[Rivet] HepMC units in AGILe

[Frank Siegert]

Andy Buckley, Wednesday 21 April 2010:
> On 21/04/10 17:29, Frank Siegert wrote:
> > I have the bad suspicion that AGILe doesn't respect the default HepMC
> > units when filling HepMC events. When using AGILe compiled with
> > HepMC from GENSER it produces events that claim to have MEV/MM units,
> > but the numbers are still GEV numbers (e.g. E=900 for a Tev-I beam).
> > 
> > Am I wrong in my diagnosis (and have just killed 230 jobs in vain) or
> > is this indeed a relatively serious problem because the bootstrap
> > script uses HepMC 2.05 from GENSER by default?
> 
> Could be a serious problem... I can't remember if AGILe is specifying
> the HepMC units properly, since the experiment version was always a
> blocker on what was done. I think that Rivet should handle the incoming
> HepMC units properly, but AGILe quite possibly doesn't write them.

Well, the problem is not that it doesn't write them. Quite the opposite:
It does write them (simply because HepMC::IO_GenEvent writes them) but the
numbers don't correspond to them.
So when I feed such a generated file into Rivet, it assumes all numbers are
MeV, even though they were meant to be GeV by AGILe. And of course
everything will be wrong from there.
The problem might be that the HepMC::IO_HEPEVT converter doesn't respect
the default units? I know that at least FPythia in AGILe is affected.

Frank