747675 - Investigate if max forgetSkippable times could be decreased

Status: RESOLVED FIXED

(Core :: XPCOM, defect)

Description

[Olli Pettay [:smaug][[e-mail]]]

Attachment: patch

I'd like to try something like this. Basically split the slowest
forget skippable to two parts.
Would have to look at telemetry data to see how this affects to
max forgetSkippable times, and whether CC times are affected too.

Comment 1

[Olli Pettay [:smaug][[e-mail]]]

https://tbpl.mozilla.org/?tree=Try&rev=d1e73c1ce6d6

Comment 2

[Andrew McCreight [:mccr8]]

Comment on attachment 617253 [details] [diff] [review]
patch

Looks reasonable.

Are the max forget skippable times in telemetry actually bad?  Have you tried profiling to see how much each of these things you are splitting up takes?

Comment 3

[Olli Pettay [:smaug][[e-mail]]]

UnmarkGrayJSListenersInCCGenerationDocuments and xpc_UnmarkSkippableJSHolders tend
both to be somewhat slow, at least when I have lots of tabs open.
But I'll do some profiling again (haven't done any for few weeks).

Comment 4

[Olli Pettay [:smaug][[e-mail]]]

https://hg.mozilla.org/mozilla-central/rev/4b6514ffcf16

Comment 5

[Olli Pettay [:smaug][[e-mail]]]

Backed out in order to investigate Bug 749461
https://hg.mozilla.org/mozilla-central/rev/0aa393eef856

Comment 6

[Olli Pettay [:smaug][[e-mail]]]

Based on the initial telemetry data, the patch doesn't affect too much to the
already low median times, but helps significantly with the worst 5%, so mean time decreases quite nicely.

Comment 7

[Olli Pettay [:smaug][[e-mail]]]

https://hg.mozilla.org/mozilla-central/rev/f622f1ccef2b