TEST(scheduling, fuseBug) {
    Tensor<double> A("A", {4, 4}, {Dense, Dense});
    Tensor<double> B("B", {4, 4}, {Dense, Dense});
    Tensor<double> C("C", {4, 4}, {Dense, Dense});
    IndexVar i("i"), j("j"), f("f");
    C(i, j) = A(i, j) * B(i, j);
    auto stmt = C.getAssignment().concretize();
    C.compile(stmt.fuse(i, j, f));
}

int compute(taco_tensor_t *A, taco_tensor_t *B, taco_tensor_t *C) {
  int A1_dimension = (int)(A->dimensions[0]);
  int A2_dimension = (int)(A->dimensions[1]);
  double* restrict A_vals = (double*)(A->vals);
  double* restrict B_vals = (double*)(B->vals);
  int C1_dimension = (int)(C->dimensions[0]);
  int C2_dimension = (int)(C->dimensions[1]);
  double* restrict C_vals = (double*)(C->vals);

  for (int32_t f = 0; f < (C1_dimension * C2_dimension); f++) {
    int32_t i = f / C2_dimension;
    if (i >= C1_dimension)
      continue;

    int32_t j = f % C2_dimension;
    if (j >= C2_dimension)
      continue;

    int32_t jA = i * A2_dimension + j;
    A_vals[jA] = B_vals[jB] * C_vals[jC];
  }
  return 0;
}