⨯ You can sign up to get a daily email of our articles, see the Mailing List page.
Stadia Devs on Linux schedulers BMQ is the fastest then Muqqs
Koopacabras Jan 3, 2020
this is a nice read indeed

https://www.phoronix.com/scan.php?page=news_item&px=Linux-2020-Scheduler-Bugs-Stadia

also check the original blog post that has some benchmarks

https://probablydance.com/2019/12/30/measuring-mutexes-spinlocks-and-how-bad-the-linux-scheduler-really-is/


had to try to Bitmap Queue Scheduler, I compiled it using this patchset
https://gitlab.com/post-factum/pf-kernel

(there's also PKGBUILD for Arch that includes BMQ from Tk-Glitch
https://github.com/Tk-Glitch/PKGBUILDS/tree/master/linux54-tkg )

Actually, I didn't notice any improvement on fps, not sure what improvement should I expect since the original blog post is too technical for me, but I already knew that modifying the yield_tipe on the Muqqs Scheduler can improve performance in some wine games (Con Kolivas mentions that in the comments of the blog).

Hope that stadia devs improve the scheduler, it's really great that so many companies are looking to improve gaming performance on Linux these days.

Last edited by Koopacabras on 3 January 2020 at 3:02 am UTC
Koopacabras Jan 3, 2020
also check Philip Rebohle (valve dev) comments his saying that lowering kernel.sched_migration_cost_ns can improve performance on Ryzen Systems, also sched_min_granularity_ns.
Koopacabras Jan 3, 2020
a better scheduler is coming courtesy of the google stadia devs??

QuoteBut it’s not a fair mutex: If a thread takes a while to wake up, a faster thread can always sneak in and take the lock instead. It’s a very interesting model and I actually want to write something similar on Linux just to see how it performs.

Quotebut the Linux schedulers have problems. This was known for a while simply because audio can stutter on Linux when all cores are busy (which doesn’t happen on Windows) but it’s good to have benchmarks for this. It is really bad that a lock can sit idle for milliseconds even though somebody is trying to get it.
Now that game developers are slowly moving onto Linux, I predict that hiccups like that will go away within a couple years, as more work will be done on the scheduler.

Last edited by Koopacabras on 3 January 2020 at 3:41 am UTC
dvd Jan 3, 2020
There were some critiques of this article on it's comment section.
Koopacabras Jan 4, 2020
Quoting: dvdThere were some critiques of this article on it's comment section.

now I see what you mean I deleted my previous reply because it was out of place. There's some ppl saying that the benchmark is flawed and also that, if it's so like the benchmarks reflect... why then, in other benchmarks windows does so bad against linux; like specially in highly cpu bound scenarios like calculations and stuff like that. So the scheduler (that mostly manages the cpu) it's not doing that bad job as portrait really.

Last edited by Koopacabras on 4 January 2020 at 9:52 am UTC
Shmerl Jan 5, 2020
Follow up from Linus:

* https://www.realworldtech.com/forum/?threadid=189711&curpostid=189723
* https://www.realworldtech.com/forum/?threadid=189711&curpostid=189747
* https://www.realworldtech.com/forum/?threadid=189711&curpostid=189752
While you're here, please consider supporting GamingOnLinux on:

Reward Tiers: Patreon. Plain Donations: PayPal.

This ensures all of our main content remains totally free for everyone! Patreon supporters can also remove all adverts and sponsors! Supporting us helps bring good, fresh content. Without your continued support, we simply could not continue!

You can find even more ways to support us on this dedicated page any time. If you already are, thank you!
Login / Register

Or login with...
Sign in with Steam Sign in with Google
Social logins require cookies to stay logged in.
Register Forgot Login?
★ Support Us
Popular this week
Search or view by category
Contact
Latest Comments
Latest Forum Posts
Misc