[Rivet] min bias analysis in Rivet

[Andy Buckley]

Sami Kama wrote:
> Hello Andy,

Hi Sami,

I'm copying this message to the Rivet mailing list as well: that's the
best place to ask for support since if I'm busy then someone else may
get round to replying first. And many sets of eyes are better than one!

> I'm going to update the Phojet and re-tune it.  Judith suggested that I
> prepare the analysis in Rivet so that it can be reproduced later. I
> wanted to check the Rivet however I couldn't be able to compile it. I
> tried both tarball and svn versions.
> 
> I downloaded and installed
> libgsl-1.11
> boost-1_37_0
> HepMC-2.04.00
> fastjet-2.3.4
> lhapdf-5.6.0
> pythia-8108
> sherpa-1.1.3
> ThePEG-1.4.0
> Herwig++-2.3.0
> also GENSER directory with the agile-genser-bootstrap script
> 
> Everything up to this point compiles and installs without any problem.

That's good: the Herwig++ and ThePEG are newer than we've yet tested,
though.

> They are all installed with --prefix=/opt/products except pythia and
> agile-genser-bootstrap which is in /opt/products/GENSER.
> 
> I check-out  agile with
> svn co http://svn.hepforge.org/agile/trunk/ agile
> configure with
> 
> ./configure --prefix=/opt/products/
> --with-boost=/opt/products/include/boost-1_37/
> --with-boost-incpath=/opt/products/include/boost-1_37/
> --with-pythia8=/opt/products/GENSER/pythia8108/
> 
> after that AGILe compilation fails with
> 
> CCHerwig.cc:16: error: wrong number of template arguments (4, should be 6)

Ok, so this problem is just in AGILe, which is actually separate from
Rivet itself and not a strict dependency. Furthermore, I'm not surprised
that the Herwig++ interface breaks, since they haven't been very careful
about keeping it invariant. In fact, for precisely this sort of reason
we are dropping Herwig++ support in the next AGILe release!

I recommend that you configure AGILe with the extra --disable--ccherwig
flag: the build should then be okay. And you can proceed with compiling
Rivet with AGILe support. Let me know whether this is successful.

> I'm using SL5, x86_64. Which versions of dependencies do I need to
> compile Rivet?

To compile Rivet itself you don't actually *need* AGILe: it's an
"optional extra" and the 1.1.3 Rivet release will make this break
completely --- that release will get rid of the "rivetgun" tool, and do
everything via the "rivet" command instead. To be able to run the
Fortran generators, your life will be easiest if you disable all the C++
generator builds via the AGILe configure arguments. For testing C++
generators, you don't even need AGILe if building from SVN, since
rivetgun will be built such that it can only read HepMC files and not
run generators natively.

If you need any more help, are interested in testing the newer (and
nicer) "rivet" command method, or want to make Phojet a supported
generator in AGILe, let me know!

> PS:the numbers Judith sent before are the HepData Ids  like
> http://durpdg.dur.ac.uk/cgi-hepdata/hepreac/875503
> http://durpdg.dur.ac.uk/cgi-hepdata/hepreac/1583476

Ah, that's confusing! Those IDs are actually the Spires IDs, since I
presume Judith found the papers by searching Spires and then clicking on
the HepData link: you can look up those numbers directly in Spires with
the "key <num>" search command. HepData's own IDs are much smaller ---
~10000 at most. Anyway, I know what you mean ;)

Best wishes,
Andy

-- 
Dr Andy Buckley
Institute for Particle Physics Phenomenology
Durham University
0191 3343798 | 0191 3732613 | www.insectnation.org