Function logical_reduce

include/cute/algorithm/tensor_reduce.hpp:86–105

          class TargetProfile,
          class BinaryOp = cute::plus>
CUTE_HOST_DEVICE constexpr
void
logical_reduce(Tensor<SrcEngine, SrcLayout> const& src,
               Tensor<DstEngine, DstLayout>      & dst,
               TargetProfile                const& target_profile,
               BinaryOp op = {})
{
  // Precondition
  assert(compatible(target_profile, shape(src)));

  auto diced_layout = dice(target_profile, src.layout());
  auto sliced_layout = slice(target_profile, src.layout());

  auto red_mode = conditional_return<rank(diced_layout) == Int<0>{}>(Layout<_1,_0>{}, diced_layout);
  auto batch_mode = conditional_return<rank(sliced_layout) == Int<0>{}>(Layout<_1,_0>{}, sliced_layout);

  auto src_tensor = make_tensor(src.data(), make_layout(red_mode, batch_mode));

  batch_reduce(src_tensor, dst, op);
}

} // end namespace cute