[Rivet] [Rivet-svn] rivet: 8 new changesets

[Andy Buckley]

I agree... the numerics can be dangerous, and it's a problem someone 
else has solved.

I don't remember if Eigen1 has those solvers, though. Maybe needs to 
wait for an upgrade to Eigen 3.

Andy

PS. Just spotted the branch -- d'oh, I'd assumed this was happening on 
the default branch so we could target it for 2.6.0 rather than 2.5.3. 
It's hopefully a safe change, but for future reference I don't think we 
should be changing "big picture" things like build dependencies in minor 
releases. Maybe this forces us to go straight to 2.6.0, but we had other 
things earmarked for that which aren't near completion...


On 23/11/16 09:09, David Grellscheid wrote:
> Hi,
>
> If we're keeping Eigen(3) around anyway, wouldn't it be cleaner to use
> their built-in eigenvalue calculations everywhere? That avoids us having
> to worry about numerical stability issues and typos at each usage point.
>
> See you,
>
>   David
>
> On 22/11/2016 16:59, Holger Schulz wrote:
>> Hi David,
>>
>> removing/replacing eigen (with eigen3) is the next logical step.
>>
>> Holger
>>
>>
>> On 22/11/16 15:51, David Grellscheid wrote:
>>> Hi Holger,
>>>
>>> that looks good! Have you also removed the dependency on the "Eigen"
>>> package now, which Rivet ships internally?
>>>
>>> See you,
>>>
>>>   David
>>>
>>>
>>> On 22/11/2016 16:30, Rivet Mercurial wrote:
>>>> details:   https://rivet.hepforge.org/hg/rivet/rev/ec1c137e5bca
>>>> branches:  release-2-5-x
>>>> changeset: 5603:ec1c137e5bca
>>>> user:      Holger Schulz <holger.schulz at durham.ac.uk>
>>>> date:      Tue Nov 22 15:11:46 2016 +0000
>>>> description:
>>>> Remove gsl from configure.ac
>>>>
>>>> details:   https://rivet.hepforge.org/hg/rivet/rev/0a8a2634ba4a
>>>> branches:  release-2-5-x
>>>> changeset: 5604:0a8a2634ba4a
>>>> user:      Holger Schulz <holger.schulz at durham.ac.uk>
>>>> date:      Tue Nov 22 15:12:28 2016 +0000
>>>> description:
>>>> Remove MatrixDiag to get rid of GSL
>>>>
>>>> details:   https://rivet.hepforge.org/hg/rivet/rev/e1fbe22bff42
>>>> branches:  release-2-5-x
>>>> changeset: 5605:e1fbe22bff42
>>>> user:      Holger Schulz <holger.schulz at durham.ac.uk>
>>>> date:      Tue Nov 22 15:13:54 2016 +0000
>>>> description:
>>>> Replace gsl dependence by using simply eigenvalue and eigenvector calculation
>>>>
>>>> details:   https://rivet.hepforge.org/hg/rivet/rev/29fe4ba6e54b
>>>> branches:  release-2-5-x
>>>> changeset: 5606:29fe4ba6e54b
>>>> user:      Holger Schulz <holger.schulz at durham.ac.uk>
>>>> date:      Tue Nov 22 15:14:24 2016 +0000
>>>> description:
>>>> Remove gsl from pyext
>>>>
>>>> details:   https://rivet.hepforge.org/hg/rivet/rev/c371a2458e8f
>>>> branches:  release-2-5-x
>>>> changeset: 5607:c371a2458e8f
>>>> user:      Holger Schulz <holger.schulz at durham.ac.uk>
>>>> date:      Tue Nov 22 15:15:42 2016 +0000
>>>> description:
>>>> use trivial eigenvalue calculation to get rid of gsl
>>>>
>>>> details:   https://rivet.hepforge.org/hg/rivet/rev/1f5c88158565
>>>> branches:  release-2-5-x
>>>> changeset: 5608:1f5c88158565
>>>> user:      Holger Schulz <holger.schulz at durham.ac.uk>
>>>> date:      Tue Nov 22 15:16:51 2016 +0000
>>>> description:
>>>> Remove build of FoxWolfram as it is not used anywhere TODO: replace legendre polynomials from gsl with own code
>>>>
>>>> details:   https://rivet.hepforge.org/hg/rivet/rev/7b7cf5654277
>>>> branches:  release-2-5-x
>>>> changeset: 5609:7b7cf5654277
>>>> user:      Holger Schulz <holger.schulz at durham.ac.uk>
>>>> date:      Tue Nov 22 15:18:28 2016 +0000
>>>> description:
>>>> Also print collider in debug
>>>>
>>>> details:   https://rivet.hepforge.org/hg/rivet/rev/bc71333e8b76
>>>> branches:  release-2-5-x
>>>> changeset: 5610:bc71333e8b76
>>>> user:      Holger Schulz <holger.schulz at durham.ac.uk>
>>>> date:      Tue Nov 22 15:18:56 2016 +0000
>>>> description:
>>>> Use event weights in analysis
>>>>
>>>> diffs (truncated from 283 to 50 lines):
>>>>
>>>> --- a/bin/rivet	Mon Nov 21 22:38:36 2016 +0000
>>>> +++ b/bin/rivet	Tue Nov 22 15:18:56 2016 +0000
>>>> @@ -283,6 +283,8 @@
>>>>                              msg += "  [" + " ".join(a.keywords()) + "]"
>>>>                          if a.luminosityfb():
>>>>                              msg += "  [ \int L = %s fb^{-1} ]"%a.luminosityfb()
>>>> +                        if a.collider():
>>>> +                            msg += " [ Collider: %s ] "%a.collider()
>>>>                  print msg
>>>>                  #os.write(tf, msg + "\n")
>>>>          # if os.path.getsize(tfpath) > 0:
>>>> --- a/configure.ac	Mon Nov 21 22:38:36 2016 +0000
>>>> +++ b/configure.ac	Tue Nov 22 15:18:56 2016 +0000
>>>> @@ -145,12 +145,12 @@
>>>>
>>>>
>>>>  ## GNU Scientific Library
>>>> -AC_SEARCH_GSL
>>>> -AC_CEDAR_HEADERS([gsl], , , [AC_MSG_ERROR([GSL (GNU Scientific Library) is required])])
>>>> -oldCPPFLAGS=$CPPFLAGS
>>>> -CPPFLAGS="$CPPFLAGS -I$GSLINCPATH"
>>>> -AC_CHECK_HEADER([gsl/gsl_vector.h], [], [AC_MSG_ERROR([GSL vectors not found.])])
>>>> -CPPFLAGS=$oldCPPFLAGS
>>>> +#AC_SEARCH_GSL
>>>> +#AC_CEDAR_HEADERS([gsl], , , [AC_MSG_ERROR([GSL (GNU Scientific Library) is required])])
>>>> +#oldCPPFLAGS=$CPPFLAGS
>>>> +#CPPFLAGS="$CPPFLAGS -I$GSLINCPATH"
>>>> +#AC_CHECK_HEADER([gsl/gsl_vector.h], [], [AC_MSG_ERROR([GSL vectors not found.])])
>>>> +#CPPFLAGS=$oldCPPFLAGS
>>>>
>>>>
>>>>  ## Disable build/install of standard analyses
>>>> @@ -238,7 +238,7 @@
>>>>  ## Set default build flags
>>>>  AM_CPPFLAGS="-I\$(top_srcdir)/include -I\$(top_builddir)/include"
>>>>  #AM_CPPFLAGS="$AM_CPPFLAGS -I\$(top_srcdir)/include/eigen3"
>>>> -AM_CPPFLAGS="$AM_CPPFLAGS \$(GSL_CPPFLAGS)"
>>>> +#AM_CPPFLAGS="$AM_CPPFLAGS \$(GSL_CPPFLAGS)"
>>>>  dnl AM_CPPFLAGS="$AM_CPPFLAGS \$(BOOST_CPPFLAGS)"
>>>>  AM_CPPFLAGS="$AM_CPPFLAGS -I\$(YODAINCPATH)"
>>>>  AM_CPPFLAGS="$AM_CPPFLAGS -I\$(HEPMCINCPATH)"
>>>> --- a/include/Rivet/Math/Matrices.hh	Mon Nov 21 22:38:36 2016 +0000
>>>> +++ b/include/Rivet/Math/Matrices.hh	Tue Nov 22 15:18:56 2016 +0000
>>>> @@ -5,6 +5,5 @@
>>>>  #include "Rivet/Math/MatrixN.hh"
>>>>  #include "Rivet/Math/Matrix3.hh"
>>>>  #include "Rivet/Math/LorentzTrans.hh"
>>>> -#include "Rivet/Math/MatrixDiag.hh"
>>>>
>>>>  #endif
>>>> _______________________________________________
>>>> Rivet-svn mailing list
>>>> Rivet-svn at projects.hepforge.org
>>>> https://www.hepforge.org/lists/listinfo/rivet-svn
>>>>
>>> _______________________________________________
>>> Rivet mailing list
>>> Rivet at projects.hepforge.org
>>> https://www.hepforge.org/lists/listinfo/rivet
>>
> _______________________________________________
> Rivet mailing list
> Rivet at projects.hepforge.org
> https://www.hepforge.org/lists/listinfo/rivet
>


-- 
Dr Andy Buckley, Lecturer / Royal Society University Research Fellow
Particle Physics Expt Group, University of Glasgow