What is the fastest way of Boost::compute vector partial sum

I have a vector of 10M float. I want to know sum of every 100 elements, so 10000 sums in total. What is the fastest way to do this?

I'd recommend using reduce_by_key algorithm, fancy iterators and Boost.Compute lambda expr. Every 100 elements are marked with the same key and reduced. I'm not sure if you can replace keys_output with a discard_iterator to save some performance.

boost::compute::vector<int> keys_output(values_input.size()/100, context);
boost::compute::vector<int> values_output(values_input.size()/100, context);

boost::compute::reduce_by_key(
    boost::compute::make_transform_iterator(
      boost::compute::make_counting_iterator<int>(0),
      boost::compute::_1 / 100
    ),
    boost::compute::make_transform_iterator(
      boost::compute::make_counting_iterator<int>(values_input.size()),
      boost::compute::_1 / 100
    ),
    values_input.begin(),
    keys_output.begin(), 
    values_output.begin(),
    queue
 );