Update reduction

Bram Veenboerrequested to merge
update-reduction into main

This merge request has many changes related to parallel reductions. These changes have now been separated to create this new merge request.

The main changes are:

  • Move reduction code out of QuantizeOutput.cu into reduction.cuh
  • Make the reduce_sum function more robust (e.g. correct regardless of block size)
  • Add tReduction to test the functionality provided in reduction.cuh
Edited by Bram Veenboer

Merge request reports