#include <metal_stdlib>
using namespace metal;

struct SimParams {
  float deltaT;
  float rule1Distance;
  float rule2Distance;
  float rule3Distance;
  float rule1Scale;
  float rule2Scale;
  float rule3Scale;
};

template<typename T, size_t N>
struct tint_array {
  const constant T& operator[](size_t i) const constant { return elements[i]; }
  device T& operator[](size_t i) device { return elements[i]; }
  const device T& operator[](size_t i) const device { return elements[i]; }
  thread T& operator[](size_t i) thread { return elements[i]; }
  const thread T& operator[](size_t i) const thread { return elements[i]; }
  threadgroup T& operator[](size_t i) threadgroup { return elements[i]; }
  const threadgroup T& operator[](size_t i) const threadgroup { return elements[i]; }
  T elements[N];
};

struct Particle {
  float2 pos;
  float2 vel;
};

struct Particles {
  tint_array<Particle, 5> particles;
};

struct tint_arg_buffer_struct_0 {
  const constant SimParams* params [[id(0)]];
  device Particles* particlesA [[id(1)]];
  device Particles* particlesB [[id(2)]];
};

void comp_main_inner(uint3 gl_GlobalInvocationID, const constant SimParams* const params, device Particles* const particlesA, device Particles* const particlesB) {
  uint index = gl_GlobalInvocationID.x;
  if ((index >= 5u)) {
    return;
  }
  float2 vPos = (*particlesA).particles[min(index, 4u)].pos;
  float2 vVel = (*particlesA).particles[min(index, 4u)].vel;
  float2 cMass = float2(0.0f);
  float2 cVel = float2(0.0f);
  float2 colVel = float2(0.0f);
  int cMassCount = 0;
  int cVelCount = 0;
  float2 pos = 0.0f;
  float2 vel = 0.0f;
  {
    uint i = 0u;
    while(true) {
      if ((i < 5u)) {
      } else {
        break;
      }
      if ((i == index)) {
        {
          i = (i + 1u);
        }
        continue;
      }
      pos = (*particlesA).particles[min(i, 4u)].pos.xy;
      vel = (*particlesA).particles[min(i, 4u)].vel.xy;
      if ((distance(pos, vPos) < (*params).rule1Distance)) {
        cMass = (cMass + pos);
        cMassCount = as_type<int>((as_type<uint>(cMassCount) + as_type<uint>(1)));
      }
      if ((distance(pos, vPos) < (*params).rule2Distance)) {
        colVel = (colVel - (pos - vPos));
      }
      if ((distance(pos, vPos) < (*params).rule3Distance)) {
        cVel = (cVel + vel);
        cVelCount = as_type<int>((as_type<uint>(cVelCount) + as_type<uint>(1)));
      }
      {
        i = (i + 1u);
      }
      continue;
    }
  }
  if ((cMassCount > 0)) {
    float2 const v = cMass;
    float const v_1 = float(cMassCount);
    float2 const v_2 = (v / float2(v_1, float(cMassCount)));
    cMass = (v_2 - vPos);
  }
  if ((cVelCount > 0)) {
    float2 const v_3 = cVel;
    float const v_4 = float(cVelCount);
    cVel = (v_3 / float2(v_4, float(cVelCount)));
  }
  vVel = (((vVel + (cMass * (*params).rule1Scale)) + (colVel * (*params).rule2Scale)) + (cVel * (*params).rule3Scale));
  vVel = (normalize(vVel) * clamp(length(vVel), 0.0f, 0.10000000149011611938f));
  vPos = (vPos + (vVel * (*params).deltaT));
  if ((vPos.x < -1.0f)) {
    vPos.x = 1.0f;
  }
  if ((vPos.x > 1.0f)) {
    vPos.x = -1.0f;
  }
  if ((vPos.y < -1.0f)) {
    vPos.y = 1.0f;
  }
  if ((vPos.y > 1.0f)) {
    vPos.y = -1.0f;
  }
  (*particlesB).particles[min(index, 4u)].pos = vPos;
  (*particlesB).particles[min(index, 4u)].vel = vVel;
}

kernel void comp_main(uint3 gl_GlobalInvocationID [[thread_position_in_grid]], const constant tint_arg_buffer_struct_0* tint_arg_buffer_0 [[buffer(0)]]) {
  tint_arg_buffer_struct_0 const v_5 = (*tint_arg_buffer_0);
  comp_main_inner(gl_GlobalInvocationID, v_5.params, v_5.particlesA, v_5.particlesB);
}