you type lrn ##... Of course we expect to simplify this process
with dst and automated crunch but we are not there yet.
Recently, we have found repeatedly the following problems with the run
list:
_ runs labelled as finished were not finished
_ runs that crashed or ended abnormally we analyzed just as a good run
_ test runs, empty runs, cosmic runs, all sorts of junk runs have been
analyzed just to see if we can screw up the list of good runs and charge
files.
- 3 or more jobs were sent to the same spud
_ the runlist was simply not updated at all or very consistently.
note: type "freecpu" to get a list of crunching jobs, and where
type "manyterm" to get a number of nicely order xterms to start jobs
and please avoid crunching from emacs, java or from other operating systems....
________________________________________________________________________________
Tancredi Botto, phone: +1-617-253-9204 mobile: +1-978-490-4124 research
scientist MIT/Bates, 21 Manning Av Middleton MA, 01949
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
This archive was generated by hypermail 2.1.2 : Mon Feb 24 2014 - 14:07:30 EST