search

Alex313031 / Mercury

Firefox fork with compiler optimizations and patches from Librewolf, Waterfox, and GNU IceCat.
https://thorium.rocks/mercury
Mozilla Public License 2.0
1.06k stars 25 forks source link

Speedometer Performance Regression #13

Open gianni-rosato opened 1 year ago

gianni-rosato commented 1 year ago

Hello, I am experiencing bad Speedometer performance compared to standard Firefox on the latest Mercury 111 from the AUR.

Mercury Firefox

➜ ~ inxi -v CPU: 16-core (8-mt/8-st) 13th Gen Intel Core i7-13700K (-MST AMCP-) speed/min/max: 1450/800/5400 MHz Kernel: 6.2.8-zen1-1-zen x86_64 Up: 17m Mem: 4548.1/31868.8 MiB (14.3%) Storage: 6.83 TiB (22.7% used) Procs: 499 Shell: Zsh inxi: 3.3.26

Troubleshooting Information (about:support)

Let me know if you need any more information. Thank you!

Alex313031 commented 1 year ago

@gianni-rosato Hmmmm. That's interesting, because this release is based on stable firefox, just with compiler optimizations, so it should be more performant, not less.

gianni-rosato commented 1 year ago

Is there a particular compiler flag that could change this? Should I try building from source?

BlueSwordM commented 1 year ago

Very weird, but I'm getting the same speedometer numbers as you do, with a much weaker CPU (3700X).

gianni-rosato commented 1 year ago

Blue also told me they tested various other benchmarks & didn't observe any differences. @Alex313031, can you try to replicate what we're seeing here?

sandboiii commented 1 year ago

Tested the newest v.112.0.3 build on my laptop. Seems to be much better. Firefox - 129 Mercury - 135

sandboiii commented 1 year ago

v115.0.0 Firefox - 136 Mercury - 157

EmberJS tests seem to be slow, looks like on Firefox they pass faster

sandboiii commented 10 months ago

v115.3.0 Firefox - 153 Mercury - 159

Looks like they are almost the same speed. I guess this issue can be closed now

afonsofrancof commented 9 months ago

Having the same issue. Firefox 118.0.2 Mercury 115,4,0 Windows 10 image

gz83 commented 9 months ago

There are differences between different versions and systems, and the performance test is for reference only. In addition, the best test is to compare versions with the same version number.