I'm guessing this and power issues are due to the non-native compositer which pcwalton wants to fix in 2019

https://pcwalton.github.io/2018/12/07/plans-for-2019.html

If this release makes it actually usable on MacOS I would be so happy. Everyone says to use FireFox here, but they don't realize that it runs horribly on machines that a lot of people use to develop on.

Reading the release notes:

  Improved performance for Mac and Linux users, by enabling link time optimization (Clang LTO). (Clang LTO was enabled for Windows users in Firefox 63.)

Doesn't seem like this fixes the high CPU issue on MacOS.

Maybe in another few dozen releases they'll fix it. Doesn't Mozilla realize how many people develop on MacOS? Everyone I know develops on a Mac.

When I interned at Mozilla, most of the FF devs developed on a Mac. They are not ignoring the platform. I suspect the issue is as other said, low incidence, high impact.

It seems like they are aware of the issue: https://bugzilla.mozilla.org/show_bug.cgi?id=1404042

To be clear, this doesn't affect all macOS users. My battery reliably lasts a whole day with Firefox running throughout. That said, the subset of users who are affected, like you, suffer a lot - low prevalence, high impact.

In my case, CPU usage goes to the roof if I use a scaled resolution instead of the default one. If that is the case with all macOS users with non-default resolution, I wouldn't call it low prevalence.

You can work around this (not ideal I know) by enabling "Low Resolution" mode for Firefox only: https://support.apple.com/en-us/HT202471

(I have the same issue and I'd like to switch to Firefox)

Not an issue for me. I have a imac 5K and a retina mac book pro. Both with image scaling on by default. I think most recent macs have retina now.

I use Firefox exclusively and have done so for the last 2 years or so. This sounds like it could be a driver issue for specific macs. Both my macs have AMD Radeon chipsets & quad core i7s.

Anyway, I'm sure this issue is real. Bugzilla ticket numbers probably exist for it and might be more helpful than vague complaints about things being slow.

Of course some web sites are a bit unreasonably javascript heavy these days. The downside of a large screen is that pushing a lot of pixels around is not free. Usually closing any offending tabs immediately restores any cpu usage I see. I'd suggest using the new task manager thingy (page menu->more->task manager)

Anecdotal, but I've found it happens when image scaling by a factor other than 2. Affects my setup of a 4K display at 2650x1440 effective resolution, or 1.5x scaling.

Exactly this. I run at 1920x1080 effective resolution on my 15" rMBP, and 1680x945 on my 13". I see the issue on both.

When I switch to the "default" scaling (1440x900 and 1280x800, respectively), it stops. Only the 15" has a discrete GPU, eliminating that as potentially causal.

The "problem" scaling is 1.5x. The default is 2x.

I use only Macs by default. It runs great and has for years. There's probably a real issue somewhere but it's far from as universal as your post implies.

Likewise. Not aware of any weird issues with performance in relation to Firefox. I use it exclusively. I'm on the beta channel. I can't remember the last time I had a browser crash. I generally restart it to apply new updates every few days or so.

If you have performance issues; you might want to check whether you need to blame the browser or some of your extensions.

Seconding your last point — almost every time I've had someone complain about generic Firefox or Chrome performance the problem went away as soon as they restarted it without extensions. There are certainly exceptions but misattribution is common enough that I'm not surprised to see browser vendors adding the UI to make it easier to discover.

This specific problem is one of those exceptions.

This is a known issue, and has been around since at least v57. There are multiple Bugzilla issues on it. The cause is known. It isn't extensions. The fix is just invasive, and apparently ongoing.

That’s true but not what this thread was about. The person I responded to above was making a very broad claim, which is wrong, and jillesvangurp agreed with the observation and added a general point which is correct. There is a specific issue affecting a subset of people with less common configurations but that doesn’t make the sweeping claim true or the recognition that browser performance issues are notoriously poorly attributed untrue.

No, it's specifically what this thread is about.

The thread-parent absolutely cast a wider net than warranted, but everyone in this discussion who has experienced this problem knows exactly which one we're talking about, and the rest are all, "I've never seen a problem!" or "It's probably just extensions." Meanwhile, there's reliably a sub-thread somewhere in the discussion on nearly every article about Firefox, about this problem.

I find it profoundly ironic that you comment down-thread that "humans are very prone to confusing things which affect them personally with the general case" about a thing which you haven't personally experienced. You talk as if those of us for whom this is a 100% reproducible problem are an edge case, based AFAICT solely on your own not suffering it, coupled with your (not incorrect) beliefs about people poorly attributing performance problems in general.

That juxtaposition is really galling.

I was aware of the issue already but that’s also why I knew that the biggest impact comes from a non-default setting. I never said that it wasn’t a real problem, or that it doesn’t warrant attention — only that it wasn’t as broad as claimed and that is a very common problem with browser issues because everyone uses them but people who don’t have problems generally don’t go around posting that everything is fine whereas the percentage of users who are affected will complain regularly.

Scaled resolution option is so widely used that it might as well be broken entirely.

Do you have data supporting that claim?

Do you require a scientific study for any fact? If you bothered to do any googling, or whatever Mozilla equivalent of google is - this is in the top 3 results. https://9to5mac.com/2016/12/02/15-inch-macbook-pro-screen-re... Scaled resolution is defaulted on a huge amount of Macs.

I asked for data because humans are very prone to confusing things which affect them personally with the general case. Since Mozilla uses telemetry heavily I tend to trust their prioritization more than random self-selected commenters.

> Improved performance for Mac and Linux users, by enabling link time optimization (Clang LTO). (Clang LTO was enabled for Windows users in Firefox 63.)

Anecdotally, it seems to be flying through pages it used to struggle on when I last tried it (~version 60). Tab switching seems faster too, but I haven't had enough time to put it through its paces yet. Fairly promising so far, though!

I does feel faster. I still can't watch a 1080p60 twitch stream without droping a ton of frames though.

It's been like this for at least 2 years now on my MBP 2015.

I've been suffering from these performance issues since I started trying Firefox again around version 57. I can say that this version finally feels like they've fixed the performance issues, at least for me.

That said, I've thought this before and turned out to be wrong. But it looks like it might be real this time.

Not for me. CPU usage and perforamnce in general is signifanctly worse than Chrome or Safari.

Same here. Had hope, but the fans of my RMBP state otherwise.

I have used Firefox since the early 2000s, and it has been frustrating to see performance get worse even as my hardware gets better. I realize that this is mainly the fault of bloated websites, but I can also see that Chrome and Safari are faster than Firefox.

I recently switched to Brave Developer Edition [1], which now supports Chrome extensions (which I need for work/pleasure). It runs circles around Firefox and Chrome, and maintains my privacy better.

Previously it was a tradeoff of getting speed (Chrome) or maintaining privacy (Firefox). For me, Brave does better on both counts, and now that it supports extensions I'm completely sold.

1: https://brave.com/download-dev/

If Brave is running circles around Chrome for you in general, that's likely a result of cognitive bias, since last I checked it's the same rendering engine.

For pageloads Brave might do better if it blocks various stuff by default, of course.

> if it blocks various stuff by default

This is one of the key differentiating factors of Brave. Maybe next time don’t lead with accusations of cognitive bias.

I am quite aware of what Brave's value proposition is. Generally when people are only talking about pageload performance they say so, because there are many other performance metrics relevant to web browsers, some of which are arguably more important for most users...

I'm curious about the downvotes — anyone care to enlighten?

The Release Notes say it does. I'm going to try it out later.

It does not.

So are you accusing the release notes of lying, or do you have actual facts to back up this assertion?

The release notes state otherwise.