#include "../lmath/lmath.h"
#include "../lmath/lm_vector_sse.h"
#include "../lmath/lm_vector_avx.h"
#include <iostream>
#include <amp.h>
#include <intrin.h>

using namespace concurrency;
using namespace lm;

void TestCPU() restrict(cpu) {

	double4_avx d0(1.0, 0, 1, 4);
	auto dm0 = d0.get<0>();
	auto dm1 = d0.get<1>();
	auto dm2 = d0.get<2>();
	auto dm3 = d0.get<3>();

	double dt = lm::dot(d0, d0);

	auto pp = lm::min(double4_avx(1, 2, 3, 4), double4_avx(4, 3, 2, 1));
	auto ppp0 = pp.get<0>();
	auto ppp1 = pp.get<1>();
	auto ppp2 = pp.get<2>();
	auto ppp3 = pp.get<3>();

	auto pplen = pp.length();

	auto f_sse3 = float4_sse(3) * float4_sse(4);

	auto l = float4_sse(0, 1, 2, 3);

	auto ls = lm::sum(l);

	auto lx = l.get<0>();

	Vector<uint16_t, 3> p00(0U);
	Vector<uint16_t, 3> p0;
	Vector<uint16_t, 5> p1((uint16_t)0U, (uint16_t)1U, (uint16_t)2U, (uint16_t)3U, (uint16_t)4U);

	Vector<uint16_t, 2> p3((uint16_t)0U, (uint16_t)1U);
	Vector<uint16_t, 3> p4((uint16_t)0U, (uint16_t)1U, (uint16_t)2U);
	Vector<uint16_t, 4> p5((uint16_t)0U, (uint16_t)1U, (uint16_t)2U, (uint16_t)3U);

	Vector<uint16_t, 2>::unitX();
	Vector<uint16_t, 3>::unitX();
	Vector<uint16_t, 4>::unitX();

	auto p3x = p3.y();


	p0.lengthSquared();
//	p0.length();

	auto pp0 = p0 / 2.0f;
	auto pp1 = p0 * 2.0f;
	auto pp2 = p0 + 2.0f;
	auto pp3 = p0 - 2.0f;

	p0 /= 2;
	p0 *= 2;
	p0 += 2;
	p0 -= 2;
	//p0.normalized();

	auto& v = p0.slice<1, 2>();

	Vector<float, 3> v0(1.0f, 2.0f, 3.0f);
	v0 = -v0;

	v0 == v0;
	v0 != v0;

	auto v1 = cross(v0, v0 + 1.0f);
	auto v2 = dot(v0, v0 + 1.0f);
	auto v3 = normalize(v0);
}

void TestAMP() restrict(amp) {
	Vector<uint32_t, 3> p00(0U);

	Vector<uint32_t, 3> p0;
	Vector<uint32_t, 5> p1(0U, 1U, 2U, 3U, 4U);

	Vector<uint32_t, 2> p3(0U, 1U);
	Vector<uint32_t, 3> p4(0U, 1U, 2U);
	Vector<uint32_t, 4> p5(0U, 1U, 2U, 3U);

	Vector<uint32_t, 2>::unitX();
	Vector<uint32_t, 3>::unitX();
	Vector<uint32_t, 4>::unitX();

	auto& v = p0.slice<1, 2>();



	auto pp0 = p0 / 2.0f;
	auto pp1 = p0 * 2.0f;
	auto pp2 = p0 + 2.0f;
	auto pp3 = p0 - 2.0f;

	p0 /= 2;
	p0 *= 2;
	p0 += 2;
	p0 -= 2;

	Vector<float, 3> v0(1.0f, 2.0f, 3.0f);
	v0 = -v0;

	v0 == v0;
	v0 != v0;

	auto v1 = cross(v0, v0 + 1.0f);
	auto v2 = dot(v0, v0 + 1.0f);
	auto v3 = normalize(v0);
}

int main() {
	TestCPU();

	Vector<uint16_t, 3> p0;

	std::vector<lm::float3> amp_vector;
	std::vector<lm::float3> amp_result_vector(amp_vector.size());

	array_view<lm::float3, 1> arr(amp_vector);
	array_view<lm::float3, 1> arr_result(amp_result_vector);
	arr_result.discard_data();


	parallel_for_each(arr.extent, [=](index<1> idx) restrict(amp) {
		TestAMP();
	});


	/*float4x4 ta;
	float4x4 tb;
	ta = tb;

	float4 pp(1.0f, 2.0f, 3.0f, 4.0f);
	float4 pn = -pp;

	auto& xyzw = pp.xyzw();
	auto& xyz = pp.xyz();
	auto& xy = pp.xy();

	auto& yzw = pp.yzw();
	auto& yz = pp.yz();

	auto& zw = pp.zw();

	auto& xx = pn.x();
	auto& yy = pn.y();
	auto& zz = pn.z();

	float2 fff;

	auto& xf = fff.x();
	auto& yf = fff.y();
	//auto& zf = fff.z();

	float* pf = (float*)pn;

	pp.slice<1, 2>() = float2(7.0f, 8.0f);

	Matrix<float, 4, 4> vvvv(
		Vector<float, 4>(1.0f, 0.0f, 0.0f, 0.0f),
		Vector<float, 4>(0.0f, 1.0f, 0.0f, 0.0f),
		Vector<float, 4>(0.0f, 0.0f, 1.0f, 0.0f),
		Vector<float, 4>(0.0f, 0.0f, 0.0f, 1.0f));

	auto deg1 = lm::degrees(lm::pi_d);
	auto deg2 = lm::degrees(lm::pi_d / 2.0);
	auto deg3 = lm::degrees(double2(pi_d / 2.0, pi_d / 3.0));

	float expf = lm::exp(1.0f);
	auto expv  = lm::exp(float2(2.0f, 3.0f));

	float logf = lm::log(1.0f);
	auto logv = lm::log(float2(2.0f, 3.0f));

	float minf1 = lm::min(1.0f, 2.0f);
	float minf2 = lm::min(2.0f, 1.0f);
	float maxf1 = lm::max(1.0f, 2.0f);
	float maxf2 = lm::max(2.0f, 1.0f);

	auto clampt = lm::clamp(float2(1.0f, 2.0f), float2(1.5f), float2(3.0f));

	auto minv1 = lm::min(float2(1.0f, 2.0f), float2(5.0f, 1.0f));
	auto minv2 = lm::min(float2(2.0f, 1.0f), float2(1.0f, 0.0f));
	auto maxv1 = lm::max(float2(1.0f, 2.0f), float2(0.0f, 0.0f));
	auto maxv2 = lm::max(float2(2.0f, 1.0f), float2(3.0f, 7.0f));

	auto ux = float3::unitX();
	auto pow_11 = lm::pow(float4(1.0f, 2.0f, 3.0f, 4.0f), 2.0f);

	Vector<Vector<float, 2>, 2> vv;
	vv[0] = float2(1.0f, 2.0f);
	vv[1] = float2(3.0f, 4.0f);
	auto pow_22 = lm::pow(vv, 2.0f);


	test_float2();
	


	float3 mul_self_0(1.0f, 2.0f, 3.0f);
	mul_self_0 *= 2;
	mul_self_0 *= float3(2.0f, 3.0f, 4.0f);
	mul_self_0 /= 2;
	mul_self_0 /= float3(4.0f, 4.0f, 4.0f);
	mul_self_0 += 1;
	mul_self_0 += float3(4.0f, 4.0f, 4.0f);

	mul_self_0 -= 2;
	mul_self_0 -= float3(2.0f, 2.0f, 2.0f);
	
	//AMP test
	parallel_for_each(arr.extent, [=](index<1> idx) restrict(amp) {


//		Half h1;

	float3x3 m1;
	
	auto col1 = m1.getColumn(0);
	auto m2 = float3x3::identity();
	auto m3 = float4x4::identity();

	auto b2 = float3x3::identity();
	auto m4 = lm::mul(m1, m2);

	float2x2 m22;


	auto det1 = determinant(m22);
	auto det2 = determinantAffine(m1);
	auto det3 = determinantAffine(m3);

	auto mi = inverse(m3, true);

	mi[0] = float4{ 0.0f,1.0f,2.0f,3.0f };
	mi[1] = float4{ 4.0f,5.0f,6.0f,7.0f };
	mi[2] = float4{ 8.0f,9.0f,10.0f,11.0f };
	mi[3] = float4{ 12.0f,13.0f,14.0f,15.0f };

	auto mm1 = (float2x2)mi;
	auto mm2 = (float3x3)mi;
	auto mm3 = (double4x4)mi;
	

	float3 t3(1.0f);
	auto len_1 = t3.lengthSquared();
	auto len_2 = t3.length();

	auto norm_1 = t3.normalized();
	auto dot_1 = dot(t3, t3);
	auto cross_1 = cross(float3::unitY(), float3::unitX());

	auto pow_1 = lm::pow(float4(1.0f, 2.0f, 3.0f, 4.0f), 2.0f);
	auto abs_1 = lm::abs(float4(1.0f, -2.0f, 3.0f, -4.0f));
	auto acos_1 = lm::acos(float4(1.0f, 0.0f, -1.0f, -0.75f));

	auto asin_1 = lm::asin(float4(1.0f, 0.0f, -1.0f, -0.75f));
	auto atan_1 = lm::atan(float4(1.0f, 0.0f, -1.0f, -0.75f));

	auto all_1 = lm::all(float4(0.0f, 0.0f, 0.0f, 0.0f));
	auto all_2 = lm::all(float4(0.0f, 0.0f, 0.0f, 1.0f));
	auto all_3 = lm::all(float4(0.0f, 0.0f, 1.0f, 1.0f));
	auto all_4 = lm::all(float4(0.0f, 1.0f, 1.0f, 1.0f));
	auto all_5 = lm::all(float4(1.0f, 1.0f, 1.0f, 1.0f));

	auto any_1 = lm::any(float4(0.0f, 0.0f, 0.0f, 0.0f));
	auto any_2 = lm::any(float4(0.0f, 0.0f, 0.0f, 1.0f));
	auto any_3 = lm::any(float4(0.0f, 0.0f, 1.0f, 1.0f));
	auto any_4 = lm::any(float4(0.0f, 1.0f, 1.0f, 1.0f));
	auto any_5 = lm::any(float4(1.0f, 1.0f, 1.0f, 1.0f));*/

/*	arr_result[idx] = lm::normalize(t3) + 2;

	float4x4 mat;
	mat.rows[0] = float4(1, 0, 0, 1);
	mat.rows[3] = float4(3);
	auto col0 = mat.get_column(0);

	auto floor_1 = lm::floor(float4(3.3f, 4.4f, 5.5f, 6.6f));
	auto mat33 = (float3x3)mat;
	auto mat44 = (float4x4)mat;

	auto mat33_2 = (float3x3)mat;

	auto is_mat1 = lm::matrix_traits::is_matrix<float4x4>::value;
	auto is_mat2 = lm::matrix_traits::is_matrix<float4>::value;

	auto mmul1 = lm::matrix_traits::can_multiply<float3x3, float3x3>::value;
	auto mmul2 = lm::matrix_traits::can_multiply<float3x3, float3x4>::value;
	auto mmul3 = lm::matrix_traits::can_multiply<float3x3, float4x3>::value;

	auto mul_r1 = lm::mul(mat33, mat33_2);

	auto is_identity_constexpr_4x4 = noexcept(float4x4::identity());
	auto is_identity_constexpr_3x3 = noexcept(float3x3::identity());
	auto id0 = float4x4::identity();
	auto id1 = float3x3::identity();
	auto id2 = float2x2::identity();

	typedef Matrix<double, 5, 5> double5x5;
	auto is_identity_constexpr_5x5 = noexcept(double5x5::identity());
	auto id3 = double5x5::identity();

	auto rc = double5x5::rows_count;

	auto mul_r = lm::mul(float3x4(1), float4x3(1));

	auto mul_r2 = lm::mul(float3x3::identity(), float3x3::identity());

	auto mul_r3 = lm::mul(
	float2x3(float3(1, 3, 2), float3(0, 4, -1)),
	float3x4(float4(2, 0, -1, 1), float4(3, -2, 1, 2), float4(0, 1, 2, 3)));


	auto clamp_1 = lm::clamp(test_vector, 1.0f, 2.0f);
	auto clamp_2 = lm::clamp(test_matrix, 0.0f, 7.0f);


	auto cosh_1 = lm::cosh(test_vector);
	auto cosh_2 = lm::cosh(test_matrix);

	auto cross_2 = lm::cross(float3(0, 1, 0), float3(0, 0, 1));
	auto cross_3 = lm::cross(float3(0, 0, 1), float3(0, 1, 0));

	auto degrees_0 = lm::degrees(float3((float)LM_PI / 2.0f, (float)LM_PI / 1.0f, (float)LM_PI / 4.0f));

	auto det_1 = lm::determinant(float2x2(float2(1, 2), float2(-1, 3)));

	auto det_2 = lm::determinant(float3x3::identity());



	auto norm_2 = lm::normalize(float3(1, 2, 4));


	
	});*/

	/*arr_result.synchronize(access_type_auto);

	auto val = std::is_same<float, typename float3::element_type>::value;

	static_assert(sizeof(float2) == sizeof(float) * 2, "Float2 size incorrect");
	static_assert(sizeof(double2) == sizeof(double) * 2, "double2 size incorrect");

	static_assert(sizeof(float3) == sizeof(float) * 3, "Float2 size incorrect");
	static_assert(sizeof(double3) == sizeof(double) * 3, "double3 size incorrect");

	static_assert(sizeof(float4) == sizeof(float) * 4, "Float4 size incorrect");
	static_assert(sizeof(double4) == sizeof(double) * 4, "double4 size incorrect");

	//CPU test
	test_float2();
	test_float3();
	test_float4();

	auto asdouble_0 = lm::asdouble(123, 456);
	auto asdouble_1 = lm::asdouble(Vector<uint32_t, 2>(2, 2), Vector<uint32_t, 2>(300, 400));


	float3x3 m1;
	auto col1 = m1.get_column(0);
	auto m2 = float3x3::identity();
	auto m3 = float4x4::identity();
	auto m4 = lm::mul(m1, m2);
	auto det1 = lm::determinant(m4);

	float3 t3(1);
	auto len_1 = lm::lengthSquared(t3);
	auto len_2 = lm::length(t3);
	auto norm_1 = lm::normalize(t3);
	auto dist_1 = lm::distance(t3, norm_1);
	auto dot_1 = lm::dot(lm::normalize(t3), lm::normalize(t3));




	std::cin.get();


	//static_assert(noexcept(Test_amp()), "Not constexpr (");

	static constexpr float3 f3 = float3::up();

	static constexpr float4 f4 = float4(1, 0, 0, 1);

	static constexpr float4 f4_value = float4(1);

	static constexpr float4 f4_init = { 1.0f,2.0f,3.0f,4.0f };



	double3 d_add(4);
	float3 f_add(1);
	f_add = f_add + 2;

	auto b0 = f_add + d_add;
	auto b1 = f_add - d_add;
	auto b2 = f_add * d_add;
	auto b3 = f_add / d_add;

	auto b4 = f_add + 2;
	auto b5 = f_add - 2;
	auto b6 = f_add * 2;
	auto b7 = f_add / 2;

	float3 fup = float3::up() * 2;
	double3 dup = double3::up() * 2;

	auto flen_s = lm::lengthSquared(fup);
	auto flen = lm::length(fup);


	auto dlen_s = lm::lengthSquared(dup);
	auto dlen = lm::length(dup);

	auto pow_1 = lm::pow(float4(1, 2, 3, 4), 2.0f);
	auto abs_1 = lm::abs(float4(1, -2, 3, -4));
	auto acos_1 = lm::acos(float4(1, 0, -1, -0.75f));

	float4x4 mat;
	mat.rows[0] = float4(1, 0, 0, 1);
	mat.rows[3] = float4(3);
	auto col0 = mat.get_column(0);


	float4x4 mWorld;

	mWorld = float4x4::identity();


	auto mat33 = (float3x3)mat;
	auto mat44 = (float4x4)mat;

	auto mat33_2 = (float3x3)mat;

	auto is_mat1 = lm::matrix_traits::is_matrix<float4x4>::value;
	auto is_mat2 = lm::matrix_traits::is_matrix<float4>::value;

	auto mmul1 = lm::matrix_traits::can_multiply<float3x3, float3x3>::value;
	auto mmul2 = lm::matrix_traits::can_multiply<float3x3, float3x4>::value;
	auto mmul3 = lm::matrix_traits::can_multiply<float3x3, float4x3>::value;

	auto mul_r1 = lm::mul(mat33, mat33_2);


	auto is_identity_constexpr_4x4 = noexcept(float4x4::identity());
	auto is_identity_constexpr_3x3 = noexcept(float3x3::identity());
	auto id0 = float4x4::identity();
	auto id1 = float3x3::identity();
	auto id2 = float2x2::identity();


	typedef Matrix<double, 5, 5> double5x5;
	auto is_identity_constexpr_5x5 = noexcept(double5x5::identity());
	auto id3 = double5x5::identity();

	auto rc = double5x5::rows_count;

	auto mul_r = lm::mul(float3x4(1), float4x3(1));

	auto mul_r2 = lm::mul(float3x3::identity(), float3x3::identity());

	auto mul_r3 = lm::mul(
		float2x3(float3(1, 3, 2), float3(0, 4, -1)),
		float3x4(float4(2, 0, -1, 1), float4(3, -2, 1, 2), float4(0, 1, 2, 3)));

	auto is_square1 = lm::matrix_traits::is_square<double5x5>::value;




	auto asin_1 = lm::asin(float4(1, 0, -1, -0.75f));
	auto atan_1 = lm::atan(float4(1, 0, -1, -0.75f));

	auto all_1 = lm::all(float4(0, 0, 0, 0));
	auto all_2 = lm::all(float4(0, 0, 0, 1));
	auto all_3 = lm::all(float4(0, 0, 1, 1));
	auto all_4 = lm::all(float4(0, 1, 1, 1));
	auto all_5 = lm::all(float4(1, 1, 1, 1));

	auto any_1 = lm::any(float4(0, 0, 0, 0));
	auto any_2 = lm::any(float4(0, 0, 0, 1));
	auto any_3 = lm::any(float4(0, 0, 1, 1));
	auto any_4 = lm::any(float4(0, 1, 1, 1));
	auto any_5 = lm::any(float4(1, 1, 1, 1));




	float4 test_vector = float4(3.3f, 4.4f, 5.5f, 6.6f);
	float3x3 test_matrix = float3x3(float3(1.1f, 2.2f, 3.3f), float3(4.4f, 5.5f, 6.6f), float3(7.7f, 8.8f, 9.9f));

	auto ceil_1 = lm::ceil(test_vector);
	auto ceil_2 = lm::ceil(test_matrix);

	auto clamp_1 = lm::clamp(test_vector, 1.0f, 2.0f);
	auto clamp_2 = lm::clamp(test_matrix, 0.0f, 7.0f);

	auto cos_1 = lm::cos(test_vector);
	auto cos_2 = lm::cos(test_matrix);

	auto cosh_1 = lm::cosh(test_vector);
	auto cosh_2 = lm::cosh(test_matrix);

	auto cross_1 = lm::cross(float3(0, 0, 1), float3(1, 0, 0));
	auto cross_2 = lm::cross(float3(0, 1, 0), float3(0, 0, 1));
	auto cross_3 = lm::cross(float3(0, 0, 1), float3(0, 1, 0));

	auto degrees_0 = lm::degrees(float3((float)LM_PI / 2.0f, (float)LM_PI / 1.0f, (float)LM_PI / 4.0f));

	auto det_1 = lm::determinant(float2x2(float2(1, 2), float2(-1, 3)));

	auto det_2 = lm::determinant(float3x3::identity());



	auto norm_2 = lm::normalize(float3(1, 2, 4));

	auto p1 = float3::up() / lm::length(float3::up());




	auto dot_2 = lm::dot(float3::up(), float3::up());
	auto dot_3 = lm::dot(float3::up(), lm::normalize(float3::up() + float3::right()));

	auto deg_1 = lm::degrees(std::acos(dot_1));
	auto deg_2 = lm::degrees(std::acos(dot_2));
	auto deg_3 = lm::degrees(std::acos(dot_3));



	*/
	return 0;
}