#pragma once

#include "GLenv/include/glad/glad.h"
#include "StgSharpGraphic.h"

#include <stdio.h>
#include <stdalign.h>

#include <xmmintrin.h>
#include <pmmintrin.h>
#include <immintrin.h>



/*

Basic functions as a copy of SIMD intrisic function

*/
SSGC_API __m128* __cdecl load(float x, float y, float z, float w)
{
    float array[4] = {x,y,z,w};
    __m128* ptr = _aligned_malloc(sizeof(__m128),16);

    if (NULL != ptr)
    {
        *ptr = _mm_loadu_ps(&array);
        return ptr;
    }
    else
    {
        return -1;
    }
}

SSGC_API __m128* __cdecl load_cpy(__m128* ptr)
{
    __m128* ret = _aligned_malloc(sizeof(__m128*),16);

    if (NULL != ret)
    {
        *ret = *ptr;
    }

    return ret;
}

SSGC_API __m128* __cdecl addps(__m128* ptr1, __m128* ptr2)
{
    
    __m128* ret = _aligned_malloc(sizeof(__m128*),16);

    if (NULL != ret)
    {
        *ret = _mm_add_ps(*ptr1, *ptr2);
        return ret;
    }
    else
    {
        return (__m128*) - 1;
    }

}

SSGC_API __m128* __cdecl subps(__m128* ptr1, __m128* ptr2)
{
    __m128* ret = _aligned_malloc(sizeof(__m128*), 16);

    if (NULL != ret)
    {
        *ret = _mm_sub_ps(*ptr1, *ptr2);
        return ret;
    }
    else
    {
        return -1;
    }
}

SSGC_API __m128* __cdecl mulps_const(__m128* ptr, float value)
{
    float c[4] = { value, value, value, value };
    __m128* ret = _aligned_malloc(sizeof(__m128),16);

    __m128 valueVec = _mm_loadu_ps(&c);

    if (NULL != ret)
    {
        *ret = _mm_mul_ps(*ptr,valueVec);
        return ret;
    }
    else
    {
        return -1;
    }
}

SSGC_API __m128* __cdecl mulps(__m128* ptr1, __m128* ptr2)
{
    __m128* ret = _aligned_malloc(sizeof(__m128), 16);

    if (NULL != ret)
    {
        *ret = _mm_mul_ps(*ptr1, *ptr2);
        return ret;
    }
    else
    {
        return -1;
    }
}

SSGC_API _inline float __cdecl mulps_sum(__m128* ptr1, __m128* ptr2)
{
    __m128 ret = _mm_mul_ps(*ptr1,*ptr2);
    ret = _mm_hadd_ps(ret,ret);
    ret = _mm_hadd_ps(ret,ret);
    return *(float*)&ret;
}

_inline float __cdecl mulps_sum_internal(__m128 left, __m128 right)
{
    __m128 ret = _mm_mul_ps(left, right);
    ret = _mm_hadd_ps(ret, ret);
    ret = _mm_hadd_ps(ret, ret);
    return *(float*)&ret;
}

_inline float __cdecl mulps_total_internal(__m128 vec)
{
    
}

SSGC_API __m128* __cdecl divps_const(__m128* ptr, float value)
{

    float c[4] = { value, value, value, value };
    __m128* ret = _aligned_malloc(sizeof(__m128), 16);

    __m128 valueVec = _mm_loadu_ps(&c);

    if (NULL != ret)
    {
        *ret = _mm_div_ps(*ptr, valueVec);
        return ret;
    }
    else
    {
        return -1;
    }
}

SSGC_API __m128* __cdecl divps(__m128* ptr1, __m128* ptr2)
{
    __m128* ret = _aligned_malloc(sizeof(__m128), 16);

    if (NULL != ret)
    {
        *ret = _mm_div_ps(*ptr1, *ptr2);
        return ret;
    }
    else
    {
        return -1;
    }
}

SSGC_API float __cdecl sqrtss_const(float a)
{
    __m128 ret = _mm_sqrt_ss(*(__m128*) & a);

    return ret.m128_f32[0];
}



/*
Some complex functions for matrix calculations.
*/

SSGC_API __m256i* __cdecl load_matmap(uint64_t c0, uint64_t c1, uint64_t c2, uint64_t c3)
{
    __m256i* retPtr = _aligned_malloc(sizeof(__m256i), sizeof(__m256i));

    uint64_t ptr[4] = {c0,c1,c2,c3 };

    if (NULL != retPtr)
    {
        *retPtr = _mm256_loadu_epi64(&ptr);
    }

    return retPtr;

}



SSGC_API __m128* __cdecl set_m128ptr_default()
{
    return defaultVecPtr;
}





/*
Functions to clean up colum and matrix resources
*/

SSGC_API void __cdecl deinitM128(__m128* ptr)
{
    _aligned_free(ptr);
}