Skip to content

Latest commit

 

History

History
28 lines (22 loc) · 2.49 KB

README.md

File metadata and controls

28 lines (22 loc) · 2.49 KB

bamboo examples based on RDataFrame tutorials with open data

Binder

These can be run directly on Binder through the badge above, or installed locally (see the documentation for more details).

In case the link above does not work, this one can be used to launch the session, where this repository can be cloned with git clone https://github.com/pieterdavid/bamboo-opendata-examples.git.

Currently, the following examples are available:

  • CMS H->4l search, based on the df103 tutorial. The default configuration file uses the skimmed (51MB) input data, for performance testing the full (12GB) version higgs4l_tutorial_CMSOpenData_full.yml (if file access is fast enough multithreading, e.g. --threads 4, works very nicely)
    bambooRun -m higgs4l_tutorial_CMSOpenData.py:Higgs4L higgs4l_tutorial_CMSOpenData.yml -o test_out/df103
  • ATLAS W boson analysis, based on the df105 tutorial
    bambooRun -m wanalysis_tutorial_ATLASOpenData.py:WAnalysis wanalysis_tutorial_ATLASOpenData.yml -o test_out/df105
  • the IRIS-HEP analysis description language benchmarks, also available in a separate repository
    bambooRun -m adl_benchmarks.py:IRISHEP_ADLBenchmarks adl_benchmarks.yml -o test_out/adl_benchmarks

Some of these, especially of the ADL benchmarks, go through enough combinatorics to benefit from implicit multithreading in RDataFrame (DOI 10.5281/zenodo.260230); this can be enabled by passing --threads N (with N three or four).