#include <metal_stdlib>
using namespace metal;

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 S {
  simdgroup_float8x8 l;
  simdgroup_float8x8 r;
};

struct S_Nested {
  S s;
};

struct tint_module_vars_struct {
  device tint_array<float, 1>* tint_member;
  const constant tint_array<uint4, 1>* tint_storage_buffer_sizes;
};

void foo(thread simdgroup_float8x8* const m, thread tint_array<simdgroup_float8x8, 4>* const m_array, thread tint_array<tint_array<simdgroup_float8x8, 4>, 4>* const m_nested_array, thread S* const m_struct, thread S_Nested* const m_nested_struct, tint_module_vars_struct tint_module_vars) {
  simdgroup_float8x8 const v = (*m);
  if ((((0u + (64u * 7u)) + 8u) <= ((*tint_module_vars.tint_storage_buffer_sizes)[0u].x / 4u))) {
    simdgroup_store(v, (&(*tint_module_vars.tint_member)[0u]), ulong(64u), ulong2(0ul), false);
  }
  simdgroup_float8x8 const v_1 = (*m_array)[0u];
  if ((((0u + (64u * 7u)) + 8u) <= ((*tint_module_vars.tint_storage_buffer_sizes)[0u].x / 4u))) {
    simdgroup_store(v_1, (&(*tint_module_vars.tint_member)[0u]), ulong(64u), ulong2(0ul), false);
  }
  simdgroup_float8x8 const v_2 = (*m_nested_array)[1u][2u];
  if ((((0u + (64u * 7u)) + 8u) <= ((*tint_module_vars.tint_storage_buffer_sizes)[0u].x / 4u))) {
    simdgroup_store(v_2, (&(*tint_module_vars.tint_member)[0u]), ulong(64u), ulong2(0ul), false);
  }
  simdgroup_float8x8 const v_3 = (*m_struct).l;
  if ((((0u + (64u * 7u)) + 8u) <= ((*tint_module_vars.tint_storage_buffer_sizes)[0u].x / 4u))) {
    simdgroup_store(v_3, (&(*tint_module_vars.tint_member)[0u]), ulong(64u), ulong2(0ul), false);
  }
  simdgroup_float8x8 const v_4 = (*m_nested_struct).s.r;
  if ((((0u + (64u * 7u)) + 8u) <= ((*tint_module_vars.tint_storage_buffer_sizes)[0u].x / 4u))) {
    simdgroup_store(v_4, (&(*tint_module_vars.tint_member)[0u]), ulong(64u), ulong2(0ul), false);
  }
}

kernel void v_5(device tint_array<float, 1>* v_6 [[buffer(0)]], const constant tint_array<uint4, 1>* tint_storage_buffer_sizes [[buffer(30)]]) {
  tint_module_vars_struct const tint_module_vars = tint_module_vars_struct{.tint_member=v_6, .tint_storage_buffer_sizes=tint_storage_buffer_sizes};
  simdgroup_float8x8 m = make_filled_simdgroup_matrix<float, 8, 8>(0.0f);
  tint_array<simdgroup_float8x8, 4> m_array = {};
  tint_array<tint_array<simdgroup_float8x8, 4>, 4> m_nested_array = {};
  S m_struct = {};
  S_Nested m_nested_struct = {};
  foo((&m), (&m_array), (&m_nested_array), (&m_struct), (&m_nested_struct), tint_module_vars);
}