Notes on building & running UCL width programs ============================================== *** SIMULATION WITHIN CDF SOFTWARE *** [10 Aug 2005] Building : ---------- [Example area : nucl05:/cdf/data1/dwaters/w_width] (1) Check out UCL pacakge from the head. (2) > gmake UCL.nobin (3) Modify TBIN target in "UCL/main/standalone/GNUmakefile" to select required binaries. (4) > gmake UCL.tbin (5) > ./bin/Linux2-KCC_4_0/ucl_simulation < UCL/cards/ucl_simulation.cards In the cards file the templates that you wish to generate are specified. These are tied to the process - e.g. cannot generate GW templates when running Z's. The output file containing the templates is named : MCTemplates_.root with a separate subdirectory for each template type. *** RUNNING OVER DATA *** [11 Aug 2005] (X4) > ./bin/Linux2-KCC_4_0/root_analysis *** STAND-ALONE FITTER *** *** WILL REQUIRE ROOT COMPILED WITH SAME g++ VERSION ETC. *** Building : ---------- [Example area : pc71:/cdf/data1/dwaters/w_width] (1) Check out UCL pacakge from the head. Copy the UCL directory structure into a fresh directory (not a CDF test release). (2) Specify which binaries to build by listing the main build files at the top of "UCL/scripts/build_nonCDF.sh" (3) Copy the following CERN libraries : libkernlib.a libmathlib.a libpacklib.a libpdflib.a into the directory "lib/standalone". (4) Do the following *without* setting up the CDF software : > ./UCL/scripts/build_nonCDF.sh Running : --------- Modify as necessary the file : UCL/card/ucl_fitter.cards Things to look out for : (1) The histo-name for the templates (last field in the line TEMPLATE_NAME_LOCATION) is the stem of each histogram name. (2) The name specified in the line TEMPLATE_VARS is the name of the histogram in the same directory as the templates, that defines the parameter value(s) corresponding to each template. It should have a name like "GAMMA_W" for width-fitting. The run : > ./bin/standalone/ucl_fitter < UCL/cards/ucl_fitter.cards