[Rivet] pilemc improvements

[James Monk]

Dear Riveteers,

The pilemc package that was originally developed by Christian Ruhr sometimes comes in handy as a tool for Rivet pile up studies, but I found it didn’t work so well for <mu> = 100s (i.e. LHC upgrade and beyond) because you need to generate 100s of min bias input events per output event.  Therefore I’ve made a few improvements that might make it a bit more useful.  

Specifically, you can now turn on “flitting” (it flits around the input file - all the other good option names were already taken) so it will combine random events from a file, rather than sequential events.  The random seed is also now settable.  It also allows multiple input pile up files, including a cross section specification, so you can add a higher pT file to get decent stats in the tail and avoid the same jets turning up repeatedly in your pile up.   Finally, it allows gzip input (if configured with —enable-gzip), but not in the random mode because gzip input is not random access.  This all makes it a lot easier to produce decent sized high <mu> pile up samples for particle-level studies with Rivet and friends.

Now if only someone would improve the documentation at pilemc.hepforge.org

Here ends the advertisement :)

cheers,

James