1604570 - Enable low priority thread pool on Linux

Status: RESOLVED INVALID

(Core :: Graphics: WebRender, task, P3)

Description

[Bert Peers [:bpeers]]

The code was submitted; enable it by default so we can see Talos numbers for the combined result.

Comment 1

[Bert Peers [:bpeers]]

Attachment: Bug 1604570 - Enable low priority thread pool on Linux

Comment 2

[Pulsebot]

Pushed by [e-mail]:
https://hg.mozilla.org/integration/autoland/rev/d39132658a17
Enable low priority thread pool on Linux r=jrmuizel

Comment 3

[Oana Pop-Rus]

https://hg.mozilla.org/mozilla-central/rev/d39132658a17

Comment 4

[Alexandru Ionescu (needinfo me) [:alexandrui]]

Attachment: Backed out changeset d39132658a17 (Bug 1604570)

Comment 5

[Pulsebot]

Backout by [e-mail]:
https://hg.mozilla.org/integration/autoland/rev/a8197fa66cfd
Backed out changeset d39132658a17 as requested by Bert Peers for causing regressions.

Comment 6

[Bert Peers [:bpeers]]

At least on Windows, if I look at the affinity of everything else, then I don't see anything that's also insisting on being on core 0..7. So in theory the work that's causing ~3% slowdown could migrate to other cores and shouldn't hurt perf. So I'm still not sure what happened there. (Edit: perhaps >8 cores for Talos, and affinity prevents the pool from spreading out, causing -3% due to hyperthreading?)

As for the extra 15% slowdown when actually enabling the low priority pool... If that's really Linux saying "oh you wanted SCHED_BATCH threads, no problem" and they get 15% less CPU time, then... it's working great, I guess :) One option would be to semi-cheat and disable the pool on Talos: if Talos tests are for detecting regressions in raw SVG rasterization speed, then the throughput-for-latency trade-off is irrelevant there. Another option is to put the "low priority" pool back to normal priority, and instead boost the "normal" pool to higher-than-normal. In theory, with nothing running on the higher priority pool, that should keep perf level at +- minus 3% from affinity.

(Reminder that all of this is just a temporary workaround until we have a unified scheduler and/or yieldable SVG tasks.)

Comment 7

[Release mgmt bot [:sylvestre / :calixte / :marco for bugbug]]

There's a r+ patch which didn't land and no activity in this bug for 2 weeks.
:bpeers, could you have a look please?
For more information, please visit auto_nag documentation.

Comment 8

[Bert Peers [:bpeers]]

Invalid, no longer using low priority threads.

https://hg.mozilla.org/mozilla-central/rev/ec6587c0da30