Minutes of CDF Detector Simulation Meeting (14 March 2002) ********************************************************** ------------------------------------------------------------------------- (1) Tutorial type introduction on how to run MC data production on the farm (Susana Cabrera) This talk was a tutorial type introduction on how to run MC data production on the farm. the best thing is look at her transparencies. She explained which tcl's you have to provide if you want people to run your MC job on the farm. She pointed to the following web-page for more information: http://cdfsga.fnal.gov/internal/mcProduction/ One thing you need to provide beforehand is the number of events that you will want. To check that a developed tcl's work you have to test the job yourself first anyway; this will then give you a good estimate of the the number of events to be generated on the farm. Sometimes when you have trouble getting somewhere it might be because the kahuna is down. You can check its status on http://rutpc7.fnal.gov/kahuna/ and if it is down send email to cdf_software_help@fnal.gov ------------------------------------------------------------------------- (2) Resolving infrastructure issues for smooth operations on the farm Example: Problems with Pythia crashes, upgrade to Pythia 6.2 (Stan Thompson/Elena Gerchtein) Up to version 4.4.0 Pythia 6.155 was used, from 4.5.0 on version 6.203 is used. Stan tells us that he put in new accessors and that this version within CDF software uses the CDF random number generator in pyr, not the internal Pythia one. Elena mentioned about problems with Pythia crashes: When generating events with Pythia and then running them through cdfSim and Production the job crashes at the end (which is rather unpleasant when running on the farms). The problem appears only for Pythia and none of the other generators (Isajet, Herwig, Fake event...) A way around this for the time being is to run Pythia together with cdfSim and then write the file out and read it in to production, so basically split the job into two parts. The step of going to Pythia 6.2 is rather important! It is the first Pythia version that is fully compliant with the Les Houches Accords. More about those: hep-ph/0109068. Basically during the Physics at TeV Colliders Workshop in Les Houches in 2001 people agreed on having a generic format for the transfer of parton level event configurations from matrix element generators and hadronization generators. Pythia 6.203 for instance can work as hadronization generator for events generated with CompHep. So going to a new version of Pythia was important also because it is needed for CompHep and Grappa. ------------------------------------------------------------------------- (4) Plans on how to import stand-alone generator output in HEPEVT format into the CDF system and into production on the farm (Sebastian Carron) This lead to the next talk about CompHep and how to import stand-alone generator output in HEPEVT format into the CDF system and into production on the farms. Info on CompHep is available at http://www-d0.fnal.gov/~dudko/mcdb/ How to install CompHep can be found at: http://hep.uchicago.edu/cdf/comphep.html There you can find information on how to run CompHep, too. CompHep can then be input for showering into Pythia 6.2 as was said before and the output is in accordance with the Les Houches Accords. The importance of such a generation is that the format is detector independent: CDF, theorists,D0 can all have a look and use it. So, theorists, D0, CDF would share common MC datasets (and theorists can give their ok to specific datasets). Elena build the interface between StDHEP and HEPG which can be found in generatorMods. What was also mentioned is that apparently CompHep does a good job of teaching you physics. It has latex plots of Feynman diagrams; so it is rather easy to see which processes you switch on and off and to analyse what happens in the different setups! One drawback so far is that it doesn't pass intermediate particles to showering. ------------------------------------------------------------------------- (5) Update on Les Houches Accords and survey of available PDF's and what CDF should be using (Joey Huston) Joey's slides are pretty self-explanatory and colourfull. So, best is to have a look there. ------------------------------------------------------------------------- (6) Status of Trigsim (Simona Rolli) Best is to look through her transparencies. The main things are at: http://ncdf70.fnal.gov:8001/trigsim/trgsim.html It explains a lot and if you want to run it always check TriggerMods/test/run_TRGSim++.tcl This is the one and only tcl file Simona is guaranteed to update ! TRGSim basically emulates the hardware trigger. ------------------------------------------------------------------------- (10) Plans from B group: Tuning of Pythia (Rick Field) Then we were running out of time and special request was to hear Rick Field first. Also, look at his nice slides. Main issues are that he wants to be able to combine MC runs using weighted histograms in pT^hat (luminosity weighted). His concern was whether the pT^hat information is stored so that indeed runs can be combined. At the moment though it is not saved in HEPEVT... work will go into making this happen. He emphasized again that we need to tune Pythia. Pythia changed, and we can't use the MIT tuning anymore to fit the underlying event. He said he would look into the tuning provided he is told which version of Pythia and which structure functions he should use. The third issue raised is that if we want to get all three sources of b-quarks, he will need some help in using filters. To generate some of the b-quark processes in Pythia you need to run the full QCD option and then filter out the events you are interested in. These are then passed to the detector simulation. Minutes kindly provided by Tatjana Unverhau.