Why is Pale Moon so slow?

[Gemmaugr]

Off-topic:

I think that should be Global submarine cable map?

Pic

Google submarine cables Interactive Map.png

Google Global Network comprises 7.5+ million kilometers of fiber optics, 100+ interconnection facilities, and 70+ data centers.
As much as their haters would like it, Google is one of those corporations that can say "We literally build the internet." With a devilish laugh, of course.

Ah, it was named google because that was the search term for everything google was involved in. Got it.

Looking a little closer though, most of those are where google just has shares, together with a lot of other companies.

It's also divided into online, project, in development, and proposals. The actual length of their existing cables are... (+1, divided by x, add the two..)

13 470 km + 2 000 km + 390 km + 15 000 km + 6 600 km + 10 500 km. So ca 47 000 km. + 20 000km with Meta.

Still, that's yet another internet access point they will control. Not good indeed. Thanks for the information (and to MC for the DNS I forgot!)

[UCyborg]

It's little faster until you load it with extensions and fully embrace "Your browser, your way".

[frostknight]

For me, its mostly fast, except on specific websites where they try to put too much googlish crap on it.

Ads, heavy javascript, etc...

No idea whose idea it was to make the web out of javascript, but it was a VERY terrible idea.

But anywho 80% of sites i go on, are pretty quick.

AVX is increased speed from SSE2 and even that was fast for me

[UCyborg]

JavaScript is Mozilla thing. And Rust also originates from Mozilla.

So a lot of complaints on this forum about these languages are quite ironic.

UXP also supports weird non-standard JavaScript that would throw on other engines.

So who are you for judging Google for rolling their syntax and functions in their engine?

[UCyborg]

Speedometer says PM's pretty slow, along with my experience with real-life web apps. It's not for the impatient, but in my case, it just happens to work a bit better than other browsers due to their inflexibility in places I take for granted. Although performance of Pale Moon is still often aggravating.

And it can be additionally slowed down with extensions and UI styling.

In my case, startup takes 14 seconds! While empty non-customized profile takes 3. That's 11 second difference, which is not insignificant. Recent (customized as much as still possible) Firefox on the other hand takes 5. These results are when browser files are already cached in RAM. Firefox takes about 24 if files are yet to be loaded from disk.

[Lucio Chiappetti]

Are you (all) seriously concerned about speed ?
I still guess it may depend on the PC, the OS and DE, on the network connection speed, and on the site (connection and site structure).
I do not care at all about PM startup speed. I start it only once (from the WM) when I log in, with just one (remote) home page, and then I manually restore the last Session Manager saved session (usually 3-4 windows with some 30 tabs) ... it takes the time it should
Then most of the sites I frequent are fast, and for those who aren't I blame the site.

[Gemmaugr]

I've been watching a lot more youtube vids this week, and that's the slowest site ever. Yet we're talking about a few seconds (3-5) of initial "activation" at most.
When it eventually does break down, as it's very bloated and doing the whole bad muxed vp9 and deprecated dom 0 "mistakes" thing, I restart the browser and it too only takes a few seconds (3-5).
I don't get the whole instant gratification of site-must-preload and 0ms load on click. Patience is a virtue. Most other sites load just fine in 1-2 seconds.

[andyprough]

In my case, startup takes 14 seconds! While empty non-customized profile takes 3. That's 11 second difference, which is not insignificant. Recent (customized as much as still possible) Firefox on the other hand takes 5. These results are when browser files are already cached in RAM. Firefox takes about 24 if files are yet to be loaded from disk.

Pale Moon takes 1 second for startup for me, Mullvad Browser takes about 1 second, and Firedragon takes about 2 seconds. The past couple of years I've been running browsers with no theming and only one or two extensions. I used to try all the themes and run with a lot of extensions enabled, that appears to have been a significantly more slow and resource demanding way to run them. Firedragon has a bunch of built-in theming, that's probably why it's the slowest to startup of the three.

In terms of the speed for rendering web pages, Firedragon has some noticeable zip to it.

[Lucio Chiappetti]

I insist: do you *really* feel the difference between 1 and 2 sec on browser or video startup ? I do not care.

I do care when in mysql I issue a "select" statement and I haven't set up proper indices on some table. There I feel the difference between 0.x sec ("instantaneous") or 5-10 sec or more ("slow" ... and sometimes my perception is "much slower" than the time indicated as execution time)

[frostknight]

JavaScript is Mozilla thing. And Rust also originates from Mozilla.

So a lot of complaints on this forum about these languages are quite ironic.

UXP also supports weird non-standard JavaScript that would throw on other engines.

So who are you for judging Google for rolling their syntax and functions in their engine?

Google is a corporation, they should have no right to push their own agenda... corporations usually have bad intentions which is why this is bs what you said.

Do not defend google or you will be likely be attacked over their bs.

I cannot believe anyone is this arrogant as you just were.

Besides, don't forget this:

https://en.wikipedia.org/wiki/Don%27t_be_evil

They dropped that motto silently... so its clear they are evil.

[andyprough]

I insist: do you *really* feel the difference between 1 and 2 sec on browser or video startup ? I do not care.

No, I was just responding to @UCyborg's time measurements with my own observations, but I agree, they are meaningless.

[frostknight]

No, I was just responding to @UCyborg's time measurements with my own observations, but I agree, they are meaningless.

Takes me 2 & half seconds to boot palemoon on my one laptop with i5-8250U

[UCyborg]

Then there's also delayed response when switching tabs. Not as much on startup, but gets worse.

I've been watching a lot more youtube vids this week, and that's the slowest site ever. Yet we're talking about a few seconds (3-5) of initial "activation" at most.

YouTube in its current form is considered unusable for me. I only still access it through Project VORAPIS.

Patience is a virtue.

I've been patient for a long time, but it's run out and my nerves are no longer as good as they used to be. These seconds pile up.

Off-topic:

I do care when in mysql I issue a "select" statement and I haven't set up proper indices on some table. There I feel the difference between 0.x sec ("instantaneous") or 5-10 sec or more ("slow" ... and sometimes my perception is "much slower" than the time indicated as execution time)

Non-optimal InnoDB buffer size can also make things pretty slow.

Then adding new columns to large tables is notoriously slow. And to some extent creating tables too. Especially on Windows. Watching software at work's log when it's creating tables (currently 176 of them) on MySQL, taking its sweet time. While on MS SQL, it just blazes through.

I once compared migrating data from one larger SQLite database, migration to MySQL took over 50 min while just little over 10 min to MS SQL.

We only recommend MySQL to cheapskates who refuse to pay for pro version of MS SQL. Though MS SQL Express was recommended in the past as well and I guess it's OK if their database isn't growing too quickly or they don't mind having to delete old data.

Google is a corporation, they should have no right to push their own agenda... corporations usually have bad intentions which is why this is bs what you said.

What was bs? I merely stated what garden JavaScript and Rust originally grew on.

As for both Goanna and Blink supporting constructs the other doesn't understand, that's the fact too.

At least this one thing doesn't work in Gecko or Blink:

Code: Select all

try {
  // code where error may occur
} catch (ex if ex.expression) {
  // error handling if condition is true
}

Or:

Code: Select all

var users = {
  jon: { username: 'Jon', genrePref: 'rock' },
  lucy: { username: 'Lucy', genrePref: 'pop' },
  mike: { username: 'Mike', genrePref: 'rock' },
  luke: { username: 'Luke', genrePref: 'house' },
  james: { username: 'James', genrePref: 'house' },
  dave: { username: 'Dave', genrePref: 'bass' },
  sarah: { username: 'Sarah', genrePref: 'country' },
  natalie: { username: 'Natalie', genrePref: 'bass' }
};
for (let [k, v] in new Iterator(users)) {
  console.log(`${v.username} | ${v.genrePref}`);
}

Though perhaps this one just doesn't behave like documentation describes (Iterator is supposedly not constructable on its own) rather than being some specific Mozillaism.

Don't know if there's any more such constructs.

[frostknight]

What was bs? I merely stated what garden JavaScript and Rust originally grew on.

You made it sound like what right do we have to question Google's motives for rolling their syntax and functions in their engine.

You made it sound like we have no right to judge them which is patently a falsehood.

Google is evil and should be judged by everyone for being the vile pit of moral bitrot they are just like microsoft, apple and many other corporations.

That's what angered me to my response.

[Gemmaugr]

Then there's also delayed response when switching tabs. Not as much on startup, but gets worse.

I've been watching a lot more youtube vids this week, and that's the slowest site ever. Yet we're talking about a few seconds (3-5) of initial "activation" at most.

YouTube in its current form is considered unusable for me. I only still access it through Project VORAPIS.

This might seem too obvious, but have you gone over your addons, or are you using a computer with failing hardware?

I've seen that you seem to struggle with Pale Moon in some areas, while most others have no issues at all.

(I even seem to recall that it's been brought up before, in relation to PassiFox)

[dinosaur]

Yes, PM can be slow, super-slow even. And strangely, after visiting some slow sites and closing the corresponding tab, PM stays slow and needs a File -> Restart...

It has not always been the case and, in the past, 2x.y PM versions used to be much faster and did not suffer from that closed-tab-post-lag effect.

[Moonchild]

It has not always been the case and, in the past, 2x.y PM versions used to be much faster and did not suffer from that closed-tab-post-lag effect.

They would also utterly fail to render the sites that are now causing those slowdowns.
I'm pretty sure this is less of an issue for Chrome and Firefox because they just kill the content process that is messed up by the bad scripting and has a bunch of zombie events running circles on the CPU. Closing a tab there is effectively closing the browser. Opening a new tab is spinning up a new browser application.

[Kand_in_Sky]

And strangely, after visiting some slow sites and closing the corresponding tab, PM stays slow and needs a File -> Restart...

That's the one thing me do nnot understand why this behaviour. It happens every time visting twitter/x. Only restart brings help, also this occurs with Basilisk.
Dactyloidae not fully tested, while it seems to be a little better, only the heavy-js tab, like twitter/x, seems to slow down. Which may be no wonder, Dactyloidae has more instances open, what can bee seen in taskmanager.

[Gemmaugr]

That's the one thing me do nnot understand why this behaviour. It happens every time visting twitter/x. Only restart brings help, also this occurs with Basilisk.
Dactyloidae not fully tested, while it seems to be a little better, only the heavy-js tab, like twitter/x, seems to slow down. Which may be no wonder, Dactyloidae has more instances open, what can bee seen in taskmanager.

If you're only reading twitter/X, then there are some frontends that works great. If you're posting though.. yeah..

[Moonchild]

The issue with slowdowns caused by certain sites is because they create a complex document tree that is interlinked with itself. It becomes impossible for the garbage collector to trace and clean up, so it gives up and even if you no longer use the website, it will remain in memory and the garbage collector still tries to clean it up.
The "ghostbuster" tries to address this but it's not really successful in the case of the biggest culprits due to the sheer size of allocations involved with those sites.
I've been trying to find a solution for this and I have some things I want to try when I have time to do so, but so far I've not had the opportunity (among other things because of the deluge of sec bugs I have had to audit and pressing web compatibility problems).

Other browsers are able to recover better because they "junk the entire car when the ash tray is full" to make a really outdated analogy.
If Dactyloidae is using e10s on a UXP basis then that is really dangerous because there are literally thousands of vulnerabilities in e10s that can cause exploitable crashes. So caution strongly advised.