Thomas.h File Reference

#include "types.h"
#include "mangling.h"

Include dependency graph for Thomas.h:

Go to the source code of this file.

Functions
void	API_cgttsv (fortran_int *N, fortran_complex *A, fortran_complex *B, fortran_complex *X, fortran_int *INFO)
void	API_dgttsv (fortran_int *N, fortran_double *A, fortran_double *B, fortran_double *X, fortran_int *INFO)
void	API_sgttsv (fortran_int *N, fortran_real *A, fortran_real *B, fortran_real *X, fortran_int *INFO)
void	API_zgttsv (fortran_int *N, fortran_double_complex *A, fortran_double_complex *B, fortran_double_complex *X, fortran_int *INFO)
void	thomas (fortran_int *N, fortran_real *A, fortran_real *B, fortran_real *X, fortran_int *INFO)
	Thomas algorithm (tridiagonal solver) for A * X = B.
void	thomas (fortran_int *N, fortran_double *A, fortran_double *B, fortran_double *X, fortran_int *INFO)
	Thomas algorithm (double-precision).
void	thomas (fortran_int *N, fortran_complex *A, fortran_complex *B, fortran_complex *X, fortran_int *INFO)
	Thomas algorithm (single-precision complex).
void	thomas (fortran_int *N, fortran_double_complex *A, fortran_double_complex *B, fortran_double_complex *X, fortran_int *INFO)
	Thomas algorithm (double-precision complex).

Function Documentation

API_cgttsv()

void API_cgttsv	(	fortran_int *	N,
		fortran_complex *	A,
		fortran_complex *	B,
		fortran_complex *	X,
		fortran_int *	INFO
	)

API_dgttsv()

void API_dgttsv	(	fortran_int *	N,
		fortran_double *	A,
		fortran_double *	B,
		fortran_double *	X,
		fortran_int *	INFO
	)

API_sgttsv()

void API_sgttsv	(	fortran_int *	N,
		fortran_real *	A,
		fortran_real *	B,
		fortran_real *	X,
		fortran_int *	INFO
	)

API_zgttsv()

void API_zgttsv	(	fortran_int *	N,
		fortran_double_complex *	A,
		fortran_double_complex *	B,
		fortran_double_complex *	X,
		fortran_int *	INFO
	)

thomas() [1/4]

void thomas	(	fortran_int *	N,
		fortran_complex *	A,
		fortran_complex *	B,
		fortran_complex *	X,
		fortran_int *	INFO
	)

Thomas algorithm (single-precision complex).

This is an overloaded member function, provided for convenience. It differs from the above function only in what argument(s) it accepts.

Parameters

[in]	N	Pointer to the matrix size (N x N).
[in,out]	A	Flat row-major tridiagonal matrix of size N*N (modified on output).
[in,out]	B	Right-hand side vector of size N (modified on output).
[out]	X	Solution vector of size N.
[out]	INFO	Pointer to the return code.

Definition at line 104 of file Thomas.h.

thomas() [2/4]

void thomas	(	fortran_int *	N,
		fortran_double *	A,
		fortran_double *	B,
		fortran_double *	X,
		fortran_int *	INFO
	)

Thomas algorithm (double-precision).

This is an overloaded member function, provided for convenience. It differs from the above function only in what argument(s) it accepts.

Parameters

[in]	N	Pointer to the matrix size (N x N).
[in,out]	A	Flat row-major tridiagonal matrix of size N*N (modified on output).
[in,out]	B	Right-hand side vector of size N (modified on output).
[out]	X	Solution vector of size N.
[out]	INFO	Pointer to the return code.

Definition at line 90 of file Thomas.h.

thomas() [3/4]

void thomas	(	fortran_int *	N,
		fortran_double_complex *	A,
		fortran_double_complex *	B,
		fortran_double_complex *	X,
		fortran_int *	INFO
	)

Thomas algorithm (double-precision complex).

This is an overloaded member function, provided for convenience. It differs from the above function only in what argument(s) it accepts.

Parameters

[in]	N	Pointer to the matrix size (N x N).
[in,out]	A	Flat row-major tridiagonal matrix of size N*N (modified on output).
[in,out]	B	Right-hand side vector of size N (modified on output).
[out]	X	Solution vector of size N.
[out]	INFO	Pointer to the return code.

Definition at line 118 of file Thomas.h.

thomas() [4/4]

void thomas	(	fortran_int *	N,
		fortran_real *	A,
		fortran_real *	B,
		fortran_real *	X,
		fortran_int *	INFO
	)

Thomas algorithm (tridiagonal solver) for A * X = B.

Direct solver for tridiagonal linear systems A * X = B using the Thomas algorithm. A is supplied as a full N x N matrix in flat row-major storage and is overwritten during computation.

On output, X contains the solution vector. No pivoting is performed; zero diagonal entries produce failure.

Parameters

[in]	N	Pointer to the matrix size (N x N).
[in,out]	A	Flat row-major tridiagonal matrix of size N*N (modified on output).
[in,out]	B	Right-hand side vector of size N (modified on output).
[out]	X	Solution vector of size N.
[out]	INFO	Pointer to the return code: 0: success < 0: zero diagonal detected at row |INFO|

Definition at line 76 of file Thomas.h.