HAMR: The hierarchical adaptive mesh refinement system. The hierarchical adaptive mesh refinement (HAMR) system is an autonomous, general purpose, flexible, extensible environment that allows the rapid convenient conversion of unigrid solvers into structured adaptive mesh refinement (SAMR) implementations. HAMR’s design is based on a theoretical foundation that describes data, methods, their management, and the relationships between them. These elements allow the development not just of SAMR algorithms, but of SAMR algorithms that require absolutely no knowledge about the application to which they are applied.par The theoretical foundation consists of several parts: a complicated data structure representing the strata of a grid hierarchy; scope and extent characteristics; a variety of data types; several kinds of metadata attributes associated with data items, and rules governing them; a specification lookup table that makes the grid hierarchy self-describing and therefore autonomous; and a user-provided declaration, divided into modules, that describes the data and methods associated with an application, and the relationships between them. Data management of these pieces is actually rather uncomplicated.par Some implementation issues are also discussed. A contemporary physics experiment is run using HAMR, and HAMR’s performance is examined.
References in zbMATH (referenced in 1 article )
Showing result 1 of 1.