April 3, 1998

A QUICK TUTORIAL ON RUNNING THE QCDWZ NTUPLE MAKER

The QCDWZ ntuple maker resides on the alpha cluster. So any work should be done on those machines. For information on the tags in the ntuple please consult D0note 3124. This guide is by no means a complete description of the ntuple maker but intended only to allow the user to make his/her own ntuples.

Getting Started:

In your OWN disk area create the following subdirectories [.NT.RCP] and [.NT.OUT]. Copy the command file RUN_JOB.COM from TMP$ROOT355:[ERIC.NT] to your directory [.NT]. The command file RUN_JOB.COM is all you will need in order to make your own ntuples. The rcp files that you will have to modify to create your own ntuples are:

• QCDWZ_NTUPLE_V11.RCP -- this controls the various blocks of data which will be written to the ntuple
• QCDWZ_L2FILT_CUT_1B.RCP -- this controls the triggers that are allowed into the ntuple
• QCDWZ_TRIGGERS_1B.RCP -- this is simply a list of triggers which allows the user to identify which triggers have fired when using your ntuple

Copy above three rcp files from PRJ$ROOT313:[QCD_4.QCD_WZ.MAKE_NT.V11.RCP] to your directory [.RCP]. You may then modify QCDWZ_NTUPLE_V11.RCP and QCDWZ_L2FILT_CUT_1B.RCP to suit your needs. The rcp files are reasonably well documented.

Modifying RUN_JOB.COM:

In the command file RUN_JOB.COM you will have to define two logicals. They are called MYRCP and NTOUT. The MYRCP logical should point to [.NT.RCP] in your area. The NTOUT logical should point to [.NT.OUT] in your area. You may also wish to change the filename of the ntuple from OUTPUT_NTUPLE_NAME.NT to something else. The final thing to change is the name of the input data file. In the sample command file the following line appears:
"$ define FILE d0$data$d0dad$trigfilt:run1b_filt_em1_eistrkcc_ms.d0dadf".
This input file is the Run 1b microdst's for the em1_eistrkcc_ms trigger. The user has to redefine this to the trigger he/she wants for their data analysis. To find out what is available the user should do a directory of d0$data$d0dad$trigfilt:.

The previous example only used the microdst's in the d0dad format. If the user wishes to run over dst's or individual files he/she should make a file containing a list of the files to be run over, e.g. MY_LIST_OF_EVENTS.LIST. In the command file RUN_JOB.COM he/she should change:
"$ define FILE d0$data$d0dad$trigfilt:run1b_filt_em1_eistrkcc_ms.d0dadf" to
"$ define FILE_NAMES MY_LIST_OF_EVENTS.LIST" and change
"FILE !- <input file name>" to "FILE_NAMES !-<input file name>.

Once this is done thein user should test to see everything is working by running over a limited number of events. After that you should submit the job in batch mode e.g. SUBMIT/CHARA=(NEW_PRJ)/NOTIFY/NOPRINT RUN_JOB.COM. Good luck and may the force be with you...always.

To send mail to Eric, click here