commonware-runtime 2026.9.0

Execute asynchronous tasks with a configurable scheduler.
Documentation 
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36

//! Benchmarks for `runtime::iobuf`.
//!
//! This entry point registers four suites:
//!
//! - [`iobuf`]: fixed-size decode benchmarks comparing `Bytes` with `IoBuf`
//!   backed by external `Bytes` or native heap storage. `Vec<u8>` modes
//!   provide the deep-clone baseline.
//! - [`allocation`]: end-to-end steady-state `BufferPool` allocation and
//!   reuse, compared against direct aligned allocation.
//! - [`reuse`]: end-to-end global reuse with thread-local caching disabled,
//!   plus forced thread-local cache refill and spill cycles.
//! - [`freelist`]: microbenchmarks of the global freelist that stores free
//!   pooled buffers shared across threads, compared against `Mutex<Vec<_>>` and
//!   `ArrayQueue`.
//!
//! Shared threading and measurement helpers live in [`utils`].
//!
//! Run with: `cargo bench --bench iobuf -p commonware-runtime --features bench`

use criterion::{criterion_group, criterion_main};

mod allocation;
mod freelist;
mod iobuf;
mod reuse;
mod utils;

criterion_group!(
    benches,
    iobuf::bench,
    allocation::bench,
    reuse::bench,
    freelist::bench
);

criterion_main!(benches);