Release 1.1 (Dezember 2015)

  • Fixed wrong parameter assignment in ‘kneighbors’ method of both neighbors/kd_tree/ and neighbors/buffer_kdtree/
  • Added Multi-GPU support to brute-force approach (for benchmark purposes)
  • Adapted parameter settings for buffer k-d tree implementation
  • Added benchmark example

Release 1.0.2 (September 2015)

  • Adapted building process

Release 1.0.1 (September 2015)

  • Adapted building process
  • Fixed small bugs

Release 1.0 (September 2015)

  • First major release
  • Python wrappers for three implementations (‘brute’, ‘kd_tree’, ‘buffer_kd_tree’)
  • Performance improvements for both kd-tree based implementations
  • Multi-OpenCL-Device support for ‘buffer_kd_tree’ implementation
  • Large-scale construction for ‘buffer_kd_tree’ implementation
  • Multi-OpenCL-Device support for query phase (queries are processed in chunks)
  • Added Sphinx documentation