Question How fast is your browser?

[Magic Carpet]

This benchmark simulates user actions for adding, completing, and removing to-do items using multiple examples in TodoMVC. Each example in TodoMVC implements the same todo application using DOM APIs in different ways. Some call DOM APIs directly from ECMAScript 5 (ES5), ECMASCript 2015 (ES6), ES6 transpiled to ES5, and Elm transpiled to ES5. Others use one of eleven popular JavaScript frameworks: React, React with Redux, Ember.js, Backbone.js, AngularJS, (new) Angular, Vue.js, jQuery, Preact, Inferno, and Flight. Many of these frameworks are used on the most popular websites in the world, such as Facebook and Twitter. The performance of these types of operations depends on the speed of the DOM APIs, the JavaScript engine, CSS style resolution, layout, and other technologies.

Speedometer 2.0

Core M-5Y10c @ 52.02

Images aren't necessary, but please state your cpu speed. The web browser of your choice. Mine is Edge 83. Thank you.

 

[GodisanAtheist]

Firefox ??? On a Pixel 4 5g

 

[moinmoin]

I am also curious as to why this is in the CPU forum when it is more GPU oriented?

It's not. While different browsers do have different GPU acceleration for painting content, all of them have to be able to fall back to software solutions to some degree depending on hardware, version and OS.

 

[TurtleCrusher]

311 on speed 2.0

 

[winr]

Core i 2600 .... 148

eege Browser and youtube playing heavy rain in the background

 

[Harry_Wild]

Much of the speed is determine by ones internet connection. I have only a 500mbp download speed fiber line. When I go to big tech corporation that equipped with a 30 terabyte per second line and use their ethernet line, it like night and day! Everything is instantaneous!

 

[Makaveli]

Much of the speed is determine by ones internet connection. I have only a 500mbp download speed fiber line. When I go to big tech corporation that equipped with a 30 terabyte per second line and use their ethernet line, it like night and day! Everything is instantaneous!

I don't believe your internet speed matters for this benchmark.

i'm on 1.5 Gbps down and 1 Gbps up fiber connection and i don't see a difference if I test on a different connection.

 

[Eug]

iPad Pro 11” M4 (binned 9-core), 8 GB RAM, Safari on iPadOS 17.5.1

Speedometer 2.0:



Speedometer 3.0:

 

[Eug]

I didn't bother running Chrome on the iPad Pro since I thought Chrome still used the same WebKit core as Safari on iOS/iPadOS. However, I see that as of iPadOS 17.4, third party browsers are now allowed to use alternative browser engines. I'm on iPadOS 17.5.1. Has Chrome changed to its own yet outside of the EU, or is it still using WebKit? The article below says it's only in the EU.

Apple is finally allowing full versions of Chrome and Firefox to run on the iPhone

The EU browser wars are about to take off in a big way.

www.theverge.com

 

[SPBHM]

this is where I do a lot of my web browsing! android 9 + firefox + snapdragon 636!

my n3350 mini pc on win 10 + chrome achieved 2.35!!!
and the Q6600@3Ghz on win10 chrome x86 5.64

 

[Eug]

Speedometer 2.0 on a 2 GB iPad Air 2 running iPadOS 15.8.2.

Safari:



Chrome 125:



I did post this message from this iPad, but otherwise I cannot comfortably surf on this machine. It’s just way too slow. And it apparently is starting to have compatibility problems. Neither browser could complete the Speedometer 3.0 bench. (Both run WebKit.)

 

[whoshere]

JavaScript based online benchmarks are invariably useless because web browsers have different JS engines and performance, not to mention that Mozilla has been dragging its feet in this department and Chrome has been leap frogging: https://v8.dev/blog

Secondly, even if we all decide to use the same web browser and its version, this particular benchmark also depends on the OS and even your GPU drivers.

With that, here are my results for Ryzen 7 7840HS under Fedora 40 running in Google Chrome 126.0.6478.126:

 

[AdamK47]

The speed of a browser can be measured using benchmarks like Speedometer 2.0, which simulates user interactions to evaluate performance. Different browsers may excel in various areas depending on the JavaScript engine, CSS rendering, and DOM handling capabilities. To assess your browser's speed, run these benchmarks and compare the scores with those of other popular browsers.

4 years and 14 pages in and we finally have an explanation of what this benchmark does. We've been running this thing to "make number go higher" without any clue as to its purpose. Thank you.

 

[winr]

8.93 using the new one (3.0 )

Core i7 2600 3.4 gig... Old Intel work station type mobo ...32 gig DDR 3 ... MSI AERO 1070 Ti video ... SATA drive ....Firefox


Ricky.

 

[Det0x]

Zen5 + Chrome

Speedometer 3.0:

Not tryhard, just daily 24/7 settings

 

[Hitman928]

Zen5 + Chrome

Speedometer 3.0:
View attachment 105020
Not tryhard, just daily 24/7 settings

SMT on or off?

 

[Det0x]

SMT on or off?

SMT enabled. Maybe i should try with disabled after what Techpowerup just posted..
Just have to finish my Karhu run first, testing some new mem settings

 

[Det0x]

Alrighthy, did some more testing in light the Zen5 SMT enabled/disable review posted at Techpowerup
My 16core Zen5 in not affected by this but, so hopefully its just the 8core 9700X that need a fix

Tested the following:
Win11 VS Win10 = ~same score
SMT OFF VS SMT ON tested in both OS = ~same score
6600/2200 VS 8000/2200 = ~same score

But did update my scores with a higher OC

Speedometer 2.0 in Chrome + Zen5


Speedometer 3.0 in Chrome + Zen5


*edit*
Hmm maybe i should try to disable a CCD to test SMT ON/OFF as a simulated 8core

 

[MS_AT]

Do your daily setting include a core overclock? I am asking just to know if you get this just with memory tuning or some additional PBO, Max frequency etc tunings

 

[Det0x]

Do your daily setting include a core overclock? I am asking just to know if you get this just with memory tuning or some additional PBO, Max frequency etc tunings

My "daily 24/7" settings is @ 8000/2200 with a light PBO CO overclock
Haven't bothered to do any finetuning on this cpu since i'm planning on getting a retail 9950X when they go to sale

 

[poke01]

Highest score ever in Jetsteam 2 and Speedometer 3.0, Safari. M4 Max with High power mode enabled.



Source: Alex Ziskind

Zen5 9950X3D will match or beat M4 Max I believe

 

[Eug]

Highest score ever in Jetsteam 2 and Speedometer 3.0, Safari. M4 Max with High power mode enabled.
View attachment 111642

I got 492 on my M4 Mac mini. There is no high power setting on this machine.

 

[DigDog]

*saved*

 

[DigDog]

oh wow my brave is NOT fast, 21.9 on Speedometer 3.0 - WX LTSC, 5600X, RX590.

 

[Doug S]

I thought this was interesting in light of the differences in results we've seen across browsers here and the different approaches major browsers have taken to minimize unintentional sharing of cross site information. That Safari is the most aggressive about avoiding such sharing and has been doing this for over a decade, versus more recent changes to Chrome and Firefox, might account for Safari being behind performance-wise in the past. At least, as I recall, the x86 Mac days Safari seemed to always be noticeably slower on benchmarks than other browsers on x86 Macs, though they'd been closing the gap in the last couple years before ARM Macs appeared.

Wish there was more information about the differences other than Google's claims (which are based only on the way they implement this not on how Safari/Firefox do, and show instrumented metrics rather than actual performance differences) to see how big the real impact is. Not that I would want to give up such safety in the name of faster browser performance, but I'm still curious about the real world numbers.

https://addyosmani.com/blog/double-keyed-caching/

 

[igor_kavinski]

https://addyosmani.com/blog/double-keyed-caching/

The easiest way to solve this problem is to hash the resource being cached and compare the hash value to that of another resource with same name and different URL and if it matches, just point to the original resource instead of storing duplicates. I don't know how privacy would be compromised by storing most reused web components in this manner.