Geek Answers Handbook

C ++ Matrix Multiplication - understanding the logic of a method optimized for it

I recently read the Matrix Tutorials with openGL and came across an optimized Matrix Multiplication method that I cannot understand.

//Create an allias type for a Matrix Type      
typedef struct Matrix
{
    float m[16];
} Matrix;

//default matrix
static const Matrix IDENTITY_MATRIX = { {
    1, 0, 0, 0,
        0, 1, 0, 0,
        0, 0, 1, 0,
        0, 0, 0, 1
} };

Matrix MultiplyMatrices(const Matrix* m1, const Matrix* m2)
{
    Matrix out = IDENTITY_MATRIX;
    unsigned int row, column, row_offset;

    for (row = 0, row_offset = row * 4; row < 4; ++row, row_offset = row * 4)
        for (column = 0; column < 4; ++column)
            out.m[row_offset + column] =
            (m1->m[row_offset + 0] * m2->m[column + 0]) +
            (m1->m[row_offset + 1] * m2->m[column + 4]) +
            (m1->m[row_offset + 2] * m2->m[column + 8]) +
            (m1->m[row_offset + 3] * m2->m[column + 12]);

    return out;
}

Here are the questions I have:

  • In the MultiplyMatrices method, why is there a pointer to the m1 and m2 parameters? If you just copy your values ​​and return a new matrix, why use a pointer?

  • Why is the condition of the for loop identical to its increment?

for (row = 0, row_offset = row * 4 ; row <4; ++ row, row_offset = row * 4 )

+4
source share
2 answers

MultiplyMatrices . , . ,

typedef struct Matrix
{
    float m[16];
} Matrix;

4 4 1-D . , 4 . 1- . , .

, for:

for (row = 0, row_offset = row * 4; row < 4; ++row, row_offset = row * 4)

row_offset 0. , row_offset row. , 1-D 2-D- a [i] [j] :

a[i][j] = a[i*num_col+j]

num_col 4. , . - . - reset row_offset, 1.

+2

MultiplyMatrices m1 m2? , ?

, , -? out, m1 m2. .

for ?

- ++row row . "" row < 4, - , .

0

More articles:


All Articles