tensor.

header only tensor math library for c. optimized for intel haswell and amd broadwell (2013+ processors) with sse & avx instruction extensions. (except for avx512, since my cpu doesn't support it.)

the affine transformations such as m4x4_make_rotate, m4x4_make_translate, m4x4_make_scale, m4x4_make_perspective are written with opengl in mind i.e. column-major order so remember that! you'll have to transpose the matrix if you're using row-major order (like in vulkan, directx etc.)

usage.

• open tensor.h file and find this line #define TENSOR_IMPLEMENTATION. comment the line above this line says "COMMENT ME TO USE THE HEADER FILE"
• then include the header file in your project. below is an example.

#define TENSOR_IMPLEMENTATION
#include "./include/tensor.h" // or whereever you put it.

int main() {
    // example usage.
    vec3 a = {1.0f, 2.0f, 3.0f};
    vec3 b = {4.0f, 5.0f, 6.0f};
    vec3 r;
    vec3_add(&r, &a, &b);
    print_vec3(&r);
    return 0;
}

• compile and run the program. (include the math lib -lm and -march=native for simd intrinsics)

$ gcc -I./include -o main.o -c main.c -lm -march=native
$ gcc -I./include -o main main.o -lm -march=native
$ ./main

test.

all functions are tested. to run the tests, do the following:

asumming you've already cloned this repo.

$ make clean & make && ./test && make clean

thoughts.

this is an extremely beta lib. i'm always updating it and adding new functions as i need them. if you find any bugs or have any suggestions, feel free to open an issue or a pull request. i'm always open to feedback.

the largest function that i couldn't optimize is m4x4_inverse. the function is a pain in the ass. so i'll prolly optimize it but when i run into perf issues where the function is called extensively.

license.

MIT