as much as it hurts, async is holding us back #9
Conversation
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
About my testing: for some reason, I can't push loopback further than ~40Gbit/s on my machine, no matter the connections. On the other hand, if I only flood with a single connection, I get ~10Gbit/s with sturmflut with the async version, with breakwater running at 102% cpu. With the sync version, I get ~14Gbit/s, still at 102% cpu usage. I'm not saying that I like the sync version more than the async version, but it is definitely noticeably faster. If I hammer them with 10 connections, both versions perform similar in terms of bandwidth.
|
something else to note: if I use pixelbomber, that roughly halves the amount of bytes per pixel, with the async version, I get only ~6.5Gbit/s, but with the sync version, I get 14Gbit/s, just like with sturmflut. |
|
Thx for the measurements, would you suggest switching back to non-async code? |
|
I'm still searching for the caveat in them. It's probably just that we let the kernel manage threads and interrupts, and if we use async, we do the same again, just in tokio. Feel free to verify these findings. |
|
After testing different features, I noticed, that my async performance isn't that bad. Was probably just me confusing earlier benchmarks |
|
Puh, good news! Thanks for the update! |
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.
About my testing: for some reason, I can't push loopback further than ~40Gbit/s on my machine, no matter the connections.
On the other hand, if I only flood with a single connection, I get ~10Gbit/s with sturmflut with the async version, with breakwater running at 102% cpu. With the sync version, I get ~14Gbit/s, still at 102% cpu usage.
I'm not saying that I like the sync version more than the async version, but it is definitely noticeably faster. If I hammer them with 10 connections, both versions perform similar in terms of bandwidth.
Tests are not changed to sync yet