Skip to content

Latest commit

 

History

History
100 lines (64 loc) · 3.8 KB

README.md

File metadata and controls

100 lines (64 loc) · 3.8 KB
Raw

SPADE (Static Polymorphic Algorithms for Differential Equations)

SPADE is a framework for the solution of partial differential equations in space and time. It provides a series of high-order algorithms that are intended to make the the implementation of solvers as smooth and efficient as possible.

Building and Installation

SPADE is currently a header-only, single-include implementation that utilized features available in the c++20 standard. To build a project using SPADE, simply include the header src/spade.h and ensure that the c++20 standard is used.

Examples

Examples are currently found in the runs directory. runs/channel contains an example solver for a compressible turbulent channel flow, solved using the Navier-stokes equations.

Namespaces

Everything that SPADE can do is organized into its various namespaces, all of which fall under the spade namespace, e.g. the io namespace would be used as spade::io. The list of namespaces within SPADE are as follows:

  • timing: contains various utilities for measuring the performance of implemented applications.

  • algs: i.e. "algorithms". Contains a series of general-purpose algorithms that are useful for managine the state of arrays or performing loops compactly.

  • aliases: an intermediate namespace that forwards STL types for use in CPU implementations and providing polymorphism for GPU implementations. For example, aliases::vector<double> will be std::vector<double> when compiling as a CPU application, but will be something else when compiling with CUDA.

  • array_container: A specialized namespace that handles underlying storage for grid arrays.

  • detail or [namespace]::detail: a quarantined space for highly specialized implementations that should not be called or manipulated outside very specific situations. Don't come in here!

  • block_config: provides various implementations of block configurations for grid arrays.

  • cli_args: contains utilities for handling command-line arguments

  • consts: contains mathematical constants

  • coords: contains everything to do with coordinate systems and their jacobians for use with generalized-coordinates formulations

  • ctrs: i.e. "containers". Contains various flavors of basic arrays with arithmetic operations enabled.

  • dims: soon to be deprecated!

  • finite_diff: algoriths for the calculation of finite-difference coefficients

  • grid: contains everything to do with the semantics of handling data stored on a grid

  • io: contains functions for outputting data to files

  • linear_algebra: contains implementations for numerical linear algebra algorithms for small-size problems

  • parallel: contains types and functions dealing with the use of a parallel computing context

  • utils: contains general-purpose utilities

  • partition: contains implementations of partitioning algorithms for load-balancing

  • pde_algs: contains algorithms specifically intended for the numerical solution of partial differential equations

  • reduce_ops: contains general definitions of reduction operations

  • static_math: contains implementations for compile-time mathematical calculations

  • time_integration: contains algorithms for causal integration in a single dimension, e.g. through time

  • fetch: contains specialized algorithms for buffering grid array data for the calculation of PDE terms. Careful in here!

  • convective: contains implementations of inviscid fluxes for the solution of the compressible Navier-Stokes equations (CNSE)

  • viscous: contains implementations of viscous fluxes for the solution of CNSE

  • fluid_state: contains implementations of and conversions between fluid states of a gas

  • navier_stokes_mms: soon to be deprecated!

  • viscous_laws: contains state-dependent viscous laws for CNSE

  • amr: contains algorithms for adaptive mesh refinement