Browsers Benchmark tests

[Guest]

We each choose our preferred (default) browser based on many personal parameters.

In my case, the first parameters to consider are:

• Usability
• Performances

My next choice is:

• Security/Privacy

It is not easy to determine which browser is the fastest among those we prefer.
I modified Firefox to be faster.
In fact on the “Speedometer 3.0” test it comes out faster than my Edge.

But this higher apparent speed is not true.

The rendering of media content is considerably slower than my Edge.
And this is highlighted well on the test below:

https://browserbench.org/MotionMark1.3.1/
 

I only know of one test to check the security of the browser.
Unfortunately, it is an old test and one that currently shows only inherent security, so to speak.
 

https://browseraudit.com/

A comparison of the results to the tests in my 2 installed browsers highlights that Edge (primary) is faster and safer than Firefox(secondary). 

Consider that each test should be taken for what it represents,just a test.

_____________________________________________________________________________________________________________________________________________________

If other forum members would like to recommend other tests they are of course welcome.

Edited December 2, 2024 by Sampei.Nihira

[NotHereToPlayGames]

23 minutes ago, Sampei.Nihira said:

In fact on the “Speedometer 3.0” test it comes out faster than my Edge.

But this higher apparent speed is not true.

What about testing in Speedometer 2.1?  I've always tested using 2.1 largely in part to a very long HISTORY of using it and all scores LOGGED in a spreadsheet.

I have several other benchmark "tests" that are also logged.  I'll supply them this afternoon.

[Guest]

14 minutes ago, NotHereToPlayGames said:

What about testing in Speedometer 2.1?  I've always tested using 2.1 largely in part to a very long HISTORY of using it and all scores LOGGED in a spreadsheet.

I have several other benchmark "tests" that are also logged.  I'll supply them this afternoon.

Absolute values change.
Firefox (125) is always falsely faster than Edge (73.31).
So my changes to Firefox make this test useless,obviously when comparing multiple browsers + Firefox.

[NotHereToPlayGames]

Here are several that I have used in the past.

 

Fast tests that don't drive my impatience through the roof -

https://mozilla.github.io/krakenbenchmark.mozilla.org/kraken-1.1/driver.html

http://proofcafe.org/jsx-bench/js/sunspider-1.0/driver.html

https://browserbench.org/Speedometer2.1/

 

Slow tests that I do not have the patience to run very often -

https://web.basemark.com/

https://browserbench.org/JetStream/

https://browserbench.org/MotionMark1.3.1/

[NotHereToPlayGames]

Ungoogled Chromium v122
   Speedometer 2.1  ==  152
   Sunspider  ==  228.1
   Kraken  ==  14,015.5
   DOMContentLoaded (YouTube)  ==  1.89 seconds

IceCat v115.17.0
   Speedometer 2.1  ==  96.1
   Sunspider  ==  93.2  [faster than UG v122]
   Kraken  ==  1,420.5  [faster than UG v122]
   DOMContentLoaded (YouTube)  ==  2.31 seconds

[UCyborg]

18 hours ago, Sampei.Nihira said:

The rendering of media content is considerably slower than my Edge.

What about videos specifically? Chromium was always a bit stuttery for me in that department.

Something to bring any browser to its knees, port of id Tech 4 running Doom 3 demo.

http://www.continuation-labs.com/projects/d3wasm/

But the demo version doesn't include the short demo file (gameplay recording), which can be played back with timedemo command, that runs it as fast as possible and outputs the average frame rate.

 

[NotHereToPlayGames]

43 minutes ago, UCyborg said:

Chromium was always a bit stuttery for me in that department.

Be sure to compare the exact same codec when comparing videos.

I'm hit or miss in video department.  I can get Chromium to be better, but Mozilla IS better "by defaults".

Video playback is also a gigantic BALANCE.  You can get awesome quality in Mozilla *if* you WANT the CPU to churn at 80% the entire span of the video.

[I should add that I can *not* use Hardware Acceleration.  My needs require being able to capture video-conferencing via "print screen" which does not work if HWA is enabled.]

Edited December 3, 2024 by NotHereToPlayGames

[NotHereToPlayGames]

Here's another collection of tests  --  https://dromaeo.com/

[UCyborg]

18 minutes ago, NotHereToPlayGames said:

I'm hit or miss in video department.  I can get Chromium to be better, but Mozilla IS better "by defaults".

Pretty sure I tried both VP9 and H.264, but it's been a while.

[Guest]

2 hours ago, UCyborg said:

What about videos specifically? Chromium was always a bit stuttery for me in that department.

Something to bring any browser to its knees, port of id Tech 4 running Doom 3 demo.

http://www.continuation-labs.com/projects/d3wasm/

But the demo version doesn't include the short demo file (gameplay recording), which can be played back with timedemo command, that runs it as fast as possible and outputs the average frame rate.

 

Not only video.
If a web page is full of multimedia content,even just GIF images you will notice a slowdown in Firefox compared to Edge.
Slowdown that obviously extends to the entire web page.
As I have already written this is well evident in the MotionMark test:

Edge= 703.61
Firefox= 438.03

So,in my opinion,multiple tests contribute to defining which browser, among those we prefer, is the fastest.

[Guest]

3 hours ago, NotHereToPlayGames said:

Be sure to compare the exact same codec when comparing videos.

I'm hit or miss in video department.  I can get Chromium to be better, but Mozilla IS better "by defaults".

Video playback is also a gigantic BALANCE.  You can get awesome quality in Mozilla *if* you WANT the CPU to churn at 80% the entire span of the video.

[I should add that I can *not* use Hardware Acceleration.  My needs require being able to capture video-conferencing via "print screen" which does not work if HWA is enabled.]

Yes, Hardware Acceleration can make a difference.

Maybe even the Just-in-Time (JIT) compiler.
I have never done this verification.

In my Edge JIT is disabled,I use the “strict” setting for security reasons:

https://microsoftedge.github.io/edgevr/posts/Super-Duper-Secure-Mode/

Turning it on would probably make my Edge more responsive.

 

In Firefox JIT is enabled.
And I have never proceeded to disable it.

Probably in this old article (with conclusion that will still be valid today?) you can see the difference:

https://v8.dev/blog/jitless

Edited December 3, 2024 by Sampei.Nihira

[j7n]

Well, New Moon won't win any special olympics. I use it because I like the traditional GUI.

Hardware acceleration in New Moon has problems for me: 1) font weight increases, 2) it seems to require new video adapter and drivers, 3) vsync for some reason gets disabled in other programs. Outputting a text doesn't require acceleration, nor should it need particularly new hardware. The bottleneck is the javascript almost always. Very big and simple pages like old reddit with 100 posts feel a tad slower.

It's a bit strange that print screen doesn't work. In media players it tends to always work nowadays, since we no longer use Overlay Mixer. Maybe it is crippled intentionaly for those movie playback sites.

[Guest]

1 hour ago, j7n said:

Well, New Moon won't win any special olympics. I use it because I like the traditional GUI.

Hardware acceleration in New Moon has problems for me: 1) font weight increases, 2) it seems to require new video adapter and drivers, 3) vsync for some reason gets disabled in other programs. Outputting a text doesn't require acceleration, nor should it need particularly new hardware. The bottleneck is the javascript almost always. Very big and simple pages like old reddit with 100 posts feel a tad slower.

It's a bit strange that print screen doesn't work. In media players it tends to always work nowadays, since we no longer use Overlay Mixer. Maybe it is crippled intentionaly for those movie playback sites.

There is a possible remedy.
Use dynamic filtering in uBO Legacy.
Start with “Medium Mode.”

https://github.com/gorhill/uBlock/wiki/Blocking-mode:-medium-mode

https://github.com/gorhill/uBlock/wiki/Dynamic-filtering:-quick-guide

Almost always you will have to allow only “3rd-party-scripts” rarely will blocking “3rd-party-frames” cause display problems.

Web page loading speed will be increased:

https://github.com/gorhill/uBlock/wiki/Blocking-mode

 

Edited December 3, 2024 by Sampei.Nihira

[NotHereToPlayGames]

16 hours ago, NotHereToPlayGames said:

Ungoogled Chromium v122
   Speedometer 2.1  ==  152
   Sunspider  ==  228.1
   Kraken  ==  14,015.5

The above scores were with JIT disabled (Chrome/Chromium default).

These are my scores with JIT enabled (which I don't mind the "security risk" this entails but that's because I have other mitigations).

Speedometer 2.1  ==  201  [with or without JIT, Ungoogled already scores higher than IceCat]
Sunspider  ==  83.7  [faster than IceCat]
Kraken  ==  840.0  [faster than IceCat]

To be honest, I forgot all about JIT performance gains.  I feel like my Ungoogled v122 just had its 2.2L K-car Era four-banger swapped out with a 392 SRT 6.4L Hemi V8.

Edited December 3, 2024 by NotHereToPlayGames

[NotHereToPlayGames]

How about providing about:config fixes for Firefox forks to improve performance and I'll even go so far as to REDUCE/ELIMINATE *any-and-ALL* "security/privacy" 'feature' ???

I want my browser to be configured for SPEED and PERFORMANCE.  I will set my own privacy/security mitigations !!!

My Ungoogled Chromium v122 has officially hit the end of the line for me to be able to use as my default.  

ie, my bank login (and US Post Office login) now both no longer work in *ANY* version of Ungoogled Chromium.  

This login issue is also an issue for Brave browser users that also use the same bank.  All signs indicate SERVER-SIDE blacklisting of "non-mainstream" browsers and they finally closed the hole for v122, v123, and v124.  

 

So I am in-process of configuring ICECAT as my default use-for-everything browser (I really am "too old" to have one browser for this, another browser for that, still another for this, still another for that).

So what about:config settings can I set that will make IceCat FASTER than Ungoogled v122 - EVEN IF IT MAKES IT "less secure" !!!

I want FASTER, FASTER, FASTER.  I shall define for myself what "security" or "privacy" BOTTLENECKS to enable.

 

Thanks in advance for any suggestions.