460 lines
9.7 KiB
C
460 lines
9.7 KiB
C
|
/*
|
||
|
* Copyright (c), Recep Aslantas.
|
||
|
*
|
||
|
* MIT License (MIT), http://opensource.org/licenses/MIT
|
||
|
* Full license can be found in the LICENSE file
|
||
|
*/
|
||
|
|
||
|
/*!
|
||
|
* Most of functions in this header are optimized manually with SIMD
|
||
|
* if available. You dont need to call/incude SIMD headers manually
|
||
|
*/
|
||
|
|
||
|
/*
|
||
|
Macros:
|
||
|
GLMS_MAT4_IDENTITY_INIT
|
||
|
GLMS_MAT4_ZERO_INIT
|
||
|
GLMS_MAT4_IDENTITY
|
||
|
GLMS_MAT4_ZERO
|
||
|
|
||
|
Functions:
|
||
|
CGLM_INLINE mat4s glms_mat4_ucopy(mat4s mat);
|
||
|
CGLM_INLINE mat4s glms_mat4_copy(mat4s mat);
|
||
|
CGLM_INLINE mat4s glms_mat4_identity(void);
|
||
|
CGLM_INLINE void glms_mat4_identity_array(mat4s * __restrict mat, size_t count);
|
||
|
CGLM_INLINE mat4s glms_mat4_zero(void);
|
||
|
CGLM_INLINE mat3s glms_mat4_pick3(mat4s mat);
|
||
|
CGLM_INLINE mat3s glms_mat4_pick3t(mat4s mat);
|
||
|
CGLM_INLINE mat4s glms_mat4_ins3(mat3s mat);
|
||
|
CGLM_INLINE mat4s glms_mat4_mul(mat4s m1, mat4s m2);
|
||
|
CGLM_INLINE mat4s glms_mat4_mulN(mat4s * __restrict matrices[], uint32_t len);
|
||
|
CGLM_INLINE vec4s glms_mat4_mulv(mat4s m, vec4s v);
|
||
|
CGLM_INLINE float glms_mat4_trace(mat4s m);
|
||
|
CGLM_INLINE float glms_mat4_trace3(mat4s m);
|
||
|
CGLM_INLINE versors glms_mat4_quat(mat4s m);
|
||
|
CGLM_INLINE vec3s glms_mat4_mulv3(mat4s m, vec3s v, float last);
|
||
|
CGLM_INLINE mat4s glms_mat4_transpose(mat4s m);
|
||
|
CGLM_INLINE mat4s glms_mat4_scale_p(mat4s m, float s);
|
||
|
CGLM_INLINE mat4s glms_mat4_scale(mat4s m, float s);
|
||
|
CGLM_INLINE float glms_mat4_det(mat4s mat);
|
||
|
CGLM_INLINE mat4s glms_mat4_inv(mat4s mat);
|
||
|
CGLM_INLINE mat4s glms_mat4_inv_fast(mat4s mat);
|
||
|
CGLM_INLINE mat4s glms_mat4_swap_col(mat4s mat, int col1, int col2);
|
||
|
CGLM_INLINE mat4s glms_mat4_swap_row(mat4s mat, int row1, int row2);
|
||
|
CGLM_INLINE float glms_mat4_rmc(vec4s r, mat4s m, vec4s c);
|
||
|
*/
|
||
|
|
||
|
#ifndef cglms_mat4s_h
|
||
|
#define cglms_mat4s_h
|
||
|
|
||
|
#include "../common.h"
|
||
|
#include "../types-struct.h"
|
||
|
#include "../mat4.h"
|
||
|
#include "vec4.h"
|
||
|
#include "vec3.h"
|
||
|
|
||
|
#define GLMS_MAT4_IDENTITY_INIT {GLM_MAT4_IDENTITY_INIT}
|
||
|
#define GLMS_MAT4_ZERO_INIT {GLM_MAT4_ZERO_INIT}
|
||
|
|
||
|
/* for C only */
|
||
|
#define GLMS_MAT4_IDENTITY ((mat4s)GLMS_MAT4_IDENTITY_INIT)
|
||
|
#define GLMS_MAT4_ZERO ((mat4s)GLMS_MAT4_ZERO_INIT)
|
||
|
|
||
|
/*!
|
||
|
* @brief copy all members of [mat] to [dest]
|
||
|
*
|
||
|
* matrix may not be aligned, u stands for unaligned, this may be useful when
|
||
|
* copying a matrix from external source e.g. asset importer...
|
||
|
*
|
||
|
* @param[in] mat source
|
||
|
* @returns destination
|
||
|
*/
|
||
|
CGLM_INLINE
|
||
|
mat4s
|
||
|
glms_mat4_ucopy(mat4s mat) {
|
||
|
mat4s r;
|
||
|
glm_mat4_ucopy(mat.raw, r.raw);
|
||
|
return r;
|
||
|
}
|
||
|
|
||
|
/*!
|
||
|
* @brief copy all members of [mat] to [dest]
|
||
|
*
|
||
|
* @param[in] mat source
|
||
|
* @returns destination
|
||
|
*/
|
||
|
CGLM_INLINE
|
||
|
mat4s
|
||
|
glms_mat4_copy(mat4s mat) {
|
||
|
mat4s r;
|
||
|
glm_mat4_copy(mat.raw, r.raw);
|
||
|
return r;
|
||
|
}
|
||
|
|
||
|
/*!
|
||
|
* @brief make given matrix identity. It is identical with below,
|
||
|
* but it is more easy to do that with this func especially for members
|
||
|
* e.g. glm_mat4_identity(aStruct->aMatrix);
|
||
|
*
|
||
|
* @code
|
||
|
* glm_mat4_copy(GLM_MAT4_IDENTITY, mat); // C only
|
||
|
*
|
||
|
* // or
|
||
|
* mat4 mat = GLM_MAT4_IDENTITY_INIT;
|
||
|
* @endcode
|
||
|
*
|
||
|
* @retuns destination
|
||
|
*/
|
||
|
CGLM_INLINE
|
||
|
mat4s
|
||
|
glms_mat4_identity(void) {
|
||
|
mat4s r;
|
||
|
glm_mat4_identity(r.raw);
|
||
|
return r;
|
||
|
}
|
||
|
|
||
|
/*!
|
||
|
* @brief make given matrix array's each element identity matrix
|
||
|
*
|
||
|
* @param[in, out] mat matrix array (must be aligned (16/32)
|
||
|
* if alignment is not disabled)
|
||
|
*
|
||
|
* @param[in] count count of matrices
|
||
|
*/
|
||
|
CGLM_INLINE
|
||
|
void
|
||
|
glms_mat4_identity_array(mat4s * __restrict mat, size_t count) {
|
||
|
CGLM_ALIGN_MAT mat4s t = GLMS_MAT4_IDENTITY_INIT;
|
||
|
size_t i;
|
||
|
|
||
|
for (i = 0; i < count; i++) {
|
||
|
glm_mat4_copy(t.raw, mat[i].raw);
|
||
|
}
|
||
|
}
|
||
|
|
||
|
/*!
|
||
|
* @brief make given matrix zero.
|
||
|
*
|
||
|
* @returns matrix
|
||
|
*/
|
||
|
CGLM_INLINE
|
||
|
mat4s
|
||
|
glms_mat4_zero(void) {
|
||
|
mat4s r;
|
||
|
glm_mat4_zero(r.raw);
|
||
|
return r;
|
||
|
}
|
||
|
|
||
|
/*!
|
||
|
* @brief copy upper-left of mat4 to mat3
|
||
|
*
|
||
|
* @param[in] mat source
|
||
|
* @returns destination
|
||
|
*/
|
||
|
CGLM_INLINE
|
||
|
mat3s
|
||
|
glms_mat4_pick3(mat4s mat) {
|
||
|
mat3s r;
|
||
|
glm_mat4_pick3(mat.raw, r.raw);
|
||
|
return r;
|
||
|
}
|
||
|
|
||
|
/*!
|
||
|
* @brief copy upper-left of mat4 to mat3 (transposed)
|
||
|
*
|
||
|
* the postfix t stands for transpose
|
||
|
*
|
||
|
* @param[in] mat source
|
||
|
* @returns destination
|
||
|
*/
|
||
|
CGLM_INLINE
|
||
|
mat3s
|
||
|
glms_mat4_pick3t(mat4s mat) {
|
||
|
mat3s r;
|
||
|
glm_mat4_pick3t(mat.raw, r.raw);
|
||
|
return r;
|
||
|
}
|
||
|
|
||
|
/*!
|
||
|
* @brief copy mat3 to mat4's upper-left
|
||
|
*
|
||
|
* @param[in] mat source
|
||
|
* @returns destination
|
||
|
*/
|
||
|
CGLM_INLINE
|
||
|
mat4s
|
||
|
glms_mat4_ins3(mat3s mat) {
|
||
|
mat4s r;
|
||
|
glm_mat4_ins3(mat.raw, r.raw);
|
||
|
return r;
|
||
|
}
|
||
|
|
||
|
/*!
|
||
|
* @brief multiply m1 and m2 to dest
|
||
|
*
|
||
|
* m1, m2 and dest matrices can be same matrix, it is possible to write this:
|
||
|
*
|
||
|
* @code
|
||
|
* mat4 m = GLM_MAT4_IDENTITY_INIT;
|
||
|
* glm_mat4_mul(m, m, m);
|
||
|
* @endcode
|
||
|
*
|
||
|
* @param[in] m1 left matrix
|
||
|
* @param[in] m2 right matrix
|
||
|
* @returns destination matrix
|
||
|
*/
|
||
|
CGLM_INLINE
|
||
|
mat4s
|
||
|
glms_mat4_mul(mat4s m1, mat4s m2) {
|
||
|
mat4s r;
|
||
|
glm_mat4_mul(m1.raw, m2.raw, r.raw);
|
||
|
return r;
|
||
|
}
|
||
|
|
||
|
/*!
|
||
|
* @brief mupliply N mat4 matrices and store result in dest
|
||
|
*
|
||
|
* this function lets you multiply multiple (more than two or more...) matrices
|
||
|
* <br><br>multiplication will be done in loop, this may reduce instructions
|
||
|
* size but if <b>len</b> is too small then compiler may unroll whole loop,
|
||
|
* usage:
|
||
|
* @code
|
||
|
* mat m1, m2, m3, m4, res;
|
||
|
*
|
||
|
* res = glm_mat4_mulN((mat4 *[]){&m1, &m2, &m3, &m4}, 4);
|
||
|
* @endcode
|
||
|
*
|
||
|
* @warning matrices parameter is pointer array not mat4 array!
|
||
|
*
|
||
|
* @param[in] matrices mat4 * array
|
||
|
* @param[in] len matrices count
|
||
|
* @returns result matrix
|
||
|
*/
|
||
|
CGLM_INLINE
|
||
|
mat4s
|
||
|
glms_mat4_mulN(mat4s * __restrict matrices[], uint32_t len) {
|
||
|
CGLM_ALIGN_MAT mat4s r = GLMS_MAT4_IDENTITY_INIT;
|
||
|
size_t i;
|
||
|
|
||
|
for (i = 0; i < len; i++) {
|
||
|
r = glms_mat4_mul(r, *matrices[i]);
|
||
|
}
|
||
|
|
||
|
return r;
|
||
|
}
|
||
|
|
||
|
/*!
|
||
|
* @brief multiply mat4 with vec4 (column vector) and store in dest vector
|
||
|
*
|
||
|
* @param[in] m mat4 (left)
|
||
|
* @param[in] v vec4 (right, column vector)
|
||
|
* @returns vec4 (result, column vector)
|
||
|
*/
|
||
|
CGLM_INLINE
|
||
|
vec4s
|
||
|
glms_mat4_mulv(mat4s m, vec4s v) {
|
||
|
vec4s r;
|
||
|
glm_mat4_mulv(m.raw, v.raw, r.raw);
|
||
|
return r;
|
||
|
}
|
||
|
|
||
|
/*!
|
||
|
* @brief trace of matrix
|
||
|
*
|
||
|
* sum of the elements on the main diagonal from upper left to the lower right
|
||
|
*
|
||
|
* @param[in] m matrix
|
||
|
*/
|
||
|
CGLM_INLINE
|
||
|
float
|
||
|
glms_mat4_trace(mat4s m) {
|
||
|
return glm_mat4_trace(m.raw);
|
||
|
}
|
||
|
|
||
|
/*!
|
||
|
* @brief trace of matrix (rotation part)
|
||
|
*
|
||
|
* sum of the elements on the main diagonal from upper left to the lower right
|
||
|
*
|
||
|
* @param[in] m matrix
|
||
|
*/
|
||
|
CGLM_INLINE
|
||
|
float
|
||
|
glms_mat4_trace3(mat4s m) {
|
||
|
return glm_mat4_trace3(m.raw);
|
||
|
}
|
||
|
|
||
|
/*!
|
||
|
* @brief convert mat4's rotation part to quaternion
|
||
|
*
|
||
|
* @param[in] m affine matrix
|
||
|
* @returns destination quaternion
|
||
|
*/
|
||
|
CGLM_INLINE
|
||
|
versors
|
||
|
glms_mat4_quat(mat4s m) {
|
||
|
versors r;
|
||
|
glm_mat4_quat(m.raw, r.raw);
|
||
|
return r;
|
||
|
}
|
||
|
|
||
|
/*!
|
||
|
* @brief multiply vector with mat4
|
||
|
*
|
||
|
* @param[in] m mat4(affine transform)
|
||
|
* @param[in] v vec3
|
||
|
* @param[in] last 4th item to make it vec4
|
||
|
* @returns result vector (vec3)
|
||
|
*/
|
||
|
CGLM_INLINE
|
||
|
vec3s
|
||
|
glms_mat4_mulv3(mat4s m, vec3s v, float last) {
|
||
|
vec3s r;
|
||
|
glm_mat4_mulv3(m.raw, v.raw, last, r.raw);
|
||
|
return r;
|
||
|
}
|
||
|
|
||
|
/*!
|
||
|
* @brief tranpose mat4 and store result in same matrix
|
||
|
*
|
||
|
* @param[in] m source
|
||
|
* @returns result
|
||
|
*/
|
||
|
CGLM_INLINE
|
||
|
mat4s
|
||
|
glms_mat4_transpose(mat4s m) {
|
||
|
glm_mat4_transpose(m.raw);
|
||
|
return m;
|
||
|
}
|
||
|
|
||
|
/*!
|
||
|
* @brief scale (multiply with scalar) matrix without simd optimization
|
||
|
*
|
||
|
* multiply matrix with scalar
|
||
|
*
|
||
|
* @param[in] m matrix
|
||
|
* @param[in] s scalar
|
||
|
* @returns matrix
|
||
|
*/
|
||
|
CGLM_INLINE
|
||
|
mat4s
|
||
|
glms_mat4_scale_p(mat4s m, float s) {
|
||
|
glm_mat4_scale_p(m.raw, s);
|
||
|
return m;
|
||
|
}
|
||
|
|
||
|
/*!
|
||
|
* @brief scale (multiply with scalar) matrix
|
||
|
*
|
||
|
* multiply matrix with scalar
|
||
|
*
|
||
|
* @param[in] m matrix
|
||
|
* @param[in] s scalar
|
||
|
* @returns matrix
|
||
|
*/
|
||
|
CGLM_INLINE
|
||
|
mat4s
|
||
|
glms_mat4_scale(mat4s m, float s) {
|
||
|
glm_mat4_scale(m.raw, s);
|
||
|
return m;
|
||
|
}
|
||
|
|
||
|
/*!
|
||
|
* @brief mat4 determinant
|
||
|
*
|
||
|
* @param[in] mat matrix
|
||
|
*
|
||
|
* @return determinant
|
||
|
*/
|
||
|
CGLM_INLINE
|
||
|
float
|
||
|
glms_mat4_det(mat4s mat) {
|
||
|
return glm_mat4_det(mat.raw);
|
||
|
}
|
||
|
|
||
|
/*!
|
||
|
* @brief inverse mat4 and store in dest
|
||
|
*
|
||
|
* @param[in] mat matrix
|
||
|
* @returns inverse matrix
|
||
|
*/
|
||
|
CGLM_INLINE
|
||
|
mat4s
|
||
|
glms_mat4_inv(mat4s mat) {
|
||
|
mat4s r;
|
||
|
glm_mat4_inv(mat.raw, r.raw);
|
||
|
return r;
|
||
|
}
|
||
|
|
||
|
/*!
|
||
|
* @brief inverse mat4 and store in dest
|
||
|
*
|
||
|
* this func uses reciprocal approximation without extra corrections
|
||
|
* e.g Newton-Raphson. this should work faster than normal,
|
||
|
* to get more precise use glm_mat4_inv version.
|
||
|
*
|
||
|
* NOTE: You will lose precision, glm_mat4_inv is more accurate
|
||
|
*
|
||
|
* @param[in] mat matrix
|
||
|
* @returns inverse matrix
|
||
|
*/
|
||
|
CGLM_INLINE
|
||
|
mat4s
|
||
|
glms_mat4_inv_fast(mat4s mat) {
|
||
|
mat4s r;
|
||
|
glm_mat4_inv_fast(mat.raw, r.raw);
|
||
|
return r;
|
||
|
}
|
||
|
|
||
|
/*!
|
||
|
* @brief swap two matrix columns
|
||
|
*
|
||
|
* @param[in] mat matrix
|
||
|
* @param[in] col1 col1
|
||
|
* @param[in] col2 col2
|
||
|
* @returns matrix
|
||
|
*/
|
||
|
CGLM_INLINE
|
||
|
mat4s
|
||
|
glms_mat4_swap_col(mat4s mat, int col1, int col2) {
|
||
|
glm_mat4_swap_col(mat.raw, col1, col2);
|
||
|
return mat;
|
||
|
}
|
||
|
|
||
|
/*!
|
||
|
* @brief swap two matrix rows
|
||
|
*
|
||
|
* @param[in] mat matrix
|
||
|
* @param[in] row1 row1
|
||
|
* @param[in] row2 row2
|
||
|
* @returns matrix
|
||
|
*/
|
||
|
CGLM_INLINE
|
||
|
mat4s
|
||
|
glms_mat4_swap_row(mat4s mat, int row1, int row2) {
|
||
|
glm_mat4_swap_row(mat.raw, row1, row2);
|
||
|
return mat;
|
||
|
}
|
||
|
|
||
|
/*!
|
||
|
* @brief helper for R (row vector) * M (matrix) * C (column vector)
|
||
|
*
|
||
|
* rmc stands for Row * Matrix * Column
|
||
|
*
|
||
|
* the result is scalar because R * M = Matrix1x4 (row vector),
|
||
|
* then Matrix1x4 * Vec4 (column vector) = Matrix1x1 (Scalar)
|
||
|
*
|
||
|
* @param[in] r row vector or matrix1x4
|
||
|
* @param[in] m matrix4x4
|
||
|
* @param[in] c column vector or matrix4x1
|
||
|
*
|
||
|
* @return scalar value e.g. B(s)
|
||
|
*/
|
||
|
CGLM_INLINE
|
||
|
float
|
||
|
glms_mat4_rmc(vec4s r, mat4s m, vec4s c) {
|
||
|
return glm_mat4_rmc(r.raw, m.raw, c.raw);
|
||
|
}
|
||
|
|
||
|
#endif /* cglms_mat4s_h */
|