Benchmarks Show Firefox 57 Quantum Doing Well, But Chrome Largely Winning

[profoundWHALE]

profoundWHALE Michael was very fair to firefox, even tho it's fans (and I am one of them(fans)) think he wasn't.

Are you talking to the right person? I just showed firefox did worse because despite being faster than chrome a few times, it was slower almost every other time by fairly large margins.

[labyrinth153]

Ran some benchmarks of Safari (Mac OS Sierra, fully updated) and Firefox Beta. System is iMac 27" 2015.

sunspider

firefox: 167.6ms +/- 5.1%
safari: 112.1ms +/- 5.4%

kraken

firefox: 812.4ms +/- 3.8%
Safari: 724.1ms +/- 1.6%

html5 test (not a performance benchmark)

firefox: 483
safari: 452

speedometer

firefox: 101
Safari: 163

[leipero]

Maybe the difference is just that I'm using ublock with FF57 then? No ads, so no lag? I haven't turned it off since yesterday, and its still ripping along at the same frantic pace. Chrome would normally be completely bogged down by now after remaining on for 24 hours.

Nah, I am not having lag problems with Firefox, in fact it works better (without ad blocked) than Chromium with it (also using micro block origin) when pages are still loading, when pages are loaded it's basically the same.

[pmorph]

Most important thing is that firefox developers aren't complacent about performance. If it keeps improving I'm fine, be it baby steps or bigger.

[Death Knight]

Year: 2017
Mozilla still not started to implement Linux HW video accelerators to their browser, Firefox. And they look like not really care of it indeed.

https://bugzilla.mozilla.org/show_bug.cgi?id=1210727

https://bugzilla.mozilla.org/show_bug.cgi?id=1210729

I am angry to mozilla due this.

[mmstick]

According to this CSS Benchmark, Chrome is basically no competition at all for Firefox 57:

Firefox (6.5 seconds): https://streamable.com/bbrwf
Chrome (16 seconds): https://streamable.com/9n7ds

[andyprough]

andyprough That's true, and as I said, Firefox (at leas on hardware I've tested) tends to load pages and allows you to scroll the page while loading without any stutter, so even if it's waiting on ads (the only thing that loads slower on pages) it doesn't ruin your browsing experience, Chromium/Chrome on the other hand does. However, WebGL performance of Firefox are just terrible, there is no way around it, same goes for javascript.

Maybe the difference is just that I'm using ublock with FF57 then? No ads, so no lag? I haven't turned it off since yesterday, and its still ripping along at the same frantic pace. Chrome would normally be completely bogged down by now after remaining on for 24 hours.

[leipero]

andyprough That's true, and as I said, Firefox (at leas on hardware I've tested) tends to load pages and allows you to scroll the page while loading without any stutter, so even if it's waiting on ads (the only thing that loads slower on pages) it doesn't ruin your browsing experience, Chromium/Chrome on the other hand does. However, WebGL performance of Firefox are just terrible, there is no way around it, same goes for javascript.

tomtomme Well, ofc. I know, it says in the title and on graphs, my point wasn't that, my main point was that Michael tests are far away from the "worst case" scenario. I did not test v57 because I use v55 from repositories, and I know from the last time I checked that performance improvements are not worth going outside of distribution repository, and aesthetic are not improved enough to make me switch (especially with broken extensions, while I could use "headerbar" extension for v55 and Gnome 3 theme and it looks much better than v57 by default). But, I've tested it for fun now, and here are the results in same conditions as those before:

Motionmark: C 259, F 115 (v57=117), Chromium 125% faster.
Basemark: C 470, F 316 (v57=322), Chromium 49% faster.
ARES-6: C 54ms, F 174ms (v57=135ms), Chromium 222% faster (150% faster than v57).
Speedometer: C 83, F 45 (v57=55), Chromium 84% faster (51% faster than v57).
Jetstream: C 101, F 103 (v57=105), Firefox 2% faster (v57 ~ 4% faster than C)
Octane: C 17447, F 16231 (v57=17379), Chromium 7% faster (about equal to v57).
Kraken: C 1946ms, F 1899ms (v57=1847ms), Firefox 2% faster (v57 ~ 5% faster than C).

As you can see there is no huge difference, tho i have to admit that loading time for browser is quite better, while v57 still suffers from "Slowing down pages" under some benchmarks, and it is possible that that thing alone makes scores much worse that what they should be, except WebGL ofc.

[tomtomme]

profoundWHALE Michael was very fair to firefox, even tho it's fans (and I am one of them(fans)) think he wasn't (because of javascripts etc.). He called 2.6-3.2GHz quoad core i7 a "modest CPU", and in reality it is, but look at desktop browser resolutions, and you would see that 1080p is not even 20%, while dominant resolution is still 1366x768, while "other high" usually doesn't mean >1080p. With that on mind, those PC's are from "casuals", and they usually run Intel dual cores (with HT at best), and those PC's main purpose is web browsing. With all that on mind, it is actually unfair to test browsers on even "mid range" CPU's, let alone high end,a nd since msot of the web is plagued with javascripts, it is logical to test for it the most.

With that said, I've done extensive testing on really low end hardware (stock FX-4100 with turbo disabled), and results are not better for Firefox, in fact they are much worse. One note is that I've used:

layers.acceleration.force-enabled = true
browser.tabs.remote.force-enable = true

for testing, but that should only improve performance (in Firefox 55). Another note is that resolution used is comparable to "1440x900", so it is not 1080p or 720p. Also, I've forced DRI2, results would be probably better in some areas, but still, it would be better for both...

So results of Chromium (v61) vs Firefox (v55) are:

Motionmark: C 259, F 115, Chromium 125% faster.
Basemark: C 470, F 316, Chromium 49% faster.
ARES-6: C 54ms, F 174ms, Chromium 222% faster.
Speedometer: C 83, F 45, Chromium 84% faster.
Jetstream: C 101, F 103, Firefox 2% faster.
Octane: C 17447, F 16231, Chromium 7% faster.

They speak for themselves, and even tho I am Firefox fan, msot of the time I do use Chromium, and you can see why. But it is not all bad for Firefox, on some pages (like Level1Techs forums) Firefox actually gives smoother experience, and especially with DRI3 and TearFree option enabled you can spot the difference, even on light websites such as youtube or phoronix. For example, when page is still loading (here on phoronix, adds tend to be terrible slow to load sometimes/most of the time) in Chromium, when I scroll down the page there is stutter, when page is fully loaded everything is smooth and fine, while on Firefox (and especially with TearFree) it's super smooth regardless if page is loaded or not, same goes for youtube or any other site. So, browsing experience (aside from loading slower, and javascripts) could be better on Firefox.

TL;DR: Michael should test it on his Pentiums, Celerons and his Athlon x3, adn Firefox developers too .

uhm - you know that all the buzz here is about firefox 57 and not 55?

[Kendji]

Here's the question more that of which browser is good enough or if both are, after that privacy, open-sourceness etc. plays a factor