#include <metal_stdlib>
using namespace metal;

struct tint_array_lengths_struct {
  uint tint_array_length_0_0;
};

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 SB_RW {
  /* 0x0000 */ tint_array<float, 1> arg_0;
};

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

void subgroupMatrixStore_543411(tint_array_lengths_struct tint_array_lengths, tint_module_vars_struct tint_module_vars) {
  uint arg_1 = 1u;
  simdgroup_float8x8 arg_2 = simdgroup_float8x8();
  uint arg_4 = 8u;
  uint const v = arg_1;
  simdgroup_float8x8 const v_1 = arg_2;
  uint const v_2 = max(arg_4, 8u);
  if ((((v + (v_2 * 7u)) + 8u) <= tint_array_lengths.tint_array_length_0_0)) {
    simdgroup_store(v_1, (&(*tint_module_vars.sb_rw).arg_0[v]), ulong(v_2), ulong2(0ul), true);
  }
}

kernel void compute_main(device SB_RW* sb_rw [[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{.sb_rw=sb_rw, .tint_storage_buffer_sizes=tint_storage_buffer_sizes};
  subgroupMatrixStore_543411(tint_array_lengths_struct{.tint_array_length_0_0=(((*tint_module_vars.tint_storage_buffer_sizes)[0u].x - 0u) / 4u)}, tint_module_vars);
}