sherpa is hosted by Hepforge, IPPP Durham
close Warning: Can't synchronize with repository "(default)" (/hepforge/svn/sherpa does not appear to be a Subversion repository.). Look in the Trac log for more information.

Version 2 (modified by Frank Siegert, 3 years ago) (diff)


Sherpa 2.2.4 (Cho Oyu)

Version 2.2.4 is a bugfix release and adds the support of the Recola one-loop generator

If you spot any problem, please send us an email. This will allow us to reach a mature version of the generator, suitable for all physics analyses. If you would like to be informed about patches and newer releases, please subscribe to our announcement mailing list.


The Sherpa package can be downloaded here:

In can also be accessed via svn through

A Manual is included in the Manual/ directory of the tarball. It is available in the HTML and INFO format and provides extensive documentation of all parameters. It can also be found online.


The main reference describing the physics of Sherpa is JHEP02 (2009) 007. Please cite this publication if you have used Sherpa for your studies. Additionally, each run will produce a file called Sherpa_References.tex which contains information about the modules used in this run and the corresponding publications which should be referred. If in doubt, please seek the advice of the Sherpa authors.


You can find a large collection of example setups for using SHERPA in various

different colliders/processes in the Examples/ directory of the tarball. They are also described in the manual.


Interface changes

  • New default parameters for hadronisation model to improve heavy flavour production fractions and meson/baryon ratios (thanks to Andrii Verbytskyi).
  • Updates to Recola interface with some new parameter (names), please refer to the manual.
  • Make STRICT_METS the default scale setter to guarantee full scale setting accuracy for unweighted events.

New features

  • DIRE NLO parton shower (cf. arXiv:1705.00742)
  • Higher numerical stability in parton shower kinematics (re)construction.
  • Add interface for hadron-level selectors.
  • Add OVERWEIGHT_THRESHOLD allowing user to cap maximal weight during event generation (cf. manual).
  • Add Max_N_TChannels option in Amegic.
  • Restructure gzip support such that it can be switched off at runtime (USE_GZIP).
  • Add possibility to identify partons which originate in the hard decay handler.


  • Bugfix in NLO-matched parton shower emission for non-trivial colour structures like in V+2j@NLO. This fixes an eta asymmetry which was visible at the ~2% level in such samples.
  • BSM models through UFO: fix running coupling behaviour and treatment of exotic color octets.
  • Bugfix in random generator for Poisson distributions with large mean values.
  • Avoid spuriously large weights in NLO-merged samples due to instable PDF ratios.
  • Bugfixes for segfaults and compiler warnings with gcc 7.1.
  • Bugfix for colour assignment in clustering for loop-induced merged processes.
  • Bugfix for decay showering in conjunction with NLO+PS matching.