---
title: Contiguous query access
authors: ["@Jenya705"]
pull_requests: [21984, 24181]
---
[SIMD] is a critical modern tool for performance optimization, but using it with Bevy has always been harder than it needed to be.
Table components in Bevy are already laid out flat in memory — all [`Transform`] components are stored as values contiguously in a table, exactly what SIMD wants.
However, the [`Query`] iterator wasn't exposing that structure; it handed you one entity's component at a time, and the compiler had no way to know that the underlying data was a contiguous array.
There are now functions to enable this use case. [`contiguous_iter`] and [`contiguous_iter_mut`] hand you the whole table slice at once. LLVM can see the contiguous array and auto-vectorize — or you can reach for explicit SIMD yourself.
On a bulk `position += velocity` update over 10,000 entities, this gives some serious speedups:
| Normal iteration | 5.58 µs | 5.51 µs |
| Contiguous iteration | 4.88 µs | 1.87 µs |
| Contiguous, no change detection | 4.40 µs | 1.58 µs |
If your project has CPU-heavy workloads (physics engines are a prime example), you should try this out immediately.
```rust
fn apply_health_decay(mut query: Query<(&mut Health, &HealthDecay)>) {
for (mut health, decay) in query.contiguous_iter_mut().unwrap() {
for (h, d) in health.iter_mut().zip(decay) {
h.0 *= d.0;
}
}
}
```
The `contiguous_iter*` family of methods returns `Err` if the storage strategy of any of the components is not the default, `"table"`.
Be careful unwrapping this result if you are writing generic code or working with dynamic components.
You may have noticed that the table above had *three* rows.
While change detection is a generally useful feature, it does incur measurable performance overhead.
By default, [`contiguous_iter_mut`] returns [`ContiguousMut<T>`]s.
Just like the ordinary [`Mut<T>`], it triggers change detection automatically on dereference.
If you don't care about that, `bypass_change_detection()` gives you the raw `&mut [T]` directly for even faster access.
Vroom!
[SIMD]: https://en.wikipedia.org/wiki/Single_instruction,_multiple_data
[`contiguous_iter_mut`]: https://docs.rs/bevy/0.19.0/bevy/prelude/struct.Query.html#method.contiguous_iter_mut
[`contiguous_iter`]: https://docs.rs/bevy/0.19.0/bevy/prelude/struct.Query.html#method.contiguous_iter
[`Transform`]: https://docs.rs/bevy/0.19.0/bevy/transform/components/struct.Transform.html
[`Query`]: https://docs.rs/bevy/0.19.0/bevy/ecs/system/struct.Query.html
[`ContiguousMut<T>`]: https://docs.rs/bevy/0.19.0/bevy/ecs/change_detection/struct.ContiguousMut.html
[`Mut<T>`]: https://docs.rs/bevy/0.19.0/bevy/ecs/change_detection/struct.Mut.html