About

libGrid has been written and is still developed by Sebastian Reiter (Sebas.nosp@m.tian.nosp@m..Reit.nosp@m.er@g.nosp@m.csc.u.nosp@m.ni-f.nosp@m.rankf.nosp@m.urt..nosp@m.de) at the Goethe Center for Scientific Computing at the University of Frankfurt, Germany (http://www.g-csc.de). libGrid is developed as a part of the ug4 framework.

Amongst others, the following people contributed to libGrid:
Sebastian Reiter, Martin Stepniewski, Nicolas Tessore

What is libGrid?

libGrid is a library for the manipulation of 1, 2 and 3 dimensional grids written in C++. It is designed in a flexible way that should allow its usage for all grid-related algorithms.

Main features

libGrid tries to treat grids in a flexible, but at the same time fast and efficient way. This goal is achieved by separation of topology and data.
While the elements of a grid are stored in linked lists, the associated data is stored in separate arrays. This not only improves cache-hits considerably but also allows to add (and remove) data to the elements of the grid at runtime. To avoid performance problems, data is not added to elements one by one, but to all elements of the same type (Vertices / Edges / Faces / Volumes) at once.
Data management is fully automated. The access to data is simple, typesafe and fast.

libGrid supports both normal and hierarchical grids.

See also: ug::Grid, ug::MultiGrid

A grid in libGrid can consist of vertices, edges, faces and volumes. Besides others, the following important concrete types are supported:

Vertex
Edge
Triangle, Quadrilateral
Tetrahedron, Hexahedron, Prism, Pyramid

See also: lib_grid_geometric_objects

New types can be added without much hassle. Those types can be combined as desired. A grid could for example consist of some Vertices that are connected by Triangles plus a single edge. However, since it is sometimes desireable that missing edges (or faces) are created automatically if a new face or volume is added, grids feature options through which such behaviours can be turned on and off.

Automated access to the neighbourhood of elements is another feature of libGrid. Neighbours of an element can be collected on the fly, or are optionally stored for each element. This allows the user to choose between better runtime or lower memory consumption. This again can be changed at runtime, which allows algorithms to enable or disable features as required.

Since libGrid heavily uses template-programming, efficient algorithms can be implemented with a minimal amount of code. However, while thorough knowledge of template-programming is beneficial for writing intelligent and efficient code, only a very basic understanding is required to start programming with libGrid.

LibGrid also features a mechanism to register observers at a grid. Those observers are then notified if changes to the grids topology are made. Combined with the possibility to dynamically add data to a grid at runtime, those observers are a powerful tool to create stable, robust and modular code.

See also: ug::Selector,; ug::SubsetHandler,; ug::MGSubsetHandler

Usage

An introduction to the most important ideas behind libGrid and a short overview of the coding-style can be found here:

Overview