Linear Collider Detector Project |
|||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
|
|
LCD DatasetsOverviewThe physics process simulation stage ("event generation"):Events are created with a physics event generator in files ("runs") containing 100 events. This number was chosen to be relatively small to put a reasonable limit on the cpu time needed for the detector simulation of a run. A variety of generators are available in various states of usefulness depending on the physics process of interest. For the standard "signal" processes (ee -> ZZ, WW, ff, HZ, etc) we initially used Pythia augmented with some beam and radiative corrections. Events created this way contain the phrase "pybms" in their full run names. We are moving to replace Pythia with Michael Peskin’s new Pandora generator. The run names for Pandora events contain the phrase "panpy", where the "py" indicates that the hadron fragmentation and particle decay is done by Pythia. Other generators are available for two photon processes and Daniel Shulte’s Guinea Pig is used to simulate beamsthralung backgrounds. The detector simulation stage:It is possible to feed the event generation stage output directly into an analysis program which will perform a "fastMC" detector simulation prior to performing analysis. However, to provide for detailed detector simulation we currently use the Gismo simulation program. (We will probably begin moving over to Geant4 sometime in 2000.) Gismo takes a detector description and an event generator run of events and produces 100 event output runs containing detector simulation hits and the original generator input. Naming conventions:At present these detector simulation 100 event run outputs are given descriptive names. An example of a name would be
The "pybms" as noted earlier is the event generator. In this case "WW" is the process. The "left" means that the electron beam was left hand polarized. Since the WW events are asymmetrically distributed with respect to the beam axis the "ctr" indicates that only events which fall into a central region of the detector have been retained. The "500" is the CM energy in GeV of the beams. "990115" is a date, yy-mm-dd, indicating the vintage of the detector description as well as the state of the generator, simulation program, etc. The date is roughly a release indicator. The "Small" refers to the type of detector. At present there is a Small and a Large detector description. In the future the detector name will probably contain release information as well. The next major release of the detector descriptions will include additional features and other changes and will probably be designated "S2" and "L2" for version 2 of Small and Large. The "sim" means this is a full detector simulation run and the "77" is a run number. Another example of an event sample name is
Here there is no reference required to beam polarization state since it was mixed and since the events tend to be symmetrically distributed in the detector no acceptance cuts were made. The output of the original physics event generation stage is similar to the detector simulation stage. The "Small-sim" will be replace with "gen". Thus, a sample generator output name would be
Format of output:The generator output is put in HepEvt format using the Fermilab StdHep program and the files are given the filetype ".dat". The output of the detector simulation stage is an ascii file given the filetype ".dat". An additional stage of processing uses the .dat files to produce files in a format ready for use by the JAS analysis program. The output of this stage is given the filetype ".lcd". Both these filetypes are readable on Unix, Linux, DEC and NT platforms. Major data production periods:In preparation for the Sitges meeting in late spring of 1999 a great deal of data was produced with the 990115 designation. Following Sitges it was decided that there would be a major revision of the detectors to the S2 and L2 mentioned above. However, there were a number of problems with the 990115 release that could be fixed on a shorter time scale and it was decided to rerun a 10K sample of WW, ZZ, tt and udscb events for the amended 990115 (S1 and L1 in the new parlance) version. These events have been produced and are referred to as the October 1999 dataset although they still contain the 990115 designation in their names. Dataset CatalogsThe tables contain the names of datasets for each major release along with their run range. The "Specs" column contains pointers to the actual datafiles used to drive the physics event generation ("gen") and detector simulation ("det") programs. The last columns (SLAC, UPenn, UMich, FNAL) indicate availability on the respective JAS dataservers. Pre-Sitges Production Run:
Pre-Sitges Notes:
October 1999 Production Run:
October 1999 Notes: 1 - Created on SLAC farm. Last modified: January 13, 2004 by Gary Bower, SLAC |
||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||