In the following we will generate manually a 'single kaon' event, which can be fed into OSCAR and ORCA.

The HEPEVT ntuple structure

OSCAR expects its input in an HBOOK ntuple with the following fields:
   integer NEVHEP = event number
   integer NHEP   = number of entries (particles, partons)
   integer ISTHEP(NHEP) = status code
   integer IDHEP(NHEP)  = PDG identifier
   integer JMOHEP(2,NHEP) = position of 1st and 2nd  mother
   integer JDAHEP(2,NHEP) = position of 1st and last daughter
   real PHEP(5,NHEP)   = 4-momentum and mass            (single precision in ntuple file)
   real VHEP(4,NHEP)   = vertex xyz and production time (single precision in ntuple file), OSCAR do not care about this...

   integer IRNMCP       = run number
   integer IEVMCP         = event number (must equal to NEVHEP)
   real WGTMCP         = event weight (1.0)
   real XSECN          = cross section equivalent (1.0)
   integer    IFILTER        = filter pattern (0)
   integer    NVRMCP         = number of additional variables
   real VARMCP(NMXMCP) = list of additional variables

Packed ntuple trick:

jmohep(i)=mod(isthep(i),100)*4000*4000 + jmohep(1,i)*4000+jmohep(2,i) jdahep(i)=mod(isthep(j)/100,100)*4000*4000+jdahep(1,i)*4000+jdahep(2,i)

Writing ntuples

Writing such an ntuple is straightforward in Fortran [TODO: C/C++ interface via cfortran.h], see the attached example program writehbook.f.
  1. Define the data blocks as Fortran COMMON's.
  2. Create an ntuple with the HROPEN procedure (say test.ntpl), and define its structure with HBNAME. Note: ntuple's ID must be 100 (unpacked) or 101 (packed; a bit tricky). Additonal note: Ntuple variable names have to be in capital letters. Don't ask why...
  3. Do a loop over the events, generate some particles, etc; at the end of the loop call hfnt(ID)
  4. Close the ntuple.

Compile the program with g77 -c writehbook.f and link with g77 -o writehbook.o `cernlib packlib,mathlib` This requires a running g77 system with cernlib properly installed (on debian: apt-get install cernlib)

Run the executable.

Now the file test.ntpl should contain the particles. You can test it by opening with PAW:

h/file 0 test.ntpl 4096
n/print 101
n/plot 101.idhep

TODO: this example program does not call Pythia to decay instable particles, but who cares anyway? We want now to control everything manually.

Modify (extend) an existing ntuple

In the previous section, we created a brand new ntuple containing our particles (a more common example is when the ntuple is generated by some fancy CMKIN routine). We can further tailor it, add or remove particles, change their properties etc. With an appropriate Fortran program (see extendhbook.f) read the original ntuple, do your modifications, and write out to an other ntuple.

TODO: there has to be a more elegant solution than just write out to a second ntuple. Maybe hropen(... 'U' ...) ?

Feed into OSCAR

This note assumes that you have a properly installed OSCAR environment.

Copy or symlink test.ntpl into the OSCAR working directory (in my case, it's OSCAR_3_9_9/src/Workspace).

Edit oscarrc:

OutputDataSet = /System/sim/test
EventNtplReader:NtplFileName = test.ntpl
EventNtplReader:NtplID = 101 # or 100 if you generated unpacked ntuple

Do the ritual eval `scram run -sh` (or -csh), and run oscar -c oscarrc

This will produce a lot of sim and xml files (see ProduceSimulatedEvents)

Further readings


HBOOK Reference

-- GergelyPatay - 30 May 2006

Topic attachments
I Attachment History Action Size Date Who Comment
Ff extendhbook.f r1 manage 3.7 K 2006-05-30 - 23:17 UnknownUser  
Ff writehbook.f r1 manage 2.9 K 2006-05-25 - 09:46 UnknownUser  
Edit | Attach | Watch | Print version | History: r7 < r6 < r5 < r4 < r3 | Backlinks | Raw View | Raw edit | More topic actions
Topic revision: r7 - 2007-11-11 - FerencSikler
This site is powered by the TWiki collaboration platform Powered by Perl This site is powered by the TWiki collaboration platformCopyright &© by the contributing authors. All material on this collaboration platform is the property of the contributing authors.
Ideas, requests, problems regarding TWiki? Send feedback