#pragma once

#include "concepts.h"

struct BgkCollideO {

template <concepts::Arithmetic V, concepts::Arithmetic T>
static void apply(V f_curr[19], V f_next[19], T tau) {
	V rho;
	V u[3];
	V m0 = f_curr[10] + f_curr[13] + f_curr[17];
	V m1 = f_curr[14] + f_curr[5] + f_curr[6];
	V m2 = f_curr[1] + f_curr[2] + f_curr[4];
	V m3 = m0 + m1 + m2 + f_curr[0] + f_curr[11] + f_curr[12] + f_curr[15] + f_curr[16] + f_curr[18] + f_curr[3] + f_curr[7] + f_curr[8] + f_curr[9];
	V m4 = -f_curr[11] + f_curr[7];
	V m5 = -f_curr[15] + f_curr[3];
	V m6 = T{1}/m3;
	V m7 = f_curr[0] - f_curr[18];
	rho = m3;
	u[0] = m6*(m0 + m4 + m5 - f_curr[1] - f_curr[5] - f_curr[8]);
	u[1] = m6*(m1 + m4 + m7 - f_curr[12] - f_curr[13] - f_curr[4]);
	u[2] = m6*(m2 + m5 + m7 - f_curr[14] - f_curr[16] - f_curr[17]);
	V x0 = T{1}/tau;
	V x1 = T{0.0138888888888889}*x0;
	V x2 = u[1] + u[2];
	V x3 = T{9.00000000000000}*(x2*x2);
	V x4 = T{6.00000000000000}*u[2];
	V x5 = u[1]*u[1];
	V x6 = T{3.00000000000000}*x5;
	V x7 = -x6;
	V x8 = u[0]*u[0];
	V x9 = T{3.00000000000000}*x8;
	V x10 = T{2.00000000000000} - x9;
	V x11 = x10 + x7;
	V x12 = x11 + x4;
	V x13 = u[2]*u[2];
	V x14 = T{3.00000000000000}*x13;
	V x15 = -x14;
	V x16 = T{6.00000000000000}*u[1];
	V x17 = x15 + x16;
	V x18 = T{6.00000000000000}*u[0];
	V x19 = -u[0];
	V x20 = x19 + u[2];
	V x21 = x14 + x6 + T{-2.00000000000000};
	V x22 = x21 + x9;
	V x23 = x22 - x4;
	V x24 = T{0.0277777777777778}*x0;
	V x25 = T{6.00000000000000}*x13;
	V x26 = u[0] + u[2];
	V x27 = T{9.00000000000000}*(x26*x26);
	V x28 = x15 + x18;
	V x29 = -u[1];
	V x30 = x29 + u[2];
	V x31 = x19 + u[1];
	V x32 = -x16;
	V x33 = x18 + x22;
	V x34 = T{6.00000000000000}*x5;
	V x35 = u[0] + u[1];
	V x36 = T{9.00000000000000}*(x35*x35);
	V x37 = T{6.00000000000000}*x8;
	V x38 = x9 + T{-2.00000000000000};
	V x39 = x29 + u[0];
	V x40 = -x18 + x22;
	V x41 = -u[2];
	V x42 = x41 + u[1];
	V x43 = x22 + x4;
	V x44 = x41 + u[0];
	f_next[0] = x1*(rho*(x12 + x17 + x3) - T{72.0000000000000}*f_curr[0]) + f_curr[0];
	f_next[1] = -x1*(rho*(x18 + x23 - T{9.00000000000000}*x20*x20) + T{72.0000000000000}*f_curr[1]) + f_curr[1];
	f_next[2] = x24*(rho*(x12 + x25) - T{36.0000000000000}*f_curr[2]) + f_curr[2];
	f_next[3] = x1*(rho*(x12 + x27 + x28) - T{72.0000000000000}*f_curr[3]) + f_curr[3];
	f_next[4] = -x1*(rho*(x16 + x23 - T{9.00000000000000}*x30*x30) + T{72.0000000000000}*f_curr[4]) + f_curr[4];
	f_next[5] = -x1*(rho*(x32 + x33 - T{9.00000000000000}*x31*x31) + T{72.0000000000000}*f_curr[5]) + f_curr[5];
	f_next[6] = x24*(rho*(x10 + x17 + x34) - T{36.0000000000000}*f_curr[6]) + f_curr[6];
	f_next[7] = x1*(rho*(x11 + x17 + x18 + x36) - T{72.0000000000000}*f_curr[7]) + f_curr[7];
	f_next[8] = -x24*(rho*(x18 + x21 - x37) + T{36.0000000000000}*f_curr[8]) + f_curr[8];
	f_next[9] = -T{0.166666666666667}*x0*(rho*x22 + T{6.00000000000000}*f_curr[9]) + f_curr[9];
	f_next[10] = x24*(rho*(x28 + x37 + x7 + T{2.00000000000000}) - T{36.0000000000000}*f_curr[10]) + f_curr[10];
	f_next[11] = -x1*(rho*(x16 + x33 - x36) + T{72.0000000000000}*f_curr[11]) + f_curr[11];
	f_next[12] = -x24*(rho*(x14 + x16 - x34 + x38) + T{36.0000000000000}*f_curr[12]) + f_curr[12];
	f_next[13] = -x1*(rho*(x16 + x40 - T{9.00000000000000}*x39*x39) + T{72.0000000000000}*f_curr[13]) + f_curr[13];
	f_next[14] = -x1*(rho*(x32 + x43 - T{9.00000000000000}*x42*x42) + T{72.0000000000000}*f_curr[14]) + f_curr[14];
	f_next[15] = -x1*(rho*(-x27 + x33 + x4) + T{72.0000000000000}*f_curr[15]) + f_curr[15];
	f_next[16] = -x24*(rho*(-x25 + x38 + x4 + x6) + T{36.0000000000000}*f_curr[16]) + f_curr[16];
	f_next[17] = -x1*(rho*(x4 + x40 - T{9.00000000000000}*x44*x44) + T{72.0000000000000}*f_curr[17]) + f_curr[17];
	f_next[18] = -x1*(rho*(x16 - x3 + x43) + T{72.0000000000000}*f_curr[18]) + f_curr[18];
}

};