[meta] Firefox Performance Adjustment - set up a mechanism for various parts of Firefox to adapt based on machine performance characteristics

ASSIGNED
Assigned to

Status

()

P1
normal
ASSIGNED
7 months ago
16 days ago

People

(Reporter: Gijs, Assigned: Gijs)

Tracking

(Depends on: 2 bugs, {meta})

Trunk
Points:
---
Dependency tree / graph

Firefox Tracking Flags

(firefox63 affected)

Details

(Assignee)

Description

7 months ago
Goal: end up with a simple service/component for different parts of Firefox to cheaply get information about performance characteristics of the machine, so they can adapt accordingly. Bootstrap/proof-of-concept 1 or 2 things into that mechanism.

A number of considerations apply here, and it's explicitly not the goal of this metabug to cover every single one of them, nor is it the intention of this metabug to come up with an exhaustive list of every part of Firefox that we *could* change based on machine performance.

Notable points that we *will* address as part of this project (one way or another):

- do our best (given time and other constraints) to be data-driven when coming up with a way of classifying machine performance
- ensure that determining machine performance "in the field" doesn't require a "live" performance test run on the machine (a la Windows Experience Index) and that determining the machine performance doesn't otherwise negatively impact performance.
- have a possibility for user override/choice. This also helps with testing / reproducing issues.
- somehow take privacy.resistFingerprinting into account (it's not really clear we should always do this, e.g. for UI animations in browser chrome we probably don't need to, for things like GC behavior that are web-exposed, we do)
- once we have a solid basis for classifying machine performance, consider how/whether this can be applied on mobile
(Assignee)

Updated

4 months ago
Depends on: 1503339
(Assignee)

Updated

16 days ago
Depends on: 1524299
You need to log in before you can comment on or make changes to this bug.