Performance is worse than upstream Firefox due to Mercury using ESR edition

[GrabbenD]

Problem

Absolutely love the work you're doing here and all the patches are awesome!

However, the problem with Mercury is that it uses ESR edition of Firefox which lags behind in terms of updates.. Just using a newer version of Firefox gives better performance than what Mercury can offer with all the performance enhancements:
https://www.phoronix.com/review/mercury-firefox-perf

Furthermore there's more performance to be gained:

• Arch Linux users can install firefox-vaapi (for better hardware acceleration support. It has to be compiled manually and by default /etc/makepkg.conf uses LTO as well as -march=native -O3 performance optimizations out of the box. PGO can easily be enabled as well).
• Betterfox has privacy and performance optimized profiles

Solution?

Providing upstream releases with Betterfox profile could easily overtake every other fork of Firefox in terms of privacy and performance, especially with the CPU optimizations which Mercury currently offers.

[32803]

I noticed this aswell, thought I was going insane when my own hardened firefox was beating mercury in every test.

[catbabylon]

I'm still finding with latest v121 compared to vanilla Firefox running a fresh profile is getting approx 150 on speedometer 2 compared to approx 140 on a fresh profile on Mercury (with uBo disabled so the extension doesn't affect results), I'm on a Dell XPS 9300.

Is it that some of the privacy hardening patches are more demanding than the gains made from optimization of the code? Might test against Librewolf...

EDIT...

Realised I was using speedometer 2.0 at least some of the time so started again and compared between Firefox Mercury and Librewolf

Librefox was getting approx 130, Mercury approx 150 and Firefox approx 160

So I guess this suggests some of the privacy stuff Librefox devs add are more demanding, but the optimisations on Mercury help mitigate them...

[GrabbenD]

@gz83 @Alex313031 Issue closed without any reason? 🙁

[gz83]

This is a batch closing operation (can also be regarded as a ping), I will reopen the issues that still have new responses