The DTFE method
The Delaunay Tessellation Field Interpolation (DTFE) method represents the natural way of going from discrete samples/measurements to values on a periodic grid using the maximum of information contained in the points set (that is the measurements points or particle distribution).
The DTFE method is especially suitable for astronomical data due to the following reasons:
- Preserves the multi-scale character of the point distribution. This is the case in numerical simulations of large scale structure (LSS) where the density varies over more than 6 orders of magnitude.
- Preserves the local geometry of the point distribution. This is important in recovering sharp features like the different components of the LSS (i.e. clusters, filaments, walls and voids).
- DTFE does not depend on user defined parameters or choices.
- The interpolated fields are volume weighted (versus mass weighted quantities in most other interpolation schemes). This can have a significant effect especially when comparing with analytical predictions which are volume weighted.
The DTFE public software
The DTFE public code is a C++ implementation of the DTFE grid interpolation method. The code was written by Marius Cautun at the Kapteyn Astronomical Institute, Netherlands, with the purpose of analyzing cosmological numerical simulations and galaxy redshift survey. Even though the code was designed with astrophysics in mind, it can be used for problems in a wide range of fields where one needs to interpolate from a discrete set of points to a grid.
The code was designed using a modular philosophy and with a wide set of features that can easily be selected using the different program options. The DTFE code is also written using OpenMP directives which allow it to run in parallel on shared-memory architectures.
The code comes with a complete documentation and with a multitude of examples that detail the program features. Moreover, a help desk is available for information and assistance for troubleshooting problems.