General concept

fastNLO provides the code to create, fill, read, and evaluate fast interpolation tables of pre-computed coefficients in perturbation theory for observables in hadron-induced processes. This allows fast theory predictions of these observables for arbitrary parton distribution functions (of regular shape), renormalization or factorization scale choices, and/or values of alpha_s(Mz) as e.g. needed in PDF fits or in systematic studies. Very time consuming complete recalculations are thus avoided.

Code

The new code release fastNLO Toolkit (v2.5) provides an extended grid format that may contain in so-called InfoBlocks information on the relative numerical uncertainty of each contribution of a fixed-order calculation. The tool fnlo-tk-modify has been updated to allow addition of such information to an interpolation table directly from NNLOJET .dat files, log files of fnlo-tk-statunc applied to a sample of NLOJet++ grids, or a text file. Numerous other improvements serve the adaptation to grid production with the NNLOJET program via the APPLfast interface.

New code named the fastNLO Toolkit (v2.3) has been made available in the form of a prerelease (beta release). The toolkit provides a library with all functionality to create, fill, read, and evaluate interpolation tables in the fastNLO format. It comes with a much improved structure that allows other programs to be interfaced to fastNLO. The only dependency is to the LHAPDF package for the evaluation part. Essentially, it will replace the former fastNLO Reader (v2.1) package.

The interface from NLOJet++ to fastNLO is also available as a prerelease and requires a patched version of NLOJet++ and the fastNLO Toolkit. It is recommended to use it in connection with FastJet for the various jet algorithms. DiffTop has been interfaced to the fastNLO Toolkit and the latest prerelease works with Sherpa via a yet to be relased update of the MCgrid package. An interface to Herwig++ MatchBox is in preparation.

Previously, table evaluation code was released in the form of the fastNLO Reader (v2.1) package. This fastNLO Reader package consists of two independent parts: a C++ library including a documented example, and a Fortran program inherited from the previous fastNLO version 1.4 with somewhat less capabilities. When applicable both codes give the same results at double precision accuracy. The C++ version is capable of evaluating all version 2.1 fastNLO tables (fixed-scale and flexible-scale tables except for some new HERA tables that cannot be read yet) with identical results compared to the toolkit C++ reader. The v2.1 Fortran code is also generic in the sense that it reads all fixed-scale tables within its capabilities, however, because of the lack of dynamic memory allocation it might be necessary to adapt the array dimensioning for particular scenarios, which entails a recompilation. In that case the user receives a warning and execution is stopped.

All packages can be downloaded as autotools-based distribution tar-balls via the links in the sidebar.