Online Browser Performance Test

Pesala

Some people think that performance is important for a browser. So is HTML5 feature support.

Opera developed the Acid 3 Test, which may be irrelevant now.

How much work would it take to create a simple suite of online tests to test browser performance and feature support?

Is there any way to detect Vivaldi now that its User Agent changed to Chrome?

It would be nice to be able to point users to an objective test when they say that Vivaldi's performance is poor.

greybeard

@Pesala A little #OffTopic... but my opinion is that there are too many web developers attempting to put too much "flash" on their pages. Add to that scripts, ads, trackers, fingerprinters, third party widgets, etc. we get slower responses that we expect.
My i5 6200 series works fine. My home network and WiFi are OK. My ISP is slow.

Catweazle

On HTML5 test Vivaldi has the same result of Chrome and better than Firefox and Opera.
Maybe you can test here
https://web.basemark.com/

Pesala

@Catweazle That is part of the problem. Vivaldi is detected as Chrome. I don't know if that matters, but the site incorrectly detects my monitor resolution on both 1200x1600 portrait and 1920x1080 landscape.

Chrome 81.0.4044.114 engine Blink resolution 1440x1920
Chrome 81.0.4044.114 engine Blink resolution 2304x1296

Pathduck

For HTML5 support there's this:
https://html5test.com/

Generally, in terms of standards spec support and render speed, I think most modern browsers would be pretty much equal. Any differences will be minimal and hardly noticeable to the average user. Sometimes one or the other engine is ahead, so this is a continuous battle.

However, I think the problem is more the average user's perception of performance and this comes from stuff like going full-screen on a video, or opening/closing tabs. This is done by the Vivaldi UI and not the render engine. So the user will claim that for instance Firefox is silky smooth while doing this while Vivaldi seems slower. And they will often be correct, as the Vivaldi UI can sometimes feel sluggish compared to other browsers. Not much, but a little bit is all it takes.

I think the team has some work to do on making the UI feel smoother in operation. Personally I always disable fluff like animations when opening tabs and so on, I just want things to happen the instant I press the key. But of course most users like this kind of fluff so it's a matter of tuning it to perform better.

Catweazle

@Pathduck , as a general rule, all current browsers have quite the same speed (half a second above or below), that's why I find these speed tests uninteresting. I am more interested in the tests of Browserleaks or Panopticlick, since a browser that is a tenth faster does not serve me, if instead it spreads on the network even the size of my underpants.

Hadden89

@Pesala said in Online Browser Performance Test:

How much work would it take to create a simple suite of online tests to test browser performance and feature support?

Would be actually useful? Such suite to compare UI should have at least two browser made in Vivaldi style (js/css/react).
Comparing Chromium|Firefox|Vivaldi is pointless as are made in different ways.
About web features, they are inherited from chromium (so we lose few licenced codecs, sadly) and are pretty much the same.

Is there any way to detect Vivaldi now that its User Agent changed to Chrome?

Most websites officially can't - which is actually a good thing. Only Vivaldi partners get the vivaldi user agent.
(Sadly this prevent users to check Vivaldi usage in the wild)

Pathduck said the rest on the perception paragraph. Users usually compare UI and not the actual performance (which are very similar)

Pesala

@Hadden89 At least an objective test could demonstrate to such users that their perception that the browser is slow is not backed up by the facts.

sgunhouse

But no one would believe a test hosted by the developer of browser X was actually impartial. There are any number of browser tests out there already - just post links to them somewhere.

I'm not sure why V performs faster than Opera on several speed tests. I can't see any reason why it should - they're both just versions of Chromium - but it has for a couple of years now.

Ayespy

@sgunhouse said in Online Browser Performance Test:

they're both just versions of Chromium

Except they are not, which might help explain it. Each of these two browser companies patches the engine very differently

Catweazle

Even the same browser on different PCs and different configurations and extensions can show very different results.
Other soft that are working in the background also influence, for example AV (see the issues with Symantec and Kaspersky) and naturally the bandwidth that we have contracted.
Otherwise I do not understand that users on a high-end PC can take 10s to open a web page that I open instantly on my cheap laptop and same OS.

Pesala

@Catweazle The test could also calculate the bandwidth as speedtest.net does.

Catweazle

Yes, some of them do this, but it's only one of several factors.
Anyway, if I notice poor browser performance that other users don't have, I would not go to a benchmark page, but would review my system and settings

Pesala

As usual, after a major update there are some users complaining about slow loading of pages etc.

On My Opera we frequently used to compare PeaceKeeper benchmark test results to see how Opera was doing compared to other browsers.

I think that well-designed objective tests could help to diagnose genuine problems, and offset the spread of misinformation.

Catweazle

@Pesala said in Online Browser Performance Test:

As usual, after a major update there are some users complaining about slow loading of pages etc.

On My Opera we frequently used to compare PeaceKeeper benchmark test results to see how Opera was doing compared to other browsers.

I think that well-designed objective tests could help to diagnose genuine problems, and offset the spread of misinformation.

Peacekeeper ended activity in 2015. In any case, a benchmark test for me would only be interesting if it shows the causes of possible slowness in the results. At least for me it is irrelevant if a browser is a tenth of a second faster or slower to load a page.
Most of the current browsers are more or less equal in their speed, if not, it is practically always due to external influences (extensions, security soft, malware, bad configuration, server problems, etc.)

Pesala

@Catweazle I wonder if the test could work in a private window?

As you say, there are a lot of factors that can cause slow-downs, from temporary server issues, to VPNs, to extensions, ad blocking, and user's hardware.

The purpose would be to eliminate the variables and establish some test conditions.

Catweazle

@Pesala , most of the modern browsers are Chromium based, with similar results.
But as I said before, a test that results shows the reasons for a possible slowness of a browser would be desirable.
Comparatives have more to do with sport than with real reasons. Is my browser capable of working well on pages like Chrome Experiments or The FWA (Vivaldi does)? Excellent, if not, use another browser. Real life tests.