nml is a simple matrix and linear algebra library written in standard C.

Related tags

Math c neat matrix linear-algebra gauss-elimination linear-algebra-library lu-decomposition nml gauss-jordan ansi-c linear-algorithms reduced-row-echelon-form row-echelon-form
Overview
 .----------------.  .----------------.  .----------------. 
| .--------------. || .--------------. || .--------------. |
| | ____  _____  | || | ____    ____ | || |   _____      | |
| ||_   \|_   _| | || ||_   \  /   _|| || |  |_   _|     | |
| |  |   \ | |   | || |  |   \/   |  | || |    | |       | |
| |  | |\ \| |   | || |  | |\  /| |  | || |    | |   _   | |
| | _| |_\   |_  | || | _| |_\/_| |_ | || |   _| |__/ |  | |
| ||_____|\____| | || ||_____||_____|| || |  |________|  | |
| |              | || |              | || |              | |
| '--------------' || '--------------' || '--------------' |      Neat Matrix Library
 '----------------'  '----------------'  '----------------'

nml is a simple matrix and linear algebra library written in standard C.

Code should be portable and there are no dependencies.

It currently supports:

The library is still under development, but a few thousands test cases are already implemented, covering the most complex algorithms (REF, RREF, LUP, QR, DET, INV, BACKWARD SUBSTITION, FORWARD SUBSTITION, etc.)

Table of Contents

Compile / Run Examples

The build file for the library it's called nml.sh. It's actually a bash script (not a makefile!).

Building the library

./nml.sh clean build

This will compile the library, create a dist folder where you will find *.a static library file and the header files.

gcc and ar should be available in $PATH.

If you want to use the clang compiler instead of gcc you need to manually edit the ./nml.sh file, changing the variable CC from gcc to clang. Nothing else should be changed.

# COMPILING RELATED
CC=clang #<----------------- here
CCFLAGS="-Wall -c"
CCFLAGS_EXAMPLES="-Wall"

Building the examples

Examples can be found in the ./examples folder.

To build the code examples:

./nml.sh clean examples
  1. This will create an examples/lib folder where the libnml.a and the header files will be copied;
  2. The examples/*.c will be compiled with the latest version of libnml;
  3. For each examples/*.c an executable (*.ex) will be created.

To run an example:

# ./nml.sh clean examples && ./examples/.ex
./nml.sh clean examples && ./examples/playground.ex

Running the tests

To run the tests

./nml.sh clean test
  1. This will create a test/lib folder where the libnml.a and the header files will be copied;
  2. Each test tests/*.c will be compiled with the latest version of libnml;
  3. For each test tests/*/c an executable (*.ex) will be created.

The test data was generated using sympy. In the tests/generators/ folder you can find the python3 (.py) scripts used to generate the data.

Cleaning

./nml.sh clean

This will clean everything (*.o,*.ex,*.a) and will leave the library folder in a clean state.

How to use the library

A few examples can be found in the ./examples folder folder.

Creating matrices

All the methods are interacting with the nml_mat struct:

typedef struct nml_mat_s {
  unsigned int num_rows;
  unsigned int num_cols;
  double **data;
  int is_square;
} nml_mat;

To interact the elements of the matrix:

nml_mat *m = ...
m->data[i][j] = ...

Creating a new Matrix

The methods for a creating a new matrix are:

  • nml_mat *nml_mat_new(unsigned int num_rows, unsigned int num_cols)
    • Creates a num_rows * num_cols matrix of zeroes.
  • nml_mat *nml_mat_sqr(unsigned int size)
    • Creates a square size * size matrix of zeroes.
  • nml_mat *nml_mat_eye(unsigned int size)
    • Creates an identity size * size matrix.
  • nml_mat *nml_mat_cp(nml_mat *m)
    • Returns a new identitcal copy of matrix m.

Everytime we create a matrix, we dynamically allocate memory. To free the memory please use: nml_mat_free(nml_mat *m).

#include <stdlib.h>
#include <stdio.h>

#include "lib/nml.h"

int main(int argc, char *argv[]) {


  nml_mat* m, *mx;

  printf("\nCreating an empty matrix with 2x3\n");
  m = nml_mat_new(2,3);
  nml_mat_print(m);
  nml_mat_free(m);

  printf("\nCreating a square matrix 5x5 \n");
  m = nml_mat_sqr(5);
  nml_mat_print(m);
  nml_mat_free(m);

  printf("\nCreating an ID 7x7 Matrix and copying it into another matrix:\n");
  m = nml_mat_eye(7);
  mx = nml_mat_cp(m);
  nml_mat_print(m);
  nml_mat_print(mx);
  nml_mat_free(m);
  nml_mat_free(mx);

  return 0;
}

To run the example:

./nml.sh clean examples && examples/creating_a_matrix.ex

Creating a marray from an array (double[N])

An array can be used as the "data source" for the Matrix by using:

  • nml_mat *nml_mat_from(unsigned int num_rows, unsigned int num_cols, unsigned int n_vals, double *vals)
    • num_rows and num_cols represent the dimensions of the matrix;
    • n_vals how many values to read from the vals source. If n_vals is smaller than the product num_cols * num_rows, 0.0 will be used as the default value;
    • vals the array containing double values.
#include <stdlib.h>
#include <stdio.h>

#include "lib/nml.h"

int main(int argc, char *argv[]) { 
    double array[6] = { 
        1.0, 0.2, 3.0, 4.0, 5.0, 3.1 
    };
    nml_mat* my;
    
    // 3 rows, 2 columns
    // read exactly 6 numbers from array[6]
    my = nml_mat_from(3, 2, 6, array);
    nml_mat_print(my);
    nml_mat_free(my);

    // 4 rows, 2 columns
    // read exactly 3 numbers from array[6]
    my = nml_mat_from(4, 2, 3, array);
    nml_mat_print(my);
    nml_mat_free(my);

    return 0;
}

To run the example:

./nml.sh clean examples && examples/creating_a_matrix_from_an_array.ex

Creating a Matrix from an external file

The two methods that can be used to create a matrix from a file on disk are:

  • nml_mat *nml_mat_fromfile(const char *file)
    • Create a matrix from the file path. If the file cannot be opened a NULL matrix will be returned.
  • nml_mat *nml_mat_fromfilef(FILE *f)
    • Creates a matrix from am already opened stream f. Does not automatically close the stream (FILE).

In the file, the matrix has the following format:

4 5
0.0     1.0     2.0     5.0     3.0
3.0     8.0     9.0     1.0     4.0
2.0     3.0     7.0     1.0     1.0
0.0     0.0     4.0     3.0     8.0

On the first line 4 represents the number of rows and 5 represents the number of columns of the Matrix Then next lines contain the matrix elements: 4 * 5 = 20 numbers.

Example code:

#include <stdlib.h>
#include <stdio.h>

#include "lib/nml.h"

int main(int argc, char *argv[]) {
    const char *f = "examples/data/matrix1.data";
    nml_mat *from_file = nml_mat_fromfile(f);
    nml_mat_print(from_file);
    nml_mat_free(from_file);

    // Or if the file is already opened

    FILE *m_file = fopen("examples/data/matrix2.data", "r");
    nml_mat *from_file2 = nml_mat_fromfilef(m_file);
    nml_mat_print(from_file2);
    nml_mat_free(from_file2);
    fclose(m_file);

    return 0;
}

To run the example:

./nml.sh clean examples && ./examples/creating_a_matrix_from_file.ex

Creating a matrix from user input

The nml_mat *nml_mat_fromfilef(FILE *f) can be called, with f=stdin.

Code example:

#include <stdlib.h>
#include <stdio.h>

#include "lib/nml.h"

int main(int argc, char *argv[]) {
    nml_mat *from_file2 = nml_mat_fromfilef(stdin);
    nml_mat_print(from_file2);
    nml_mat_free(from_file2);
    return 0;
}

To run the example:

./nml.sh clean examples && examples/creating_a_matrix_from_user_input.ex

Creating randomized matrices

Creating a randomized matrix can be done with the following two methods:

  • nml_mat *nml_mat_rnd(unsigned int num_rows, unsigned int num_cols, double min, double max)
    • Creates a randomized matrix of size num_rows * num_cols;
    • The random values are between min and max;
  • nml_mat *nml_mat_sqr_rnd(unsigned int size, double min, double max)
    • Creates a randomized matrix of size size * size;
    • The random values are between min and max;
#include <stdlib.h>
#include <stdio.h>
#include <time.h>

#include "lib/nml.h"

int main(int argc, char *argv[]) {
  srand(time(NULL)); // Should be called once per program
  nml_mat *m = nml_mat_rnd(5, 5, -10.0, 10.0);
  nml_mat_print(m);
  nml_mat_free(m);
  return 0;
}

To run the example:

./nml.sh clean examples && examples/create_randomized_matrix.ex

Check if two matrices are equal

There are two "equality" methods for matrices:

  • int nml_mat_eqdim(nml_mat *m1, nml_mat *m2)

    • Tests if two matrices have the same dimension.

  • int nml_mat_eq(nml_mat *m1, nml_mat *m2, double tolerance)

    • Test if two matrices are equal:
      • They have the same dimensions
      • The elements are equal or close of being equal.
    • If you want the elements to be "exactly" eqaul, tolerance=0.0
#include <stdlib.h>
#include <stdio.h>
#include <time.h>

#include "lib/nml.h"

int main(int argc, char *argv[]) {

    srand(time(NULL));
    nml_mat *m1 = nml_mat_rnd(2, 3, 1.0, 10.0);
    nml_mat *m2 = nml_mat_rnd(2, 3, 1.0, 10.0);

    if (nml_mat_eq(m1, m2, 0.001)) {
        printf("Wow, what were the oddss..\n");
    } else {
        printf("It's ok, nobody is that lucky!\n");
    }
    if (nml_mat_eqdim(m1, m2)) {
        printf("At least we know they both have the same number of rows and columns.\n");
    }

    nml_mat_free(m1);
    nml_mat_free(m2);
    return 0;
}

To run the example:

./nml.sh clean examples && ./examples/matrix_equality.ex

Accesing and modifying matrix elements

Select rows and columns

Two methods can be used to select rows and columns from a source matrix (nml_mat*):

  • nml_mat *nml_mat_col_get(nml_mat *m, unsigned int col)
  • nml_mat *nml_mat_row_get(nml_mat *m, unsigned int row)

The following code extracts every column of a given random matrix into a temporary column matrix (nml_mat*):

#include <stdlib.h>
#include <stdio.h>
#include <time.h>

#include "lib/nml.h"

int main(int argc, char *argv[]) {
  printf("\nExtract all matrix columns from a Matrix as matrices\n");
  srand(time(NULL));
  nml_mat *m = nml_mat_rnd(5, 5, -10.0, 10.0);
  nml_mat *col;
  nml_mat_print(m);
  int i = 0;
  for(i = 0; i < m->num_cols; i++) {
    col = nml_mat_col_get(m, i);
    nml_mat_print(col);
    nml_mat_free(col);
  }
  nml_mat_free(m);
  return 0;
}

To run the example:

./nml.sh clean examples && examples/select_columns.ex

Set all elements to a value

Use: void nml_mat_all_set(nml_mat *matrix, double value)

#include <stdlib.h>
#include <stdio.h>
#include <math.h>

#include "lib/nml.h"

int main(int argc, char *argv[]) {
    // Creates a matrix of zeros of size = 5
    nml_mat *pi_mat = nml_mat_sqr(5);

    // Sets all elements to PI
    nml_mat_all_set(pi_mat, M_PI);

    nml_mat_print(pi_mat);
    nml_mat_free(pi_mat);
    return 0;
}

To run the example:

./nml.sh clean examples && ./examples/set_all_elements.ex

Set the first diagonal to a value

Use: int nml_mat_diag_set(nml_mat *matrix, double value)

#include <stdlib.h>
#include <stdio.h>
#include <math.h>

#include "lib/nml.h"

int main(int argc, char *argv[]) {
    // Creates a matrix of zeros of size = 5
    nml_mat *pi_mat = nml_mat_sqr(5);

    // Sets the first diagonal to PI
    nml_mat_diag_set(pi_mat, M_PI);

    nml_mat_print(pi_mat);
    nml_mat_free(pi_mat);
    return 0;
}

To run the example:

./nml.sh clean examples && examples/set_diagonal_elements.ex

Scalar multiply the matrix

Use:

  • nml_mat *nml_mat_smult(nml_mat *m, double num)
    • Multiplies all elements of matrix m with num. A new matrix is returned.
  • int nml_mat_smult_r(nml_mat *m, double num)
    • Multiplies all elements of matrix m with num. All changes are done on matrix m.
#include <stdlib.h>
#include <stdio.h>

#include "lib/nml.h"

int main(int argc, char *argv[]) {
    nml_mat *m = nml_mat_eye(5);

    // Multiply all elements of m with 2.0
    // and return a new matrix

    nml_mat *new_m = nml_mat_smult(m, 2.0);

    if (!(nml_mat_eq(m, new_m, 0.0))) {
        printf("It's normal to see this message.\n");
    }

    // Multiply all elements of m with 2.0
    // m is modified, no new matrix is created
    nml_mat_smult_r(m, 2.0);

    if (nml_mat_eq(m, new_m, 0.0)) {
        printf("It's even more normal to see this message.\n");
    }

    nml_mat_free(m);
    nml_mat_free(new_m);
    
    return 0;
}

To run the example:

./nml.sh clean examples && examples/scalar_multiply.ex

Multiply rows

Use:

  • nml_mat *nml_mat_row_mult(nml_mat *m, unsigned int row, double num)
    • Multiplies all elements from row row in matrix m with scalar num. A new matrix is returned. m remains un-altered.
  • int nml_mat_row_mult_r(nml_mat *m, unsigned int row, double num)
    • Multiplies all elements from row row in matrix m with scalar num. The changes are done directly on matrix m.

Example:

#include <stdlib.h>
#include <stdio.h>

#include "lib/nml.h"

int main(int argc, char *argv[]) {
    nml_mat *a = nml_mat_new(4,5);
    nml_mat_all_set(a, 1.0);
    int i = 0;
    for(; i < a->num_rows; ++i) {
        // Changes are doing on matrix a
        // row[i] is multiplied with (double) i
        nml_mat_row_mult_r(a, i, (double)i);
    }
    nml_mat_print(a);

    // Create a new matrix b by multiplying row[1] 
    // in matrix a with 5.0.
    // Matrix a remains unchanged
    nml_mat *b = nml_mat_row_mult(a, 1, 5.0);
    nml_mat_print(b);
    nml_mat_free(a);
    nml_mat_free(b);
    return 0;
}

To run the example:

./nml.sh clean examples && examples/multiply_rows.ex

To run the example:

./nml.sh clean examples && examples/multiply_rows.ex

Add rows

The following methods are used to add a row to another row (with a multiplicator). This method is usally used when implementing various forms of matrix reduction or decompositions.

Use:

  • nml_mat *nml_mat_row_addrow(nml_mat *m, unsigned int where, unsigned int row, double multiplier)
    • This will do the following: m->data[where][...] *= m->data[row][...] * multiplier. The results will be kept in a new matrix. Matrix m remains unchanged.
  • int nml_mat_row_addrow_r(nml_mat *m, unsigned int where, unsigned int row, double multiplier)
    • This will do the following: m->data[where][...] *= m->data[row][...] * multiplier. The changes are done directly on m.
#include <stdlib.h>
#include <stdio.h>

#include "lib/nml.h"

int main(int argc, char *argv[]) {

    nml_mat *m = nml_mat_rnd(5, 4, 1.0, 2.0);
    nml_mat_print(m);

    // Add row[1] elements to row[2] elements

    nml_mat_row_addrow_r(m, 2, 1, 1.0);

    // Add row[1] to row[0] with a multiplier of 2.0

    nml_mat_row_addrow_r(m, 0, 1, 2.0);

    nml_mat_print(m);
    nml_mat_free(m);

    return 0;
}

To run the example:

./nml.sh clean examples && ./examples/row_plus_row.ex

Modifying the matrix structure

Remove rows and columns

To remove columns:

  • nml_mat *nml_mat_col_rem(nml_mat *m, unsigned int column)
    • A new matrix is being created, m remains the same.

To remove rows:

  • nml_mat *nml_mat_row_rem(nml_mat *m, unsigned int row)
    • A new matrix is being created, m remains the same.

Example:

#include <stdlib.h>
#include <stdio.h>

#include "lib/nml.h"

int main(int argc, char *argv[]) {

    nml_mat *m = nml_mat_sqr_rnd(4, 1.0, 2.0);
    nml_mat_print(m);

    // Remove column[1] from m
    // m remains the same
    // less_columns is another matrix 
    nml_mat *less_columns = nml_mat_col_rem(m, 1);
    nml_mat_print(less_columns);

    // Remove row[0] from less_columns
    // less_columns remains the same
    // less_rows is another matrix
    nml_mat *less_rows = nml_mat_row_rem(less_columns, 0);
    nml_mat_print(less_rows);

    nml_mat_free(m);
    nml_mat_free(less_columns);
    nml_mat_free(less_rows);

    return 0;
}

To run the example:

./nml.sh examples && ./examples/remove_columns_rows.ex

Swap rows and columns

Use:

  • nml_mat *nml_mat_row_swap(nml_mat *m, unsigned int row1, unsigned int row2)
  • int nml_mat_row_swap_r(nml_mat *m, unsigned int row1, unsigned int row2)
  • nml_mat *nml_mat_col_swap(nml_mat *m, unsigned int col1, unsigned int col2)
  • int nml_mat_col_swap_r(nml_mat *m, unsigned int col1, unsigned int col2)
#include <stdlib.h>
#include <stdio.h>
#include <time.h>

#include "lib/nml.h"

int main(int argc, char *argv[]) {
  srand(time(NULL));

  nml_mat *m = nml_mat_sqr_rnd(8, 0.0, 10.0);

  printf("m=");
  nml_mat_print(m);

  printf("m= (...after swapping col1=%d with col2=%d):\n", 1, 2);
  nml_mat_col_swap_r(m, 1, 2);
  nml_mat_print(m);

  printf("newm= (...after swapping col1=%d with col2=%d and creating a new matrix):\n", 0, 1);
  nml_mat *newm = nml_mat_col_swap(m, 0, 1);
  nml_mat_print(newm);

  printf("m= (...after swapping row1=%d with row2=%d)\n", 0, 2);
  nml_mat_row_swap_r(m, 0, 2);
  nml_mat_print(m);

  nml_mat_free(m);
  nml_mat_free(newm);

  return 0;
}

Output:

m=
3.467541		8.965948		0.685298		7.802247		2.363902		0.097955		6.325767		7.162469
9.613042		6.399923		3.512556		5.528620		9.520015		2.886378		1.363232		1.838682
2.722997		5.410983		2.397803		9.870863		9.601451		1.591262		4.335792		1.660161
2.318266		3.094436		8.189400		9.242210		3.818718		1.198454		2.423213		6.939399
0.471069		7.252268		8.869578		0.994544		5.301898		9.007567		0.172999		7.586261
2.295128		4.212713		3.072580		0.849523		7.941120		6.407214		6.049472		3.469933
9.157250		5.896395		0.705688		0.491878		6.985944		2.762460		8.673238		1.114106
4.783565		7.369584		0.592299		4.774641		7.395844		1.942471		7.115440		9.204845

m= (...after swapping col1=1 with col2=2):

3.467541		0.685298		8.965948		7.802247		2.363902		0.097955		6.325767		7.162469
9.613042		3.512556		6.399923		5.528620		9.520015		2.886378		1.363232		1.838682
2.722997		2.397803		5.410983		9.870863		9.601451		1.591262		4.335792		1.660161
2.318266		8.189400		3.094436		9.242210		3.818718		1.198454		2.423213		6.939399
0.471069		8.869578		7.252268		0.994544		5.301898		9.007567		0.172999		7.586261
2.295128		3.072580		4.212713		0.849523		7.941120		6.407214		6.049472		3.469933
9.157250		0.705688		5.896395		0.491878		6.985944		2.762460		8.673238		1.114106
4.783565		0.592299		7.369584		4.774641		7.395844		1.942471		7.115440		9.204845

newm= (...after swapping col1=0 with col2=1 and creating a new matrix):

0.685298		3.467541		8.965948		7.802247		2.363902		0.097955		6.325767		7.162469
3.512556		9.613042		6.399923		5.528620		9.520015		2.886378		1.363232		1.838682
2.397803		2.722997		5.410983		9.870863		9.601451		1.591262		4.335792		1.660161
8.189400		2.318266		3.094436		9.242210		3.818718		1.198454		2.423213		6.939399
8.869578		0.471069		7.252268		0.994544		5.301898		9.007567		0.172999		7.586261
3.072580		2.295128		4.212713		0.849523		7.941120		6.407214		6.049472		3.469933
0.705688		9.157250		5.896395		0.491878		6.985944		2.762460		8.673238		1.114106
0.592299		4.783565		7.369584		4.774641		7.395844		1.942471		7.115440		9.204845

m= (...after swapping row1=0 with row2=2)

2.722997		2.397803		5.410983		9.870863		9.601451		1.591262		4.335792		1.660161
9.613042		3.512556		6.399923		5.528620		9.520015		2.886378		1.363232		1.838682
3.467541		0.685298		8.965948		7.802247		2.363902		0.097955		6.325767		7.162469
2.318266		8.189400		3.094436		9.242210		3.818718		1.198454		2.423213		6.939399
0.471069		8.869578		7.252268		0.994544		5.301898		9.007567		0.172999		7.586261
2.295128		3.072580		4.212713		0.849523		7.941120		6.407214		6.049472		3.469933
9.157250		0.705688		5.896395		0.491878		6.985944		2.762460		8.673238		1.114106
4.783565		0.592299		7.369584		4.774641		7.395844		1.942471		7.115440		9.204845

To run the example:

./nml.sh examples && ./examples/swap_rows_and_columns.ex

Concatenate matrices

Two or more matrices can be concatenated (horizontally) or (vertically) into one matrix.

To achieve this, please use:

  • nml_mat *nml_mat_cath(unsigned int mnun, nml_mat **matrices)
    • For horizontal concatenation. A new matrix is returned.
    • num represents the number of matrices to concatenate.
    • matrices the matrices to be concatenated.
  • nml_mat *nml_mat_catv(unsigned int mnum, nml_mat **matrices)
    • For vertical concatenation. A new matrix is returned.
    • num represents the number of matrices to concatenate.
    • matrices the matrices to be concatenated.

Example:

#include <stdlib.h>
#include <stdio.h>

#include "lib/nml.h"

int main(int argc, char *argv[]) {
  nml_mat *I = nml_mat_eye(3);
  nml_mat *Ix2 = nml_mat_smult(I, 2.0);
  nml_mat *rndm = nml_mat_rnd(3, 4, 1.0, 5.0);

  nml_mat **ms = malloc(sizeof(*ms) * 2);
  ms[0] = I;
  ms[1] = Ix2;
  
  nml_mat *concats1 = nml_mat_cath(2, ms);

  ms[0] = concats1;
  ms[1] = rndm;

  nml_mat *concats2 = nml_mat_cath(2, ms);

  printf("\nConcatenate horizontally\n");
  printf("I=\n");
  nml_mat_print(I);
  printf("Ix2=\n");
  nml_mat_print(Ix2);
  printf("rndm=\n");
  nml_mat_print(rndm);
  printf("concats1=\n");
  nml_mat_print(concats1);
  printf("concats2=\n");
  nml_mat_print(concats2);

  free(ms);
  nml_mat_free(I);
  nml_mat_free(Ix2);
  nml_mat_free(concats1);
  nml_mat_free(concats2);
  nml_mat_free(rndm);

  // -------------------------------------
  // Vertical concatenation
  // -------------------------------------

  nml_mat *A = nml_mat_rnd(3, 4, 1.0, 4.0);
  nml_mat *B = nml_mat_rnd(5, 4, 10.0, 20.0);
  nml_mat *C = nml_mat_eye(4);

  nml_mat **ABarr = malloc(sizeof(*ABarr) * 2);
  ABarr[0] = A;
  ABarr[1] = B;
  nml_mat *ABCat = nml_mat_catv(2, ABarr);

  printf("\nA=\n");
  nml_mat_print(A);
  printf("\nB=\n");
  nml_mat_print(B);
  printf("\nC=\n");
  nml_mat_print(C);
  printf("\nA concat B =\n");
  nml_mat_print(ABCat);

  free(ABarr);
  nml_mat_free(A);
  nml_mat_free(B);
  nml_mat_free(C);

  return 0;
}

To run the example:

./nml.sh clean examples && ./examples/concatenate_matrices.ex

Matrices operations

Add and subtract matrices

To add or subtract two matrices, the following methods can be used:

  • nml_mat *nml_mat_add(nml_mat *m1, nml_mat *m2)
    • Adds two matrices, the results are kept in a new nml_mat*. m1 and m2 remain unchanged.
  • int nml_mat_add_r(nml_mat *m1, nml_mat *m2)
    • Add two matrices, the results are kept in m1. m2 remains unchanged.
  • nml_mat *nml_mat_sub(nml_mat *m1, nml_mat *m2)
    • Subtracts two matrices, the results are kept in a new nml_mat*. m1 and m2 remain unchanged.
  • int nml_mat_sub_r(nml_mat *m1, nml_mat *m2)
    • Subtracts two matrices, the results are kept in m1. m2 remains unchanged.

Example:

#include <stdlib.h>
#include <stdio.h>

#include "lib/nml.h"

int main(int argc, char *argv[]) {
    nml_mat *m1 = nml_mat_sqr_rnd(4, 0.0, 10.0);
    nml_mat *m2 = nml_mat_sqr_rnd(4, 0.0, 10.0);
    printf("m1=\n");
    nml_mat_print(m1);
    
    printf("m2=\n");
    nml_mat_print(m2);


    // Add the matrices to, result is kept in m3
    // m1 and m2 remain unchanged
    nml_mat *m3 = nml_mat_add(m1, m2);
    printf("m3=\n");
    nml_mat_print(m3);

    // Add the matrices, the result is kept in m1
    // m1 is modified, m2 remains unchanged
    nml_mat_add_r(m1, m2);
    printf("m1=\n");
    nml_mat_print(m1);
    
    nml_mat_free(m1);
    nml_mat_free(m2);
    nml_mat_free(m3);

    return 0;
}

To run the example:

./nml.sh examples && examples/add_matrices.ex

Multiply matrices (dot)

To multiply two matrices, the following method can be used:

  • nml_mat *nml_mat_dot(nml_mat *m1, nml_mat *m2)
    • Multiplies two matrices, the result is kept in a new nml_mat*. m1 and m2 remain unchanged.

Example:

#include <stdlib.h>
#include <stdio.h>

#include "lib/nml.h"

int main(int argc, char *argv[]) {
    nml_mat *m1 = nml_mat_sqr_rnd(4, 0.0, 10.0);
    nml_mat *m2 = nml_mat_sqr_rnd(4, 0.0, 10.0);

    printf("m1=\n");
    nml_mat_print(m1);
    
    printf("m2=\n");
    nml_mat_print(m2);

    // Multiply matrices
    nml_mat *m3 = nml_mat_dot(m1, m2);
    printf("m3=\n");
    nml_mat_print(m3);

    nml_mat_free(m1);
    nml_mat_free(m2);
    nml_mat_free(m3);

    return 0;
}

To run the example:

./nml.sh examples && examples/dot_matrices.ex

Transpose matrices

To transpose a matrix, the following method can be used:

  • nml_mat *nml_mat_transp(nml_mat *m)
    • A new nml_mat* will be created, representing the transpose matrix of m. m remains unchanged.

Example:

#include <stdlib.h>
#include <stdio.h>

#include "lib/nml.h"

int main(int argc, char *argv[]) {
    nml_mat *m1 = nml_mat_rnd(1, 5, 1.0, 10.0);
    nml_mat_print(m1);

    nml_mat *m2 = nml_mat_transp(m1);
    nml_mat_print(m2);

    nml_mat_free(m1);
    nml_mat_free(m2);

    return 0;
}

To run the example:

./nml.sh clean examples && examples/transpose.ex

Calculate trace

To calculate the trace of the matrix the following method can be used: double nml_mat_trace(nml_mat* m).

Row Echelon

Calculate Row Echelon Form using Gaussian Elimination

To bring the matrix in Row Echelon Form the following method can be used: nml_mat *nml_mat_ref(nml_mat *m).

Example:

#include <stdlib.h>
#include <stdio.h>

#include "lib/nml.h"

int main(int argc, char *argv[]) {

  double v1[9] = {
    0.0, 1.0, 2.0,
    1.0, 2.0, 1.0,
    2.0, 7.0, 8.0
  };

  nml_mat *m1 = nml_mat_from(3, 3, 9, v1);
  printf("\nm1=\n");
  nml_mat_print(m1);
  nml_mat *refm1 = nml_mat_ref(m1);
  printf("\nrefm1=\n");
  nml_mat_print(refm1);

  nml_mat_free(m1);
  nml_mat_free(refm1);
  return 0;
}

Output:

m1=

0.000000		1.000000		2.000000
1.000000		2.000000		1.000000
2.000000		7.000000		8.000000


refm1=

1.000000		2.000000		1.000000
0.000000		1.000000		2.000000
0.000000		0.000000		0.000000

To run the example:

./nml.sh examples && ./examples/row_echelon.ex

Calculate Reduced Row Echelon Form using Gauss-Jordan

To bring the matrix in Reduced Row Echelon Form the following method can be used: nml_mat *nml_mat_rref(nml_mat *m).

Example:

#include <stdlib.h>
#include <stdio.h>

#include "lib/nml.h"

int main(int argc, char *argv[]) {

  double v1[9] = {
    0.0, 1.0, 2.0,
    1.0, 2.0, 1.0,
    2.0, 7.0, 8.0
  };

  nml_mat *m1 = nml_mat_from(3, 3, 9, v1);
  printf("\nm1=\n");
  nml_mat_print(m1);
  nml_mat *rrefm1 = nml_mat_rref(m1);
  printf("\nrrefm1=\n");
  nml_mat_print(rrefm1);

  nml_mat_free(m1);
  nml_mat_free(rrefm1);
  return 0;
}

Output:

m1=

0.000000		1.000000		2.000000
1.000000		2.000000		1.000000
2.000000		7.000000		8.000000


rrefm1=

1.000000		0.000000		-3.000000
-0.000000		1.000000		2.000000
0.000000		0.000000		0.000000

To run the example:

./nml.sh clean examples && examples/reduced_row_echelon.ex

LU(P) Decomposition

To decompose a matrix using LU you can use: nml_mat_lup *nml_mat_lup_solve(nml_mat *m).

The result is a pointer nml_mat_lup*:

typedef struct nml_mat_lup_s {
  nml_mat *L;
  nml_mat *U;
  nml_mat *P;
  unsigned int num_permutations;
} nml_mat_lup;

To free the nml_mat_lup*. Please use void nml_mat_lup_free(nml_mat_lup* lu). This will also deallocate the memory for the three internal nml_mat* pointers.

LU decomposition is used for solving linear systems of equations, computing the determinant and the inverse of a matrix.

Example:

#include <stdlib.h>
#include <stdio.h>

#include "lib/nml.h"

int main(int argc, char *argv[]) {
    nml_mat *m1 = nml_mat_sqr_rnd(4, 0.0, 10.0);
    printf("m1=\n");
    nml_mat_print(m1);

    nml_mat_lup *m1_lup = nml_mat_lup_solve(m1);
    printf("L, U, P:\n");
    nml_mat_lup_print(m1_lup);

    nml_mat_free(m1);
    nml_mat_lup_free(m1_lup);

    return 0;
}

Output:

m1=

0.000078		1.315378		7.556053		4.586501
5.327672		2.189592		0.470446		6.788647
6.792964		9.346929		3.835021		5.194164
8.309653		0.345721		0.534616		5.297002

L, U, P:

1.000000		0.000000		0.000000		0.000000
0.817479		1.000000		0.000000		0.000000
0.000009		0.145116		1.000000		0.000000
0.641143		0.217108		-0.086373		1.000000


8.309653		0.345721		0.534616		5.297002
0.000000		9.064309		3.397983		0.863978
0.000000		0.000000		7.062947		4.461075
0.000000		0.000000		0.000000		3.590254


0.000000		0.000000		0.000000		1.000000
0.000000		0.000000		1.000000		0.000000
1.000000		0.000000		0.000000		0.000000
0.000000		1.000000		0.000000		0.000000

Running the example:

./nml.sh clean examples && examples/lup.ex

Matrix inverse

Calculating the inverse requires to decompose the matrix LU(P) decomposition first.

Afterwards obtaining the inverse is straightforward: nml_mat *nml_mat_inv(nml_mat_lup *m).

Example:

#include <stdlib.h>
#include <stdio.h>

#include "lib/nml.h"

int main(int argc, char *argv[]) {
  printf("\nInverse of a matrix:\n");
  double m_v[16] = {
    2.0, 7.0, 6.0, 1.0,
    9.0, 5.0, 0.0, 2.0,
    4.0, 3.0, 8.0, 3.0,
    3.0, 5.0, 1.0, 9.0
  };
  nml_mat *m = nml_mat_from(4,4,16, m_v);
  nml_mat_lup *lup = nml_mat_lup_solve(m);
  nml_mat* minv = nml_mat_inv(lup);
  nml_mat *mdotminv = nml_mat_dot(m, minv);

  printf("m=");
  nml_mat_print(m);
  printf("minv=");
  nml_mat_print(minv);
  printf("(%%e) m * minv=");
  nml_mat_printf(mdotminv, "%e\t");
  printf("(%%f) m * minv=");
  nml_mat_printf(mdotminv, "%f\t");

  nml_mat_free(m);
  nml_mat_free(minv);
  nml_mat_free(mdotminv);
  return 0;
}

Output:

m=
2.000000		7.000000		6.000000		1.000000
9.000000		5.000000		0.000000		2.000000
4.000000		3.000000		8.000000		3.000000
3.000000		5.000000		1.000000		9.000000

minv=
-0.081577		0.112583		0.065924		-0.037929
0.174895		0.013245		-0.133955		0.022276
0.001505		-0.046358		0.127935		-0.032511
-0.070138		-0.039735		0.038230		0.114991

(%e) m * minv=
1.000000e+00	4.163336e-17	4.163336e-17	-1.387779e-17
0.000000e+00	1.000000e+00	-2.775558e-17	2.775558e-17
5.551115e-17	6.938894e-17	1.000000e+00	5.551115e-17
0.000000e+00	5.551115e-17	-5.551115e-17	1.000000e+00

(%f) m * minv=
1.000000	0.000000	0.000000	-0.000000
0.000000	1.000000	-0.000000	0.000000
0.000000	0.000000	1.000000	0.000000
0.000000	0.000000	-0.000000	1.000000

To run the example:

./nml.sh clean examples && ./examples/inverse.ex

Matrix determinant

Calculating the determinant requires to decompose the matrix LU(P) decomposition first.

Afterwards obtaining the determinant is straightforward: double nml_mat_det(nml_mat_lup* lup).

#include <stdlib.h>
#include <stdio.h>

#include "lib/nml.h"

int main(int argc, char *argv[]) {
    nml_mat *m1 = nml_mat_sqr_rnd(4, 0.0, 10.0);
    nml_mat_lup *m1_lup = nml_mat_lup_solve(m1);
    
    printf("m1=\n");
    nml_mat_print(m1);
    printf("determinant=%lf\n", nml_mat_det(m1_lup));

    nml_mat_free(m1);
    nml_mat_lup_free(m1_lup);

    return 0;
}

Output:

m1=

0.000078		1.315378		7.556053		4.586501
5.327672		2.189592		0.470446		6.788647
6.792964		9.346929		3.835021		5.194164
8.309653		0.345721		0.534616		5.297002

determinant=-1909.979877

Runnning the example:

./nml.sh clean examples && examples/determinant.ex

Solve linear systems of equations

Solving A * x = B where A is lower triangular (Forward Substitution)

Use: nml_mat *nml_ls_solvefwd(nml_mat *low_triang, nml_mat *b).

Note: no validation will be performed to check is low_triang is a lower triangular matrix

#include <stdlib.h>
#include <stdio.h>

#include "lib/nml.h"

int main(int argc, char *argv[]) {
    
    FILE *input = fopen("./examples/data/matrix9_lower_triangular.data", "r");

    nml_mat *A = nml_mat_fromfilef(input);
    nml_mat *B = nml_mat_fromfilef(input);
    nml_mat *x = nml_ls_solvefwd(A, B);

    nml_mat_print(A);
    nml_mat_print(B);
    nml_mat_print(x);

    nml_mat_free(A);
    nml_mat_free(B);
    nml_mat_free(x);

    fclose(input);

    return 0;
}

To run the example:

./nml.sh clean examples && examples/forward_substition.ex

Solving A * x = B where A is upper traingular (Backward Substition)

Use: nml_mat *nml_ls_solvebck(nml_mat *upper_triang, nml_mat *b).

Note: no validation will be performed to check is upper_triang is an upper triangular matrix

#include <stdlib.h>
#include <stdio.h>

#include "lib/nml.h"

int main(int argc, char *argv[]) {
    
    FILE *input = fopen("./examples/data/matrix10_upper_triangular.data", "r");

    nml_mat *A = nml_mat_fromfilef(input);
    nml_mat *B = nml_mat_fromfilef(input);
    nml_mat *x = nml_ls_solvebck(A, B);

    nml_mat_print(A);
    nml_mat_print(B);
    nml_mat_print(x);

    nml_mat_free(A);
    nml_mat_free(B);
    nml_mat_free(x);

    fclose(input);

    return 0;
}

To run the example:

./nml.sh clean examples && examples/backward_substitution.ex

Solving A * x = B using LU(P) decomposition

Use: nml_mat *nml_ls_solve(nml_mat_lup *lup, nml_mat* b).

#include <stdlib.h>
#include <stdio.h>

#include "lib/nml.h"

int main(int argc, char *argv[]) {
    nml_mat *A = nml_mat_sqr_rnd(4, 1.0, 10.0);
    nml_mat *B = nml_mat_rnd(4, 1, 1.0, 10.0);
    nml_mat_lup *LUP = nml_mat_lup_solve(A);

    nml_mat *x = nml_ls_solve(LUP, B);
    nml_mat_print(x);

    nml_mat_free(A);
    nml_mat_free(B);
    nml_mat_free(x);
    nml_mat_lup_free(LUP);
}

To run the example:

./nml.sh clean examples && ./examples/ls_solve.ex
You might also like...
Kraken is an open-source modern math library that comes with a fast-fixed matrix class and math-related functions.

Kraken 🐙 Table of Contents Introduction Requirement Contents Installation Introduction Kraken is a modern math library written in a way that gives ac

yahya mohammed 24 Nov 30, 2022
A modern, C++20-native, single-file header-only dense 2D matrix library.
A modern, C++20-native, single-file header-only dense 2D matrix library.

A modern, C++20-native, single-file header-only dense 2D matrix library. Contents Example usage creating matrices basic operations row, col, size, sha

feng wang 62 Dec 17, 2022
Minimal matrix implementation in C++

min-matrix Minimal matrix implementation in C++. 🤗 Pull Request 🤗 😎 Post Issues 😎 Brief 😄 Eigen compiles too slow? 😄 😅 Just want something simp

Shepard 5 Dec 17, 2021
Some out-of-core matrix operations via HDF5.

hdfmat Version: 0.2-2 License: BSD 2-Clause Author: Drew Schmidt This package provides some out-of-core matrix operations via HDF5. The main purpose o

Drew Schmidt 5 Jan 9, 2022
A Code Base for Matrix operations in C++

SimpM A Code Base for Matrix operations in C++ Dependencies: GNU Bignum Library: https://gmplib.org Needed to be installed on you computer. Check your

null 3 Oct 27, 2022
MIRACL Cryptographic SDK: Multiprecision Integer and Rational Arithmetic Cryptographic Library is a C software library that is widely regarded by developers as the gold standard open source SDK for elliptic curve cryptography (ECC).

MIRACL What is MIRACL? Multiprecision Integer and Rational Arithmetic Cryptographic Library – the MIRACL Crypto SDK – is a C software library that is

MIRACL 527 Jan 7, 2023
LibTomMath is a free open source portable number theoretic multiple-precision integer library written entirely in C.

libtommath This is the git repository for LibTomMath, a free open source portable number theoretic multiple-precision integer (MPI) library written en

libtom 543 Dec 27, 2022
SymEngine is a fast symbolic manipulation library, written in C++

SymEngine SymEngine is a standalone fast C++ symbolic manipulation library. Optional thin wrappers allow usage of the library from other languages, e.

null 926 Dec 24, 2022
Header only, single file, simple and efficient C++ library to compute the signed distance function to a triangle mesh

TriangleMeshDistance Header only, single file, simple and efficient C++11 library to compute the signed distance function to a triangle mesh. The dist

Interactive Computer Graphics 100 Dec 28, 2022
Comments
  • đŸŠŗ🧙 fix nml_mat_col_swap_r validation check

    đŸŠŗ🧙 fix nml_mat_col_swap_r validation check

    I changed the validation check inside the columns swap function. It was comparing col2 against the number of rows of the matrix, when i believe it should be comparing against the number of columns.

    P.S. Loved the tutorial you made about this library :)

    opened by arthursfares 1
  • Fix for adding non-square matrices

    Fix for adding non-square matrices

    Fixed a typo where it was iterating over (m1->rows and then m2->rows) instead of (m1->rows and then m2->cols)

    It doesn't matter when adding square matrices like in the add_matrices.c example but breaks with non-square matrices

    opened by bhabas 1
  • Fix double abs in equality check

    Fix double abs in equality check

    Absolute values are unnecessary for each entry, and in fact can lead to errors for numbers with similar value but opposite sign. For example, (1) and (-1) would be considered equal in the original code.

    opened by nicknytko 1
Owner
Andrei Ciobanu
Neat code
Andrei Ciobanu
GitHub
A simple C++ complex & real matrix library, with matrix inversion, left division and determinant calculation

NaiveMatrixLib 帆帆įš„įŽ€æ˜“įŸŠé˜ĩ莥įŽ—åē“ A simple C++ stdlib-based complex & real matrix library, with matrix inversion, left division (A\b) and determinant calculat

FerryYoungFan 50 Dec 28, 2022
libmpc++ is a C++ header-only library to solve linear and non-linear MPC

libmpc++ libmpc++ is a C++ library to solve linear and non-linear MPC. The library is written in modern C++17 and it is tested to work on Linux, macOS

Nicola Piccinelli 46 Dec 20, 2022
C++ Matrix -- High performance and accurate (e.g. edge cases) matrix math library with expression template arithmetic operators

Matrix This is a math and arithmetic matrix library. It has stood many years of performing in mission critical production for financial systems. It ha

Hossein Moein 71 Oct 29, 2022
Seidel's Algorithm: Linear-Complexity Linear Programming for Small-Dimensional Variables

SDLP Seidel's Algorithm: Linear-Complexity Linear Programming (LP) for Small-Dimensions About This solver is super efficient for small-dimensional LP

ZJU FAST Lab 43 Dec 19, 2022
A matrix header-only library, uses graphs internally, helpful when your matrix is part of a simulation where it needs to grow many times (or auto expand)

GraphMat Header-only Library Matrix implemented as a graph, specially for the use case when it should be auto expanding at custom rate, specially in s

Aditya Gupta 3 Oct 25, 2021
P(R*_{3, 0, 1}) specialized SIMD Geometric Algebra Library

Klein ?? ?? Project Site ?? ?? Description Do you need to do any of the following? Quickly? Really quickly even? Projecting points onto lines, lines t

Jeremy Ong 635 Dec 30, 2022
a lean linear math library, aimed at graphics programming. Supports vec3, vec4, mat4x4 and quaternions

linmath.h -- A small library for linear math as required for computer graphics linmath.h provides the most used types required for programming compute

datenwolf 729 Jan 9, 2023
C++ library for solving large sparse linear systems with algebraic multigrid method

AMGCL AMGCL is a header-only C++ library for solving large sparse linear systems with algebraic multigrid (AMG) method. AMG is one of the most effecti

Denis Demidov 578 Dec 11, 2022
A large scale non-linear optimization library

Ceres Solver Ceres Solver is an open source C++ library for modeling and solving large, complicated optimization problems. It is a feature rich, matur

null 2.9k Jan 3, 2023
Library for nonconvex constrained optimization using the augmented Lagrangian method and the matrix-free PANOC algorithm.

alpaqa Alpaqa is an efficient implementation of the Augmented Lagrangian method for general nonlinear programming problems, which uses the first-order

OPTEC 21 Dec 9, 2022