ecs_bench_suite

ecs_bench_suite copied to clipboard

Update bevy, hecs, legion, specs and shipyard

2

zakarumych

Adding [`brood`](https://crates.io/crates/brood) to the benchmark suite. I included the benchmark results as well. From the other PRs I looked at, it seems that's the common practice, although my machine will...

Anders429

"Schedule" benchmark is heavily influenced by iteration performance

3

Systems `(A, B)`, `(C, D)`, and `(C, E)` are iterating through 40k, 10k, and 10k entites respectively. I argue this is measuring iteration performance more than it does scheduler performance,...

Ratysz

Heavy compute does not give a good comparison of parallel iter

2

Hi, I think there should be a benchmark that compares how the libraries handle parallel iteration. Currently, the closest test for this would be `heavy_compute`, but the task (inverting a...

ElliotB256

Add Edict ECS

2

Update other ECS implementations to latest versions

zakarumych

Bump legion to 0.4

1

* Updated legion to 0.4 The performance of legion 0.4 seems to have improved slightly in some test cases.

FlyingRatBull

Use faster .for_each iteration in Bevy code

1

Iterating with the `for_each` is substantially faster in Bevy as of 0.5. While many systems will not be written in this way (for loops are great!), performance critical ones will...

alice-i-cecile

Use sparse set components for fragmented_iter for Bevy

1

As discussed in the [0.5 news release](https://bevyengine.org/news/bevy-0-5/#hybrid-component-storage-the-solution), sparse set component storage was designed to make fragmented iteration faster, precisely the use case being tested in the `fragmented_iter` benchmark.

alice-i-cecile

Consider revising "Simple Insert"

1

"Simple Insert" seems to test a scenario that is very rare outside of benchmarks, and yet can uniquely optimized for in a benchmark. How often will anyone initialize 10,000 entities...

kettle11

Excellence in that is the main thing setting archetypal ECS layouts apart, so it's an odd omission.

Ralith