/*============================================================================= This file is part of FLINT. FLINT is free software; you can redistribute it and/or modify it under the terms of the GNU General Public License as published by the Free Software Foundation; either version 2 of the License, or (at your option) any later version. FLINT is distributed in the hope that it will be useful, but WITHOUT ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU General Public License for more details. You should have received a copy of the GNU General Public License along with FLINT; if not, write to the Free Software Foundation, Inc., 51 Franklin St, Fifth Floor, Boston, MA 02110-1301 USA =============================================================================*/ /****************************************************************************** Copyright (C) 2013 Mike Hansen ******************************************************************************/ ******************************************************************************* Memory management ******************************************************************************* void fq_mat_init(fq_mat_t mat, slong rows, slong cols, const fq_ctx_t ctx) Initialises \code{mat} to a \code{rows}-by-\code{cols} matrix with coefficients in $\mathbb{F}_{q}$ given by \code{ctx}. All elements are set to zero. void fq_mat_init_set(fq_mat_t mat, fq_mat_t src, const fq_ctx_t ctx) Initialises \code{mat} and sets its dimensions and elements to those of \code{src}. void fq_mat_clear(fq_mat_t mat, const fq_ctx_t ctx) Clears the matrix and releases any memory it used. The matrix cannot be used again until it is initialised. This function must be called exactly once when finished using an \code{fq_mat_t} object. void fq_mat_set(fq_mat_t mat, fq_mat_t src, const fq_ctx_t ctx) Sets \code{mat} to a copy of \code{src}. It is assumed that \code{mat} and \code{src} have identical dimensions. ******************************************************************************* Basic properties and manipulation ******************************************************************************* fq_struct * fq_mat_entry(fq_mat_t mat, slong i, slong j) Directly accesses the entry in \code{mat} in row $i$ and column $j$, indexed from zero. No bounds checking is performed. fq_struct * fq_mat_entry_set(fq_mat_t mat, slong i, slong j, fq_t x, const fq_ctx_t ctx) Sets the entry in \code{mat} in row $i$ and column $j$ to \code{x}. slong fq_mat_nrows(fq_mat_t mat, const fq_ctx_t ctx) Returns the number of rows in \code{mat}. slong fq_mat_ncols(fq_mat_t mat, const fq_ctx_t ctx) Returns the number of columns in \code{mat}. void fq_mat_swap(fq_mat_t mat1, fq_mat_t mat2, const fq_ctx_t ctx) Swaps two matrices. The dimensions of \code{mat1} and \code{mat2} are allowed to be different. void fq_mat_zero(fq_mat_t mat, const fq_ctx_t ctx) Sets all entries of \code{mat} to 0. ******************************************************************************* Printing ******************************************************************************* void fq_mat_print_pretty(const fq_mat_t mat, const fq_ctx_t ctx) Pretty-prints \code{mat} to \code{stdout}. A header is printed followed by the rows enclosed in brackets. int fq_mat_fprint_pretty(FILE * file, const fq_mat_t mat, const fq_ctx_t ctx) Pretty-prints \code{mat} to \code{file}. A header is printed followed by the rows enclosed in brackets. In case of success, returns a positive value. In case of failure, returns a non-positive value. void fq_mat_print(const fq_mat_t mat, const fq_ctx_t ctx) Prints \code{mat} to \code{stdout}. A header is printed followed by the rows enclosed in brackets. int fq_mat_fprint(FILE * file, const fq_mat_t mat, const fq_ctx_t ctx) Prints \code{mat} to \code{file}. A header is printed followed by the rows enclosed in brackets. In case of success, returns a positive value. In case of failure, returns a non-positive value. ******************************************************************************* Window ******************************************************************************* void fq_mat_window_init(fq_mat_t window, const fq_mat_t mat, slong r1, slong c1, slong r2, slong c2, const fq_ctx_t ctx) Initializes the matrix \code{window} to be an \code{r2 - r1} by \code{c2 - c1} submatrix of \code{mat} whose \code{(0,0)} entry is the \code{(r1, c1)} entry of \code{mat}. The memory for the elements of \code{window} is shared with \code{mat}. void fq_mat_window_clear(fq_mat_t window, const fq_ctx_t ctx) Clears the matrix \code{window} and releases any memory that it uses. Note that the memory to the underlying matrix that \code{window} points to is not freed. ******************************************************************************* Random matrix generation ******************************************************************************* void fq_mat_randtest(fq_mat_t mat, flint_rand_t state, const fq_ctx_t ctx) Sets the elements of \code{mat} to random elements of $\mathbb{F}_{q}$, given by \code{ctx}. int fq_mat_randpermdiag(fq_mat_t mat, fq_struct * diag, slong n, flint_rand_t state, const fq_ctx_t ctx) Sets \code{mat} to a random permutation of the diagonal matrix with $n$ leading entries given by the vector \code{diag}. It is assumed that the main diagonal of \code{mat} has room for at least $n$ entries. Returns $0$ or $1$, depending on whether the permutation is even or odd respectively. void fq_mat_randrank(fq_mat_t mat, slong rank, flint_rand_t state, const fq_ctx_t ctx) Sets \code{mat} to a random sparse matrix with the given rank, having exactly as many non-zero elements as the rank, with the non-zero elements being uniformly random elements of $\mathbb{F}_{q}$. The matrix can be transformed into a dense matrix with unchanged rank by subsequently calling \code{fq_mat_randops()}. void fq_mat_randops(fq_mat_t mat, slong count, flint_rand_t state, const fq_ctx_t ctx) Randomises \code{mat} by performing elementary row or column operations. More precisely, at most \code{count} random additions or subtractions of distinct rows and columns will be performed. This leaves the rank (and for square matrices, determinant) unchanged. void fq_mat_randtril(fq_mat_t mat, flint_rand_t state, int unit, const fq_ctx_t ctx) Sets \code{mat} to a random lower triangular matrix. If \code{unit} is 1, it will have ones on the main diagonal, otherwise it will have random nonzero entries on the main diagonal. void fq_mat_randtriu(fq_mat_t mat, flint_rand_t state, int unit, const fq_ctx_t ctx) Sets \code{mat} to a random upper triangular matrix. If \code{unit} is 1, it will have ones on the main diagonal, otherwise it will have random nonzero entries on the main diagonal. ******************************************************************************* Comparison ******************************************************************************* int fq_mat_equal(fq_mat_t mat1, fq_mat_t mat2, const fq_ctx_t ctx) Returns nonzero if mat1 and mat2 have the same dimensions and elements, and zero otherwise. int fq_mat_is_zero(const fq_mat_t mat, const fq_ctx_t ctx) Returns a non-zero value if all entries \code{mat} are zero, and otherwise returns zero. int fq_mat_is_empty(const fq_mat_t mat, const fq_ctx_t ctx) Returns a non-zero value if the number of rows or the number of columns in \code{mat} is zero, and otherwise returns zero. int fq_mat_is_square(const fq_mat_t mat, const fq_ctx_t ctx) Returns a non-zero value if the number of rows is equal to the number of columns in \code{mat}, and otherwise returns zero. ******************************************************************************* Addition and subtraction ******************************************************************************* void fq_mat_add(fq_mat_t C, const fq_mat_t A, const fq_mat_t B, const fq_ctx_t ctx) Computes $C = A + B$. Dimensions must be identical. void fq_mat_sub(fq_mat_t C, const fq_mat_t A, const fq_mat_t B, const fq_ctx_t ctx) Computes $C = A - B$. Dimensions must be identical. void fq_mat_neg(fq_mat_t A, const fq_mat_t B, const fq_ctx_t ctx) Sets $B = -A$. Dimensions must be identical. ******************************************************************************* Matrix multiplication ******************************************************************************* void fq_mat_mul(fq_mat_t C, const fq_mat_t A, const fq_mat_t B, const fq_ctx_t ctx) Sets $C = AB$. Dimensions must be compatible for matrix multiplication. $C$ is not allowed to be aliased with $A$ or $B$. This function automatically chooses between classical and KS multiplication. void fq_mat_mul_classical(fq_mat_t C, const fq_mat_t A, const fq_mat_t B, const fq_ctx_t ctx) Sets $C = AB$. Dimensions must be compatible for matrix multiplication. $C$ is not allowed to be aliased with $A$ or $B$. Uses classical matrix multiplication. void fq_mat_mul_KS(fq_mat_t C, const fq_mat_t A, const fq_mat_t B, const fq_ctx_t ctx) Sets $C = AB$. Dimensions must be compatible for matrix multiplication. $C$ is not allowed to be aliased with $A$ or $B$. Uses Kronecker substitution to perform the multiplication over the integers. void fq_mat_submul(fq_mat_t D, const fq_mat_t C, const fq_mat_t A, const fq_mat_t B, const fq_ctx_t ctx) Sets $D = C + AB$. $C$ and $D$ may be aliased with each other but not with $A$ or $B$. ******************************************************************************* LU decomposition ******************************************************************************* slong fq_mat_lu(slong * P, fq_mat_t A, int rank_check, const fq_ctx_t ctx) Computes a generalised LU decomposition $LU = PA$ of a given matrix $A$, returning the rank of $A$. If $A$ is a nonsingular square matrix, it will be overwritten with a unit diagonal lower triangular matrix $L$ and an upper triangular matrix $U$ (the diagonal of $L$ will not be stored explicitly). If $A$ is an arbitrary matrix of rank $r$, $U$ will be in row echelon form having $r$ nonzero rows, and $L$ will be lower triangular but truncated to $r$ columns, having implicit ones on the $r$ first entries of the main diagonal. All other entries will be zero. If a nonzero value for \code{rank_check} is passed, the function will abandon the output matrix in an undefined state and return 0 if $A$ is detected to be rank-deficient. This function calls \code{fq_mat_lu_recursive}. slong fq_mat_lu_classical(slong * P, fq_mat_t A, int rank_check, const fq_ctx_t ctx) Computes a generalised LU decomposition $LU = PA$ of a given matrix $A$, returning the rank of $A$. The behavior of this function is identical to that of \code{fq_mat_lu}. Uses Gaussian elimination. slong fq_mat_lu_recursive(slong * P, fq_mat_t A, int rank_check, const fq_ctx_t ctx) Computes a generalised LU decomposition $LU = PA$ of a given matrix $A$, returning the rank of $A$. The behavior of this function is identical to that of \code{fq_mat_lu}. Uses recursive block decomposition, switching to classical Gaussian elimination for sufficiently small blocks. ******************************************************************************* Reduced row echelon form ******************************************************************************* slong fq_mat_rref(fq_mat_t A, const fq_ctx_t ctx) Puts $A$ in reduced row echelon form and returns the rank of $A$. The rref is computed by first obtaining an unreduced row echelon form via LU decomposition and then solving an additional triangular system. ******************************************************************************* Triangular solving ******************************************************************************* void fq_mat_solve_tril(fq_mat_t X, const fq_mat_t L, const fq_mat_t B, int unit, const fq_ctx_t ctx) Sets $X = L^{-1} B$ where $L$ is a full rank lower triangular square matrix. If \code{unit} = 1, $L$ is assumed to have ones on its main diagonal, and the main diagonal will not be read. $X$ and $B$ are allowed to be the same matrix, but no other aliasing is allowed. Automatically chooses between the classical and recursive algorithms. void fq_mat_solve_tril_classical(fq_mat_t X, const fq_mat_t L, const fq_mat_t B, int unit, const fq_ctx_t ctx) Sets $X = L^{-1} B$ where $L$ is a full rank lower triangular square matrix. If \code{unit} = 1, $L$ is assumed to have ones on its main diagonal, and the main diagonal will not be read. $X$ and $B$ are allowed to be the same matrix, but no other aliasing is allowed. Uses forward substitution. void fq_mat_solve_tril_recursive(fq_mat_t X, const fq_mat_t L, const fq_mat_t B, int unit, const fq_ctx_t ctx) Sets $X = L^{-1} B$ where $L$ is a full rank lower triangular square matrix. If \code{unit} = 1, $L$ is assumed to have ones on its main diagonal, and the main diagonal will not be read. $X$ and $B$ are allowed to be the same matrix, but no other aliasing is allowed. Uses the block inversion formula $$ \begin{pmatrix} A & 0 \\ C & D \end{pmatrix}^{-1} \begin{pmatrix} X \\ Y \end{pmatrix} = \begin{pmatrix} A^{-1} X \\ D^{-1} ( Y - C A^{-1} X ) \end{pmatrix} $$ to reduce the problem to matrix multiplication and triangular solving of smaller systems. void fq_mat_solve_triu(fq_mat_t X, const fq_mat_t U, const fq_mat_t B, int unit, const fq_ctx_t ctx) Sets $X = U^{-1} B$ where $U$ is a full rank upper triangular square matrix. If \code{unit} = 1, $U$ is assumed to have ones on its main diagonal, and the main diagonal will not be read. $X$ and $B$ are allowed to be the same matrix, but no other aliasing is allowed. Automatically chooses between the classical and recursive algorithms. void fq_mat_solve_triu_classical(fq_mat_t X, const fq_mat_t U, const fq_mat_t B, int unit, const fq_ctx_t ctx) Sets $X = U^{-1} B$ where $U$ is a full rank upper triangular square matrix. If \code{unit} = 1, $U$ is assumed to have ones on its main diagonal, and the main diagonal will not be read. $X$ and $B$ are allowed to be the same matrix, but no other aliasing is allowed. Uses forward substitution. void fq_mat_solve_triu_recursive(fq_mat_t X, const fq_mat_t U, const fq_mat_t B, int unit, const fq_ctx_t ctx) Sets $X = U^{-1} B$ where $U$ is a full rank upper triangular square matrix. If \code{unit} = 1, $U$ is assumed to have ones on its main diagonal, and the main diagonal will not be read. $X$ and $B$ are allowed to be the same matrix, but no other aliasing is allowed. Uses the block inversion formula $$ \begin{pmatrix} A & B \\ 0 & D \end{pmatrix}^{-1} \begin{pmatrix} X \\ Y \end{pmatrix} = \begin{pmatrix} A^{-1} (X - B D^{-1} Y) \\ D^{-1} Y \end{pmatrix} $$ to reduce the problem to matrix multiplication and triangular solving of smaller systems.