eic / afterburner

Monte Carlo Afterburner for Crossing Angle and Beam Effects
0 stars 1 forks source link

afterburner

EIC MC afterburner. Software package to provide framework independent well validated crossing angle and beam effects C++ library and HepMC file converter (abconv) for Electron Ion Collider.

Physics simulated:

Software:

Installation

The converter and benchmark require libyaml-cpp. Those are available in package managers for the most of systems:

apt install libyaml-cpp-dev     # Debian/Ubuntu/Mint
yum install yaml-cpp-devel      # RHEL/CentOS/Fedora
git clone https://eicweb.phy.anl.gov/monte_carlo/afterburner.git
mkdir -p afterburner/build
cd afterburner/build

# setup cmake (c++ code lives in cpp directory) 
cmake ../cpp

# to build all
cmake --build ../cpp --target all

# to install
cmake --build ../cpp --target install -- -j 8

Preinstalled versions are available in ATHENA containers.

CLI command

Examples:

# Process my whole file, add crossing angle, beam effects, do validation plots
abconv my.hepmc

# Same, but limit the number of events to 1000 and set the output name as 'test'
abconv my.hepmc -o test -l 1000

# Just convert hepmc3 to hepmc2 don't use afterburner, don't create plots
abconv my.hepmc -f hepmc2 --ab-off --plot-off 

# Just convert hepmc3 to hepmc.root don't use afterburner, don't create plots
abconv my.hepmc -f treeroot --ab-off --plot-off 

Note: In the final example above, you want the treeroot option. The root option creates a (rather strange) flat file.

How exactly the afterburner works illustrated by abconv my.hepmc command.

Afterburner algorithm

  1. AB opens the input file, gets beam particles and extract beam energy settings (from the first event). See beam energy settings for the details below.

  2. AB awaits that there is no crossing angle between beam particles. To check this AB calculates a crossing angle between the beam particles. If the crossing angle is not zero and --exit-ca flag is set - AB exits; without the flag a warning is issued.

  3. AB processes events applying crossing angle and beam effects for each event. -s,--ev-start, -e,--ev-end, -l,--limit flags limit the number of events to process

  4. By defaults AB also creates *.hist.root file that contains validation histograms
    --plot-off flag can switch off histograms creation

P.S. Is one needs just HepMC3->HepMC2 converter one can use --ab-off flag to disable afterburner during conversion. This way abconv could be use as different HepMC formats converter.

Beam energy settings

(!) For now IP8 is a copy of IP6 parameters with 35 mrad crossing angle. It will be populated with IP8 own parameters as soon as they are officially published.

Input file requirements

All options:

Flag Description
-h,--help Print this help message and exit
-v,--version Shows package version number
-o,--output TEXT Base name for Output files ((!) no extension)
-p,--preset TEXT Beams configuration 0: IP6 High divergence[default], 1: IP6 High acceptance, 2: IP6 eAu
-i,--in-format TEXT Input format: auto [default], hepmc2, hepmc3, hpe, lhef, gz, treeroot, root
-f,--out-format TEXT Output format: hepmc3 [default], hepmc2, dot, none (no events file is saved)
-s,--ev-start INT Start event index (all previous are skipped)
-e,--ev-end INT End event index (end processing after this event)
-l,--limit UINT Limit number of events to process. (Shutdown after this number of parsed events)
--ab-off No afterburner is applied
--plot-off Don't produce validation plots
--exit-ca Check existing crossing angle and exit if CA>1mrad (1)

(1) How --exit-ca works exactly: when afterburner processes the first event it checks if it has 2 beam particles (fails with non zero code if not) and then calculates their crossing angle with the --exit-ca is set and crossing angle absolute value is > 1mrad program ends with 0 code. This method is very coarse as if in a source file the crossing angle is 0 but all beam effects do exist, beam effects will be applied twice

-p,--preset flag, values [0,1,2,3,4,5] set config and auto determine energy from source file:

The other options sets energy settings manually, not checking the source file:

ip6_hidiv_41x5ip6_hidiv_100x5ip6_hidiv_100x10ip6_hidiv_275x10ip6_hidiv_275x18
ip6_hiacc_41x5ip6_hiacc_100x5ip6_hiacc_100x10ip6_hiacc_275x10ip6_hiacc_275x18
ip6_eau_41x5 ip6_eau_110x5 ip6_eau_110x10 ip6_eau_110x18 -
ip8_hidiv_41x5ip8_hidiv_100x5ip8_hidiv_100x10ip8_hidiv_275x10ip8_hidiv_275x18
ip8_hiacc_41x5ip8_hiacc_100x5ip8_hiacc_100x10ip8_hiacc_275x10ip8_hiacc_275x18
ip8_eau_41x5 ip8_eau_110x5 ip8_eau_110x10 ip8_eau_110x18 -

Example of manual configuration setting:

abconv -p ip6_hidiv_100x5 source_file.hepmc

(!) For now IP8 is a copy of IP6 parameters with 35 mrad crossing angle. It will be populated with IP8 own parameters as soon as they are officially published.

Validation

The validation is performed by comparison of the resulting particle distributions with Pythia8 beam effects simulation. The later was also validated with Fun4All afterburner implementation.

To apply benchmarks. Head-on data (no crossing angle nor effects) is located:

https://eics3.sdcc.bnl.gov:9000/minio/eictest/ATHENA/WG/BeamEffects/

The validation plots are generated and can be viewed in python/comparison.ipynb

Essential plots from the latest validation run:

The discrepancy in low Eta has been explained by a slight phase space difference in generated headon and full beam effects files on Pythia8 side. TL;DR; It is Fine

41x5 GeV 100x10 GeV 275x18 GeV
Eta comparison Eta comparison Eta comparison
Phi comparison Phi comparison Phi comparison
atan2_pxpz hadron comparison atan2_pxpz hadron comparison atan2_pxpz hadron comparison
atan2_pypz hadron comparison atan2_pypz hadron comparison atan2_pypz hadron comparison
atan2_pxpz lepton comparison atan2_pxpz lepton comparison atan2_pxpz lepton comparison
atan2_pypz lepton comparison atan2_pypz lepton comparison atan2_pypz lepton comparison
Vertex X comparison Vertex X comparison Vertex X comparison
Vertex Z comparison Vertex Z comparison Vertex Z comparison
Vertex T comparison Vertex T comparison Vertex T comparison

References