- heppel's home page
- Posts
- 2021
- July (3)
- 2020
- February (1)
- 2019
- 2018
- 2017
- 2016
- December (2)
- November (2)
- October (3)
- September (2)
- August (1)
- July (3)
- June (5)
- May (8)
- April (4)
- March (1)
- February (2)
- January (2)
- 2015
- December (1)
- November (4)
- October (8)
- September (4)
- August (3)
- July (2)
- June (7)
- May (8)
- April (5)
- March (13)
- February (5)
- January (2)
- 2014
- December (1)
- November (2)
- September (1)
- June (3)
- May (2)
- April (1)
- March (3)
- February (2)
- January (1)
- 2013
- 2012
- 2011
- December (2)
- November (1)
- September (2)
- August (3)
- July (2)
- June (6)
- May (2)
- April (2)
- March (3)
- February (3)
- January (3)
- 2010
- December (1)
- November (2)
- September (2)
- August (1)
- July (4)
- June (3)
- May (2)
- April (1)
- March (1)
- February (2)
- January (1)
- 2009
- 2008
- My blog
- Post new blog entry
- All blogs
Distribution_of_TimeDependent_Code
As I indicated in
http://drupal.star.bnl.gov/STAR/blog/heppel/2011/mar/28/new-code-led-time-dependence
I have created new classes to characterize the time dependence of FMS gains based on LED puser data.
This note includes instructions and a new psu software release that allows the use of these classes in the batch analysis of run 11 data.
The psu batch analysis code was described in a Feb 11 posting,
http://www.star.bnl.gov/protected/spin/heppelmann/tmp/Steps_analysis.pdf
The package has morphed a bit over time but the big change is the addition of time dependend gain corrections.
Getting the new package
The new version of the package is linked here.
It is a tar "*.tar.gz" and builds the directory root12fms
Note: remember to delete the old complied lib. (root12fms/FpdRoot/Fpdchan.so)
and do a "make" to rebuild it!
tar -zxf root12fms_WithRunDep.tgz
cd root12fms
cd FpdRoot
rm *.so
make
cd ../
Review of Old batch sequence
To run a condor job to analyse several runs of data, the recommend sequence was to run the MkRun script.
to make a file physics plots file called Jchk12xxx.pdf.
Run Sequence for using new time dependent correction code
This assumes that you have experience running the above scripts that submit condor jobs. This section highlites that which now is changed in the sequence.
1) Check the environment variables in root12fms/SetFMSEnv to be sure that you have $FMSCORR set correctly, with a current set of gain corrections from root12fms/fmstxt/. A new environment variable is defined in SetFMSEnv. This variable
setenv RunDepPath "$FMSTXT/RunDep.root"
2) There are lots of trigger files on the psu disk in either ~heppel/psu/heppel/Run11/Day... or in ~heppel/psu/tobiv/Run12/Day...
Edit the root12fms/MkRun to point to 1 or more runs in one of these trigger source file directories.
2) Also edit the script root12fms/MkLed to point to the files and source directories to be analysed.
The only thing the MkLed is used for is to update the file root12fms/fmstxt/RunDep.root so the LED adc information is added to the file for the given run. This script does not use condor but runs on the local machine. It runs a short job on each segment file for a run under study and makes an entry in the RunDep.root file. If you want to look at the RunDep.root file in root, be sure to run an initialization script like root12fms/start.C to load the libraries.
3) Run the MkLed script to add entries to RunDep.root.
You don't need to run MkLed if you are running data that is already represented in the root12fms/fmstxt/RunDep.root file.
After you finish this, a root12fms/files120... directory tree is created for each run. You should delete these directories.
4) After the root12fms/fmstxt/RunDep.root file is ready, you can submit the condor jobs as before.
Run the MkRun script
The resulting file Output.root can be processed to produce physics plots with something like:
root -b -q 'CeckJet.C("12xxx","Output.root")'
to make a file called Jchk12xxx.pdf.
There is a lot of new code here and expect bugs.
Steve Heppelmann
- heppel's blog
- Login or register to post comments