Per-object thread-local storage

Efficient per-object thread-local storage implementation.

Unlike thread_local crate, it can release objects normally at the end of thread, and will not move objects between multiple threads (see playground and document).

Unlike os-thread-local crate, it only uses the std::thread_local! abstraction of rust standard library, and does not use any other platform-related APIs. This means that its capacity is not limited by PTHREAD_KEYS_MAX.

And its performance is relatively good, value access for less than N threads is completely lock-free, and has O(1) time complexity. But since we store thread id in std::thread_local!, so we will be slightly slower than std::thread_local!.

std::thread_local!      time:   [1.5702 us 1.5725 us 1.5751 us]
                        change: [-3.6008% -0.8194% +1.2594%] (p = 0.58 > 0.05)
                        No change in performance detected.
Found 7 outliers among 100 measurements (7.00%)
  7 (7.00%) high severe

per-thread-object       time:   [1.6411 us 1.6545 us 1.6697 us]
                        change: [-0.4540% +1.1403% +2.6958%] (p = 0.16 > 0.05)
                        No change in performance detected.
Found 9 outliers among 100 measurements (9.00%)
  6 (6.00%) high mild
  3 (3.00%) high severe

thread_local            time:   [2.7854 us 2.7996 us 2.8142 us]
                        change: [-3.9626% -3.0122% -1.9548%] (p = 0.00 < 0.05)
                        Performance has improved.
Found 4 outliers among 100 measurements (4.00%)
  2 (2.00%) high mild
  2 (2.00%) high severe

os-thread-local         time:   [1.5628 us 1.5708 us 1.5801 us]
                        change: [-2.7584% +0.1343% +3.1291%] (p = 0.94 > 0.05)
                        No change in performance detected.
Found 10 outliers among 100 measurements (10.00%)
  2 (2.00%) high mild
  8 (8.00%) high severe

License

This project is licensed under the MIT license.