Making W/Z Production Grids CMS

This document is a a brief explanation of how to create grids of production cross section as a function of transverse momentum (pt) and rapidity (y) for input to the CMS Monte Carlo used by various W/Z analyses at DØ.

The grids needed from CMS are obtained by combining grids made with Legacy using ResBos. Legacy itself produces several grids, each corresponding to a part of the total calculation of the cross section. ResBos adds the grids together properly, in particular, it takes care of matching the resummed+perturbative calculation to the perturbative only calculation. (Briefly, the resummed+perturbative calculation fails if pt gets large enough, around the value of Q. In general, one could just use a cut-off, switching to the perturbative only calculation at some value of pt. Doing so results in a very "unsmooth" grid in pt. The matching criteria used in ResBos is a function of both pt and y, effectively smearing out the switch over a range of pt values, and smoothing out the transition.)

In general, ResBos calculates the dilepton cross section, i.e., it decays the vector boson. Since we want the production grid for the vector boson itself, we must use the option to output the grid before doing the complete calculation.

The basic proceedure for getting a grid for CMS is:

• Obtain or generate grids from Legacy
  I recommend that you use grids generated by the experts if at all possible. There a loads of Legacy grids on the ResBos page. Additionally, Csaba and company are generally happy to make grids for you if there aren't many to make.
• Run ResBos with iMG = 1. This will tell ResBos to output the the grid to the file Matched.out
• Run the CMSHistogramMaker using the file from ResBos (i.e., Matched.out) to create a 2d histogram for use in CMS and a 1d pt plot for general consumption

Grids made for the W mass analysis

I made several grids for the W mass analysis completed in 1999. They all have 32 bins in y, 100 bins in pt, and go from 0-50 in pt.

• wp_cteq5m_mw_matched.hst
• wp_cteq5m_mw_g2-_matched.hst (g2 decreased by 0.06 to 0.52)
• wp_cteq5m_mw_g2+_matched.hst (g2 increased by 0.06 to 0.64)
• wp_mrst_mw_matched.hst
• wp_mrsr2_mw_matched.hst
• z0_cteq5m_mw_matched.hst
• z0_mrst_mw_matched.hst
• z0_mrsr2_mw_matched.hst

GridUtils

I wrote a couple of programs to facilitate creating a 2d histogram appropriate for CMS from the text table made by ResBos. I put the program in a tar file called gridutils.tar.gz . You will need to expand the archive and compile the program to run it. There is a simple Makefile provided, so you should just have to type "make" to build the executable. You should look over the Makefile and be sure that everything works on your system. Pay attention to the directory containing the cernlib product packlib. You need it to make the histograms.

The make_cms_hist program assumes that the output from ResBos has the cross section calculated at the exact mass value you want and all of the y values. It then interpolates in the pt direction. If you don't have the exact y values you want actually calculated by Legacy (ResBos just combines them), then you will have to add an interpolation in y as well.

Once you run make_cms_hist, you will have a histogram file (called histfile) containing all the 2d histograms and 1d histograms available from the ResBos output. You want to use the histograms corresponding to the CSS+Y Matched calculation -- 2dhist 14 or 1dhist 114. If you are making W grids, there will be both W+ and W- grids generated, where I assume that ResBos is outputing W+ grids.

If there are any questions, please feel free to email me at casey@pa.msu.edu

Below are some links to relevant documentation about Legacy and ResBos.

Papers

Online references:

Legacy: Pavel Nadolsky's presentation on the C++ version of Legacy

ResBos: ResBos homepage

Code

Legacy

ResBos (on homepage)