|
[Rivet] installation problems (It works!)James Monk jmonk at hep.ucl.ac.ukTue Jul 29 15:49:01 BST 2008
I just re-built AGILe against HepMC 2.03.08 and still see the seg fault when running pythia. The build of HepMC 2.03.08 has not changed (been re-built etc) in the past few months, has it? On 29 Jul 2008, at 15:30, Andy Buckley wrote: > James Monk wrote: >> On 28 Jul 2008, at 22:07, Andy Buckley wrote: >>> Holger Schulz wrote: > >>>> It seems as if Rivet does not work with the generators present at >>>> /afs/cern... >>>> >>>> Out of curiosity I installed a local copy of Pythia 6.418 >>>> (using the mkGenserArea script) on my DESY account >>>> to see if it works and it does! >>> >>> I just tested the bootstrap script at CERN and, yes, the released >>> versions of Rivet and AGILe work but the development ones segfault, >>> citing a HepMC problem. I'm not sure what's going on, but I'll look >>> into it! >> >> I re-installed rivet at CERN using the bootstrap. It did not find >> either of Herwig++ or Pythia8 but built FHerwig and FPythia. > > James, can you hack the bootstrap script to make AGILe find Herwig++ > and > Pythia 8 on AFS? It would be good to have them available, since both > now > work! > >> FPythia >> seg faults like you say, but FHerwig only spews out the following >> message repeatedly: >> >> IO_HERWIG::translate_herwig_to_pdg_id HERWIG particle 205 translates >> to zero. >> >> which I guess is not actually a problem. > > (I've copied the Genser developers in on this) > > Hmm, the 2.3.9 release was specifically to fix problems with > IO_HERWIG: > is this behaviour expected? > >> The Pythia seg fault happens >> somewhere in the HepMC_IO hepevt wrapper. I see that Genser >> installed >> the new version of HepMC (2.03.09) just last week. > > Yes that's what I get. I also get strange behaviour with the > (unreleased) Genser bootstrap version of Pythia 6, but building the > same > generator with AGILe's mkGenserArea script works! I've not tried a > stand-alone run with HepMC 2.3.9, though. > > Genser people, any comments? > > Andy >
More information about the Rivet mailing list |