ccl

blazing fast data structures for Rust

cpu performance?

Here are some benchmarks for the concurrent DHashMap hashmap. You can run them yourself by executing cargo bench in the ccl-bench directory. Benchmarks are made with the default settings.

``` Xeon 2.1Ghz 16C/32T Hetzner Cloud CXX51

dhashmapcclrayoninsertonly100ku64_u64
time: [1.7019 ms 1.7097 ms 1.7176 ms]

dhashmapcclrayoninsertonly100ku64_u128x16
time: [2.3916 ms 2.4040 ms 2.4173 ms]

dhashmapcclrayonreadonly100ku64_u64
time: [1.2478 ms 1.2487 ms 1.2497 ms]

dhashmapcclrayonreadonly100ku64_u128x16
time: [1.5210 ms 1.5259 ms 1.5329 ms] ```

todo list

• add a new interface api for managing borrows to prevent deadlocks

• explore new concurrent hashmap designs possibly based on c++ junction