[Rivet] problems with RivetGun/Alpgen

Andy Buckley andy.buckley at ed.ac.uk
Mon Apr 12 21:48:23 BST 2010


rivet-bootstrap also has a --devmode switch which will make it get the
latest head versions of both Rivet and AGILe from SVN... but it sounds
like that won't work at Fermilab!

You could use this mode at CERN, then copy the build/agile directory to
Fermilab. This *should* work, but of course there are potential snags:
you need to run "make clean" and run configure again, and be careful
with the pyext/AGILe_wrap.cc and pyext/AGILe.py files: "make clean" will
delete them and I don't know whether the Fermilab system has a valid
version of the SWIG utility to rebuild them. You'll find out if you get
compiler errors during the pyext build phase! If that happens, try
copying them from the CERN build, and please let us know (with attached
config.log and make output) if you have any further trouble.

Sorry for the inconvenience... the next AGILe release should solve your
problem, but it's just finding the time to get everything finished off
and make the tarballs that's difficult!

Best wishes,
Andy


On 12/04/10 20:31, renkel at physics.smu.edu wrote:
> Hi Andy, Regretfully, I am running this at FermiLab. I am using
> rivet-bootstrap which has an option --install-agile. This installs agile
> to my area. I don't have SVN here at FNAL, only CVS. Can you help me and
> say what I should do to get it from your SVN? I can use my lxplus
> account to install it there and then move FNAL, but will it work?
> 
> Peter.
> 
> Quoting Andy Buckley <andy.buckley at ed.ac.uk>:
> 
>> Can you try using the SVN head version of AGILe: I think this problem is
>> fixed there. We should make a new AGILe release sometime soon, but your
>> feedback on whether the development version works would be very helpful!
>>
>> Andy
>>
>>
>> On 12/04/10 20:10, renkel at physics.smu.edu wrote:
>>> Hi Andy,
>>>
>>> I am exporting AGILE_GEN_PATH. But the problem is that the system
>>> looks for
>>> /local/lib/libAGILeAlpGen.so, and I have just libAGILeAlpGenFPythia.so
>>> and libAGILeAlpGenFHerwig.so. If I try to naively rename one of these
>>> files, AGILE finds the generator, but crashes at execution. Do I need to
>>> specify something when installing AGILe to have this shared library in
>>> please? Or can I tell Alpgen to use the files I have?
>>>
>>> Thank you!
>>>
>>> Peter.
>>>
>>> Quoting Andy Buckley <andy.buckley at ed.ac.uk>:
>>>
>>>> On 09/04/10 23:30, renkel at physics.smu.edu wrote:
>>>>> Hi Holger,
>>>>>
>>>>> Sorry, but I still fight with it, The last test agile-runmc
>>>>> --list-gens
>>>>> fails - it does not give me alpgen. I was not able to track where
>>>>> agile
>>>>> finds the information on available generators. It goes into some c++,
>>>>> which I don't understand. May be, you can help me here, then I can put
>>>>> this information by hand. Thank you!
>>>>
>>>> Are you setting and exporting the AGILE_GEN_PATH variable? If you run
>>>> agile-runmc like so
>>>>
>>>> agile-runmc --list-gens -lTRACE
>>>>
>>>> you will get an enormous amount of output documenting all the paths and
>>>> permutations that the library loader searches through to try and find
>>>> each generator version... that will help you to debug whether you have
>>>> the correct structure in which to find the AlpGen, PYTHIA, HERWIG, etc.
>>>> libraries: the trick is to make it look as much as possible like the
>>>> Genser area on AFS (although it is possible to leave out the
>>>> platform/architecture tags directories)
>>>>
>>>> Cheers,
>>>> Andy
>>>>
>>>> -- 
>>>> Dr Andy Buckley
>>>> SUPA Advanced Research Fellow
>>>> Particle Physics Experiment Group, University of Edinburgh
>>>>
>>>
>>>
>>>
>>>
>>>
>>
>>
>> -- 
>> Dr Andy Buckley
>> SUPA Advanced Research Fellow
>> Particle Physics Experiment Group, University of Edinburgh
>>
> 
> 
> 
> 
> 


-- 
Dr Andy Buckley
SUPA Advanced Research Fellow
Particle Physics Experiment Group, University of Edinburgh


More information about the Rivet mailing list