KaHIP - Karlsruhe High Quality Partitioning - is a family of graph partitioning programs that tackle the balanced graph partitioning problem. It focuses on solution quality and implements flow-based methods, more-localized local searches, and several parallel and sequential meta-heuristics.
GriF is a collaborative grid framework to support computational chemistry applications. It is meant to be used as a tool to facilitate massive grid calculations and also to improve scientific collaboration. Accordingly, GriF facilitates profiling the users of grid communities in order to systematically evaluate the work carried out in a grid and to foster its sustainability.
SourceAFIS is a fingerprint recognition/matching SDK (library), or more generally an Automated Fingerprint Identification System (AFIS). It essentially compares two fingerprints and decides whether they belong to the same person. It can quickly search a large database of registered fingerprints. It comes with an easy-to-use API (pure .NET and Java) plus assorted applications and tools.
GHCN Processor is a command-line tool that reads temperature data from the Global Historical Climatic Network (GHCN) database and produces an annual or monthly temperature series in CSV format for an arbitrary set of stations. Stations are filtered based on a simple EL expression passed to the tool. For example, you can select only stations that are in the Northern Hemisphere, in hilly and rural locations. You can also select stations that started reporting in a given year, and so on. The tool supports more than one method of grid partitioning, station combination, and can use both the adjusted data and raw unadjusted data.
GRALE is a set of tools - a library and a number of accompanying applications - to study gravitational lenses. Gravitational lenses are astronomical objects so massive that their gravitational pull even deflects light rays. This can cause multiple copies of the same background object to be visible, like a cosmic mirage. The locations and shapes of these copies can provide information about the mass distribution of the gravitational lens, which GRALE can help recover using a genetic algorithm-based method. Apart from these so-called lens inversions, it's also possible to simulate gravitational lenses.
Global Paths Matching is an implementation of the global paths graph matching algorithm proposed by Maue and Sanders in "Engineering Algorithms for Approximate Weighted Matching" (WEA'07). Given a graph G=(V,E), a matching M is a set of edges without common vertices, i.e. the graph G=(V,M) has a degree of at most one. The algorithm scans the edges in order of decreasing weight (or rating), constructing a collection of paths and even length cycles. These paths initially contain no edges. While scanning the edges, the set is extended by successively adding applicable edges, which are those connecting two endpoints of different paths or two endpoints of an odd length path. Optimal solutions/matchings are computed for each path and cycle using dynamic programming.