A Quick Benchmark Of Mozilla Firefox With WebRender Beta vs. Chrome

[bwat47]

I see in the screenshot that your azure backend is not accelerated. That's also due to some rather weird GPU blacklisting. Set these keys in about:config:

layers.acceleration.force-enabled true
gfx.content.azure.accelerated true
gfx.canvas.azure.accelerated true

Without it, benchmarking is rather pointless.

Firefox Linux releases should really stop this sweeping blacklisting and should enable all this by default. Drivers situation today is very different from five years ago.

Yeah I've been running with all this enabled for like a year now on intel graphics with no issues. It's really annoying that AT THE VERY LEAST they haven't enabled hardware compositing by default

[DanL]

Say no to the Firefox browser that requires you to use pulseaudio.

You could use apulse emulation for Firefox (already packaged in Debian repo) and get on with your life.
Or you could keep on trolling Phoronix with the same, tired rhetoric for years on end.

Hmmm. I wonder which one you'll choose.

[tomtomme]

I see in the screenshot that your azure backend is not accelerated. That's also due to some rather weird GPU blacklisting. Set these keys in about:config:

layers.acceleration.force-enabled true
gfx.content.azure.accelerated true
gfx.canvas.azure.accelerated true

Without it, benchmarking is rather pointless.

Firefox Linux releases should really stop this sweeping blacklisting and should enable all this by default. Drivers situation today is very different from five years ago.

is this for nightly? my about config only has
gfx.canvas.azure.backends skia
gfx.content.azure.backends skia

as options

[cj.wijtmans]

In a real life scenario, the difference is a page loading in 0.1 or 0.2 seconds. Not really something a human being can perceive. It's the same crap as operating systems competing on boot time. Your OS booting in 6 or 7 seconds changes absolutely nothing in your life.

Anyone who cant percieve 0.1 seconds ..... is a bit slow mentally... you should even be able to make a complete sound in 0.1 seconds.

[carewolf]

Anyone who cant percieve 0.1 seconds ..... is a bit slow mentally... you should even be able to make a complete sound in 0.1 seconds.

Yeah, but if loading the page takes a few seconds beause it is bloated crap (which anything taking more than 0.1 seconds to render is), then you won't preceive whether it takes 0.1s extra to render it.

[carewolf]

Damn, it is noticeable in daily use but I still didn't think Blink is that much faster. Those graphs are insane.

Chrome is very good at benchmarks. Its major drawbacks are overhead and bloat, but they have the superior numbers in engineers, so they optimize any metric they can find, except a smaller codebase or binary.

[Weasel]

Never mind that rewriting the CSS engine in Rust made it twice as fast as the previous version. It doesn't matter, does it. You can't fight fanboyism with mere facts.

No, it doesn't. Mozilla devs obviously wrote it in something that would be called incompetent++ rather than C++. I mean, the fact they rewrote it in Rust kind of proves this by itself.

Nobody sane would rewrite real C++ into anything else other than maybe C, especially not something as crap as Rust.

Also this is not a comparison between Firefox and pre-Rust Firefox you know.

[Weasel]

He is very obviously baiting and everyone is just taking it, sometimes saying illogical things in response. This is not a serious poster. You don't need to respond to him.

Maybe, but I'm definitely not trolling more than Rust advocates in other threads so keep it in mind.

[Weasel]

You have just answered the question: What's worse than a Rust zealot?

But it's true? No instructions will always use less power than extra instructions.

A stupid bounds check when it's not needed (which is 99.999999% of the time since otherwise the app would crash 24/7) is just that: extra instructions to process for the CPU, all the time. It doesn't matter if they're done in parallel or whatever, the energy wasted goes up even if you do it in parallel and doesn't affect "performance". This is basic physics.

Which means more energy wasted, and thus, causing more global warming than no bounds checks.

It's easy to label people as trolls when you have no arguments and know it.

[Weasel]

Not really. Rust is basically just C++ best practises

C++ best practices is not real C++. Real C++ is just C with extra features (even the name says it all). You code with C mindset but with extra features. That's real C++.

"Best practise C++" is what I called incompetent++, only for incompetent morons who need "runtime safety checks" because they're too incompetent so they make everyone's CPUs (users) process more crap and waste more energy globally.