July 20, 2018
- Fixed crash (3.12.0 regression) on copying of empty bm::bvector<>
- Major optimization of bm::set2set_11_transform<> (4x times speedup).
- Reworked svsample05 to better explain use of bm::sparse_vector<> as associative container
for set-to-set remapping operations
- Improved performance of logical OR operation on bm::bvector<>
- Implemented bm::sparse_vector<>::gather() for faster selective retrieval of values from
bit-transposed containers. SSE4.2 and AVX2 optimizations for gather operation.
- Fixed some harmless but inconveninent compilation warnings
- Implemented new bm::bvector<> functions copy_range() and range copy contructor for
faster horizontal partitioning
- Fixed some warnings, cleaned code for better overall compatibility
Download v 3.12.6