Firefox 73 + Firefox 74 Beta Benchmarks On Ubuntu Linux

[phoronix]

Phoronix: Firefox 73 + Firefox 74 Beta Benchmarks On Ubuntu Linux

Given this week's release of Firefox 73 stable that also puts Firefox 74 into beta state, here are fresh Firefox browser benchmarks of Firefox 72/73/74 on Ubuntu Linux with and without WebRender as well as how it compares to the current state of Google Chrome.

http://www.phoronix.com/vr.php?view=28875

[skeevy420]

Firefox might have its ass handed to it going by most of those results, but, to me anyways, I don't notice a difference between the two doing the usual day to day internet activities.

[fagnerln]

The biggest difference between the two in my pc is that Firefox uses much more resources than Chromium in videos on youtube or twitch, the cpu/gpu temperatures increases on Firefox.

Maybe would be interesting to add cpu utilization/temp on next benchmark

[kozman]

The biggest difference between the two in my pc is that Firefox uses much more resources than Chromium in videos on youtube or twitch, the cpu/gpu temperatures increases on Firefox.

Maybe would be interesting to add cpu utilization/temp on next benchmark

Probably related to https://bugzilla.mozilla.org/show_bug.cgi?id=1610193 but the fix is not yet in 74.0b. Should be soon though.

[kozman]

Phoronix: Firefox 73 + Firefox 74 Beta Benchmarks On Ubuntu Linux

Given this week's release of Firefox 73 stable that also puts Firefox 74 into beta state, here are fresh Firefox browser benchmarks of Firefox 72/73/74 on Ubuntu Linux with and without WebRender as well as how it compares to the current state of Google Chrome.

http://www.phoronix.com/vr.php?view=28875

Ouch. Those numbers show some pretty big perf disparities and to see some negative trends even (Motionmark) is not encouraging. Although these early 74 betas haven't yet pulled in newer Rust 1.41, it might help close the perf delta or at least not trend negative. But as Skeevy420 said, in regular real world use, most users will not notice much. FF76 for Android cannot come soon enough. 68.x on Android is just painful with more than 5 tabs open. Crash city.

[kozman]

Well, looks like Rust 1.41 is targeting FF75 (https://bugzilla.mozilla.org/show_bug.cgi?id=1605188). Pity. Opportunity missed.

[andyprough]

The biggest difference between the two in my pc is that Firefox uses much more resources than Chromium in videos on youtube or twitch, the cpu/gpu temperatures increases on Firefox.

Maybe would be interesting to add cpu utilization/temp on next benchmark

Debian-based distros (not Ubuntu-based) tend to have excellent resource management with videos on Firefox. Especially two Debian-based distros without systemd for some reason - antiX and MX. And Arch-based Artix does not have systemd and has outstanding resource management with Firefox. I doubt that it has anything to do with systemd, probably other optimizations, but the difference is quite noticeable.

Also might try watching videos through invidio.us rather than through the youtube site. Youtube itself seems poorly designed, whereas the same videos are available through invidio.us which has a very simple layout.

[smitty3268]

Ouch. Those numbers show some pretty big perf disparities and to see some negative trends even (Motionmark) is not encouraging. Although these early 74 betas haven't yet pulled in newer Rust 1.41, it might help close the perf delta or at least not trend negative.

Most of those tests are just javascript based, which isn't going to be affected by Rust one way or another. Javascript and probably some DOM based testing as well, which again - no Rust code. Stylebench is presumably the big one that is using Rust code, and Firefox already has a big lead in it.

[enigmaxg2]

Probably related to https://bugzilla.mozilla.org/show_bug.cgi?id=1610193 but the fix is not yet in 74.0b. Should be soon though.

FF 75 targeted... need to wait 2 months to solve a memory leak on one of the most visited sites...are you kidding me?

I experienced crazy memory usage when doing a speedtest or while watching long twitch streams, others found leaks on deezer... bugs were filed. 4 months later and still no word from Mozilla... what a joke.

[TemplarGR]

So, Google Chrome is still faster in some (not all) javascript benchmarks.... So? A web page is not just pure Javascript code, you know. And unless your plan is to execute software and games in your browser, pure javascript performance is not the end of benchmarking. Personally i don't care about javascript that much, it is mostly used for advertisements (that i filter anyway).