covfie: vector fields made easy =============================== covfie (pronounced *coffee*) is a **co**-processor **v**\ ector **fie**\ ld library. This header-only C++ library is designed to make the use of vector fields as easy as possible, both on traditional CPUs as well as a variety of co-processors. covfie is provided in the hope that it is useful for scientists in the domain of high-energy physics, but also other fields where vector fields are common, such as atmospheric sciences. .. figure:: static/bfield.png :figwidth: 40 % :alt: map to buried treasure :align: right The `ATLAS `_ magnetic field, rendered entirely on a GPU with *covfie*. The covfie documentation consists of three major components. The first is the user guide, which establishes many of the core ideas behind covfie and details -- in a global sense -- how it can be used in applications. The second component details the benchmarking part of covfie, which is a major part of the software. The third component is the API reference, which describes in depth the different types and methods that covfie exposes. User guide ---------- The user guide is a mostly prosaic introduction to the covfie library and its use. The user guide describes the main design goals of the library, as well as a step by step introduction to using it. If you are an end user and you are looking to use covfie in a wider project, this is for you. .. toctree:: :maxdepth: 3 user/design user/installation user/integration user/quickstart user/backend user/types user/ownership user/conversion user/io Developer guide --------------- The developer guide is a mostly prosaic summary of the internal workings of covfie, which should be useful to users who want to extend covfie with their own backends. .. toctree:: :maxdepth: 3 developer/build developer/backend developer/concepts developer/contributing Benchmarking guide ------------------ In addition to being a header-only library, covfie is also -- in part -- a benchmark. The modular design of covfie makes it easy to rapidly and thoroughly test all kinds of new representations and storage strategies. API reference ------------- Coming soon...