1443066 - Try to understand why we're slower than Chrome on StyleBench "Positional pseudo classes : Adding leaf elements: Async" (& similar "Removing" measurement)

Status: NEW

(Core :: CSS Parsing and Computation, defect, P3)

Description

[Daniel Holbert [:dholbert]]

See StyleBench results in bug 1425058 Comment 26 -- we're slower than Chrome on one of the larger sets of measurements there:

Stylo with privacy.reduceTimerPrecision=false and STYLO_THREADS=1 (revision: https://hg.mozilla.org/mozilla-central/rev/190b536928f8)
Positional pseudo classes : Adding leaf elements : Async: 125.42000000000007 ms
Positional pseudo classes : Removing leaf elements : Async: 120.69999999999982 ms

Chromium (revision: https://chromium.googlesource.com/chromium/src/+/a8768c2cd0aea923b6ee86077e3b608518b44e2d)
Positional pseudo classes : Adding leaf elements : Async: 99.50000000026193 ms
Positional pseudo classes : Removing leaf elements : Async: 95.5999999996493 ms

For both measurements, our time is about 26% longer than Chrome's. This seems like the main piece of StyleBench where we're significantly behind, at the moment.  Hence, I'm spinning this bug off to cover investigating these metrics & seeing if there are any wins we can get here.

Comment 1

[Xidorn Quan [:xidorn] UTC+10]

Optimization of positional pseudo-classes is generally just the NthIndexCache, and IIRC we have a different cache strategy than Chrome. It is possible that their cache strategy is better than ours.

Comment 2

[Emilio Cobos Álvarez [:emilio]]

Well, right now our handling of dom mutations on something that has :nth-child is just restyling the whole subtree, which is... not great.

Comment 3

[Emilio Cobos Álvarez [:emilio]]

I dumped my thoughts on how to fix this in bug 1406622 comment 2.