lala::GaussSeidelIteration Class Reference

#include <fixpoint.hpp>

Public Member Functions
CUDA void	barrier ()
template<class F >
CUDA local::B	iterate (size_t n, const F &f) const
template<class F , class StopFun , class M >
CUDA size_t	fixpoint (size_t n, const F &f, const StopFun &must_stop, B< M > &has_changed)
template<class F , class StopFun >
CUDA size_t	fixpoint (size_t n, const F &f, const StopFun &must_stop)
template<class F , class M >
CUDA size_t	fixpoint (size_t n, const F &f, B< M > &has_changed)
template<class F >
CUDA size_t	fixpoint (size_t n, const F &f)

Detailed Description

A simple form of sequential fixpoint computation based on Kleene fixpoint. At each iteration, the deduction operations \( f_1, \ldots, f_n \) are simply composed by functional composition \( f = f_n \circ \ldots \circ f_1 \). This strategy basically corresponds to the Gauss-Seidel iteration method.

Member Function Documentation

barrier()

CUDA void lala::GaussSeidelIteration::barrier ( )

inline

iterate()

template<class F >

CUDA local::B lala::GaussSeidelIteration::iterate ( size_t n, const F & f ) const

inline

We iterate the function f n times: \( f(0); f(1); \ldots ; f(n); \)

Parameters

`n`	the number of call to f.
`bool	f(size_t i)returnstrueif something has changed fori. \returntrueif for somei,f(i)returnedtrue,false` otherwise.

fixpoint() [1/4]

template<class F , class StopFun , class M >

CUDA size_t lala::GaussSeidelIteration::fixpoint ( size_t n, const F & f, const StopFun & must_stop, B< M > & has_changed )

inline

We execute iterate(n, f) until we reach a fixpoint or must_stop() returns true.

Parameters

`n`	the number of call to f.
`bool	f(size_t i)returnstrueif something has changed fori. \parambool must_stop()returnstrueif we must stop early the fixpoint computation. \paramhas_changedis set totrueif we were not yet in a fixpoint. \return The number of iterations required to reach a fixpoint or untilmust_stop()returnstrue`.

fixpoint() [2/4]

template<class F , class StopFun >

CUDA size_t lala::GaussSeidelIteration::fixpoint ( size_t n, const F & f, const StopFun & must_stop )

inline

Same as fixpoint above without has_changed.

fixpoint() [3/4]

template<class F , class M >

CUDA size_t lala::GaussSeidelIteration::fixpoint ( size_t n, const F & f, B< M > & has_changed )

inline

Same as fixpoint above with must_stop always returning false.

fixpoint() [4/4]

template<class F >

CUDA size_t lala::GaussSeidelIteration::fixpoint ( size_t n, const F & f )

inline

Same as fixpoint above without has_changed and with must_stop always returning false.

---

The documentation for this class was generated from the following file:

• include/lala/fixpoint.hpp