Add something about early bevy results.

Author: nical

Diff for: _posts/2023-11-24-arcanization.md

@@ -32,7 +32,7 @@ Interestingly, `wgpu` also had to maintain internal reference counts to resource
 
 ![A simplified diagram showing resources stored via Arcs](img/arcanization-after.png)
 
-The result is much lower lock contention. If you use `wgpu` from multiple threads, this should significantly improve performance.
+The result is much lower lock contention. If you use `wgpu` from multiple threads, this should significantly improve performance. Our friends in the [bevy engine][bevy] community noted that some initial testing showed that with arcanization, the encoding of  shadow-related commands can run in parallel with the main passes, yielding 45% frame time reduction on a test scene (the famous bistro scene) compared to their single threaded configuration. Without arcanization, lock contention is too high to reliably improve performance.
 
 In addition, `wgpu`'s internals are now simpler. This change lifted some restrictions and opens the door for further performance and ergonomics improvements.
 
@@ -68,4 +68,5 @@ Thanks [@gents83]!
 [@gents83]: https://github.com/gents83
 [rp-pr]: https://github.com/gfx-rs/wgpu/issues/1453
 [arc-pr]: https://github.com/gfx-rs/wgpu/pull/3626
+[bevy]: https://bevyengine.org/