This is an experimental queue to be used in multi-threaded scenarios, like the task processors. More specifically: - internally synchronized for both consumers and producers (MPMC). - backed by an array, which is fast to access - bounded: the capacity is specified at creation
Unlike other implementations, such as crossbeam-queue, it doesn't carry a atomic bit per element.
Checked by both Miri and Loom on CI.
Note: experimental and currently slower than alternatives.