Photon-induced simulations in Herwig

Updated 568 Days AgoPublic

Actions

Here we demonstrate an example of photon-induced events in Herwig.

Using the Les Houches Event Handler

To start, create a new work directory:

mkdir photon_induced
cd photon_induced

Next, download below input/config/analysis files in your work directory.

1. MadGraph configure script: this is a script for generating a sample of unweighted yy > ll events with MadGraph.

wget https://phab-files.hepforge.org/file/download/fmev7v6idvr75ggbdth3/PHID-FILE-i3aaqziv52hgqrsztjuo/MG_make.sh

2. MadGraph parameter and run cards: parameter space and run statistics for the MadGraph run.

wget https://phab-files.hepforge.org/file/download/xjefkopi6whfilhore3q/PHID-FILE-kfvne4b34gu6thn2syt5/run_card.dat

wget https://phab-files.hepforge.org/file/download/6ox3c6if7ekcawjw33am/PHID-FILE-ynhnj545mdxfbtqytoqn/param_card.dat

3. Rivet analysis and plot setup: Rivet analysis files, to be run alongside the Herwig setup.

wget https://phab-files.hepforge.org/file/download/drv32ihccvchwfm6qrjx/PHID-FILE-74aubf5x2iktkajdepci/RivetAnalysis_yy.cc

wget https://phab-files.hepforge.org/file/download/mjsx7nxjbmu3curex3ew/PHID-FILE-p3xj34mwaixlritunlmp/RivetAnalysis_yy.plot

4. Herwig Input files: the main Herwig input file.

wget https://phab-files.hepforge.org/file/download/ccjmisiezgugnatpxbqf/PHID-FILE-66o5cobqy5u3fydkyr7u/LesHouches.in

Once the files are downloaded, please have a look at MG_make.sh. It can be run by:

chmod +x MG_make.sh
[sudo] ./MG_make.sh

Once it is done, you should be able to see a new directory named "yyll" that contains the requited unweighted sample in yyll/Events/run_01/unweighted_events.lhe.gz. Also, you should see a Rivet.so library file appearing in your work directory.

Note: MG_make.sh includes the command

export RIVET_ANALYSIS_PATH=$PWD

which allows Rivet to find a custom-made analysis. If you close your terminal at any point, remember to run this command to set up the environment variables correctly.

We are now set to sun Herwig using these samples. Run

Herwig read LesHouches.in

It should generate a run file named LesHouches.run. Now execute

Herwig run LesHouches.run -N 10000

The last part asks Herwig to run this simulation for 10000 events only. After this is done, you should be able to see several files generated in your work directory. This will generate a number of files in your work directory:

• LesHouches.log: contains the details of each produced event, for the first 10 events (can be changed to include any number of events). The details include incoming/outgoing/intermediate particles and their kinematics. This is useful information to check what is actually going on under the hood.
• LesHouches.out: Statistics for the event handler, including the number of valid events (subtracting vetoed/scrapped events) and a list of sub-process categories with their event count and cross-section.
• LesHouches.tex: A list of publications, relevant to your Herwig setup, which you should cite if you are making use of this specific simulation.
• LesHouches.yoda: Specific results, according to the Rivet analysis you called.

To produce the plots, execute

rivet-mkhtml LesHouches.yoda -o plots

where the -o flag allows you to choose an output directory. Have a look at the generated plot by double-clicking on plots/index.html. If you are unhappy with the statistics, you can improve them by running more events. This might take longer to run, but you improve your run-time by multi-threading your simulation:

Herwig run LesHouches.run -N 100000 -j 5

Here, the -j flag tells Herwig to split this simulation into 5 parallel positions, each running 20000 events. The results will be saved in 5 different sets:

LesHouches-1.yoda
LesHouches-2.yoda
LesHouches-3.yoda
LesHouches-4.yoda
LesHouches-5.yoda

You can combine these by

yodamerge LesHouches-*.yoda -o LesHouches-ALL.yoda

Using Matchbox

Another way of performing this simulation is by using Herwig's Matchbox module, which handles the matrix element generation internally. First, we try this with an LEP collider simulation with incoming photonic beams and then for an LHC simulation, using photon PDFs. In both cases, we assume 13 TeV incoming beams.

Download the following input file in your work directory

wget https://phab-files.hepforge.org/file/download/y2cim4fmtnskxlcz64jv/PHID-FILE-utfokyqh3bibbrc4rizv/LEP-Matchbox-yyll.in

Look inside this file and try to understand the structure and the commands. You can run it by

Herwig read LEP-Matchbox-yyll.in
Herwig run LEP-Matchbox-yyll.run -N 1000000 -j5
yodamerge LEP-Matchbox-yyll-*.yoda -o LEP-Matchbox-ALL.yoda

Now, try to plot LesHouches and Matchbox results together:

rivet-mkhtml LesHouches-ALL.yoda:"Les Houches" LEP-Matchbox-ALL.yoda:"LEP Matchbox" -o ALL-plots

:"Les Houches" and :"Matchbox" are optional tags for the plots. You can either include them or not. Have a look at the results by running ALL-plots/index.html.

Now, repeat the same steps for the following LHC setup file:

wget https://phab-files.hepforge.org/file/download/ahcm4ddudp2if23nomkq/PHID-FILE-yhuibu6zk3stcnccochr/LHC-Matchbox-yyll.in

Herwig read LHC-Matchbox-yyll.in
Herwig run LHC-Matchbox-yyll.run -N 100000 -j5
yodamerge LHC-Matchbox-yyll-*.yoda -o LHC-Matchbox-ALL.yoda

Finally you can plot all your simulations together

rivet-mkhtml LesHouches-ALL.yoda:"Les Houches" LEP-Matchbox-ALL.yoda:"LEP Matchbox" LHC-Matchbox-ALL.yoda:"LHC Matchbox" -o ALL-plots

and compare the results. Note that the statistics chosen for this run are significantly lower than the previous run. This is because this is a much more complex simulation and will need more computation time. If possible, you can use more cores (e.g. -j10) to speed this up or incorporate higher statistics.

Exercise

1- Change the configurations in LHC-Matchbox-yyll.in and LEP-Matchbox-yyll.in files:

read Matchbox/LO-DefaultShower.in
# read Matchbox/LO-DipoleShower.in
# read Matchbox/LO-NoShower.in

Uncomment each line separately to change the parton shower setting from Default (Angularly-ordered) to Dipole and then turn it off and produce .yoda files. Plot them together and check the difference. Does the choice of shower algorithm drastically affect the simulation results?

2 - Herwig-7.3 supports a full QCD+QED+EW parton shower. You can control the parton shower meta by adding the command

set /Herwig/Shower/ShowerHandler:Interactions ALL # options are QCD QED QEDQCD EWOnly ALL

Alternate between these options in both Les Houches and Matchbox runs and plot them against each other and observe the differences.

3 - Replace The default PDFs in the LHC-Matchbox-yyll.in file to CT14qed_proton by changing

cd /Herwig/Partons
set HardLOPDF:PDFName    LUXlep-NNPDF31_nlo_as_0118_luxqed
set HardNLOPDF:PDFName   LUXlep-NNPDF31_nlo_as_0118_luxqed
set ShowerLOPDF:PDFName  LUXlep-NNPDF31_nlo_as_0118_luxqed
set ShowerNLOPDF:PDFName LUXlep-NNPDF31_nlo_as_0118_luxqed
set MPIPDF:PDFName       LUXlep-NNPDF31_nlo_as_0118_luxqed
set RemnantPDF:PDFName   LUXlep-NNPDF31_nlo_as_0118_luxqed

to

cd /Herwig/Partons
set HardLOPDF:PDFName    CT14qed_proton
set HardNLOPDF:PDFName   CT14qed_proton
set ShowerLOPDF:PDFName  CT14qed_proton
set ShowerNLOPDF:PDFName CT14qed_proton
set MPIPDF:PDFName       CT14qed_proton
set RemnantPDF:PDFName   CT14qed_proton

and re-generate the plots. Observer the effects of changing the PDFs. You can choose from any viable photon PDFs.