HOPPET is a Fortran 95 package for carrying out DGLAP evolution and other common manipulations of parton distribution functions (PDFs).

Within HOPPET, PDFs are represented on a grid in x-space so as to avoid limitations on the functional form of input distributions. Decent speed and accuracy are obtained through the representation of splitting functions in terms of their convolution with a set of piecewise polynomial basis functions, and Runge-Kutta techniques are used for the evolution in Q.

Unpolarised evolution is provided in the MSbar scheme to NNLO in the coupling, including heavy-quark thresholds, and polarised evolution to NLO. The code is structured so as to provide simple access to the objects representing splitting function and PDFs, making it possible for a user to extend the facilities already provided.

The latest version can always be obtained from

git clone https://github.com/gavinsalam/hoppet

Details of changes are to be found in the file ChangeLog, while summaries of changes between releases are in ReleaseNotes.

You will need a Fortran 95 compiler to compile this package.

The gfortran compiler is widespread. Versions upwards of 4.3 are OK. Older versions (which might be present on legacy systems) generate incorrect code for HOPPET and should be avoided. The code is acceptably fast, though in tests some years ago was not quite as fast as with commercial compilers.

The code has also been tested with the intel (ifort, versions 8.1.033 upwards) and lahey (lf95) compilers.

For details see the INSTALL file. To get moving quickly, just specify an installation prefix and a fortran compiler (FC), and then do

./configure --prefix="..."  FC="..."
make 
make check
make install      [if you're interested]
make install-mod  [if you want the f90 module files installed too]

This is not autotools based: if you're used to more advanced usage of autotools scripts, you'll be disappointed here...

The main example program is example_f90/tabulation_example.f90

An equivalent program with the startup, pdf initialisation and evolution spread across different subroutines is given as tabulation_example_2.

An equivalent program based on the streamlined interface is given as tabulation_example_streamlined.

Some users may prefer a pure f77 interface. Corresponding examples are to be found in the example_f77/ directory. Look inside the suppplied Makefile and if need be edit it manually.

 cd ../example_f77
 # <edit the Makefile directly>
 # compile
 make tabulation_example
 # run the program. Should give output identical to that from
 # example_f90/tabulation_example
 ./tabulation_example

In the same directory there is a C++ example

 make cpp_tabulation_example
 ./cpp_tabulation_example

which again does the same things (though in this case it uses the simpler of the two streamlined initialization calls).

Other programs provided in the example_f77/ directory illustrate the use of the streamlined interface in conjunction with LHAPDF (compare_lhapdf_hoppet.f), and show how to use the feature of getting convolutions with splitting functions (convolution_example.f).

Detailed documentation is available as doc/HOPPET-v1-doc.tex. If you use HOPPET in a scientific publication, please refer to http:https://arxiv.org/abs/0804.3755.

The benchmarking/ directory contains the programs used for the full benchmarking, accuracy and precision testing.

When the hoppet library is being built, on some systems the following warnings may be reported

/usr/bin/ranlib: file: libhoppet_v1.a(hoppet_v1.o) has no symbols
/usr/bin/ranlib: file: libhoppet_v1.a(types.o) has no symbols

ranlib: file: libhoppet_v1.a(hoppet_v1.o) has no symbols
ranlib: file: libhoppet_v1.a(types.o) has no symbols

These warnings do not indicate an actual problem. They are simply a consequence of the fact that some of the source (.f90) files contain only information about interfaces and constants, so that the resulting object (.o) files are essentially empty.

The master branch provides the official HOPPET release. Two additional branches are in use for specific physics applications:

• the qed branch provides evolution including QED and mixed QED+QCD splitting functions, as used in the LUXqed project.
• the struct-func-devel branch provides tools for calculating structure functions, as used in the proVBFH project.