P G S

By the way, "PGS" is pronounced pee-gee-ess, not "pigs"...  ;)

You also need a recent version of CERNLIB , including PDFLIB.  Note: you may need to add the qualifier " -lsnl " to the compile/link script, mkpgs.

Make sure you hit the reload button on the code page before saving - Netscape caches older versions. You should get version 3.3.

All user-accessible data is stored in common blocks in the include file pgs.inc. The main program directly generates or reads in events, then performs a simple trigger simulation, generating a list of trigger objects. The program then makes a list of reconstructed photons, electrons, muons, hadronic tau decays, and jets (including b-tag information) according to the detector parameters given.

The reconstruction efficiency for physics objects (photons, electrons, muons, taus, jets) depends greatly on the environment. Thus, the best way to perform a simple simulation is to generate a list of reconstructed tracks, and generate a list of calorimeter tower energies, based on parametrizations and measurements from the actual detector and full detector simulations. Then one can perform object reconstruction much as is done in the real data analysis.

This approach works well for photons, electrons, taus, and jets. For muons, one can simply parametrize the efficiency, including fiducial coverage. For b-tagging, one can also parametrize the efficiency for jets in some ficudial region. Thus for muons and b-tags one can start with generated objects (mu or b), match to a track or jet, and then apply the efficiency.

The simulation presently does NOT simulate the following effects:

> setenv PGS    <your pgs location>
> setenv PYTHIA <your pythia location>
> setenv ISAJET <your isajet location>
> setenv STDHEP <your stdhep location>
> setenv CRNLIB <your cernlib location>

You should take one of the main programs in the examples and modify the subroutine pgs_user_analysis to perform your own analysis. Follow the template, which allows initialization, event processing, and wrap-up at the end. To compile, use the example script. For pgs_taus.f, for example, issue the command

> $PGS/scripts/mkpgs_linux taus

which will create an executable called taus.

Event Storage and I/O

To save generator files on disk, you simply need to set the value of pgs_output_file to your file name, and then, in pgs_user_analysis, call pgs_write_event('event'). Note, however, that this does NOT save the reconstructed objects, etc. This ONLY saves the generator information. So, upon re-reading the event and re-simulating, you will get different results! Also, the resulting generator files are quite large, about 40 kbyte/event.

To avoid big files, though, PGS can run PYTHIA/ISAJET directly. This is determined by what you put in your main program for the variable "optpgs"...see the examples.

Submitting code to PGS

Write your code, test it if possible, and send it to me for inclusion in pgs_library.f. Please send it as a plain text file, by email. Sending me the entire pgs_library.f file, or a tar file, just means a lot of extra work. Keep it simple, use "implicit none" and reference the pgs.inc commons only. Let's avoid having to read external files and the like. If you perceive a need for changing pgs.inc, let's discuss it - it isn't written in stone!

For photons, electrons, taus, and jets, we can use the calorimeter clusters and track list to actually do the ID as if it were the real experiment. If we have done the calorimeter job properly, the efficiencies should be realistic. For muons, we can apply simple efficiency numbers. For jets, we can use the calorimeter clusters as a start, and then apply efficiency numbers for b-tagging.

Version History