Closed formateu closed 5 years ago
std::unordered map is bottleneck of the algorithm, probably because of it's cache missing implementation. Need to exchange it for e.g. absl::flat_hash_map (when they will finally add cmake support) or change algorithm to use std::vector instead.
std::unordered map is bottleneck of the algorithm, probably because of it's cache missing implementation. Need to exchange it for e.g. absl::flat_hash_map (when they will finally add cmake support) or change algorithm to use std::vector instead.