Mozilla Firefox 53.0 Released, Drops Old Linux CPU Support

[soulsource]

As far as I know, the dropped CPU compatibility is currently just a compiler flag (they now build with SSE support by default). If you need to run Firefox on older hardware, you currently can still build from source. Also, distributions can still choose to build Firefox without SSE support, if they want.

I'm curious though: Do recent Firefox versions yield acceptable performance on such dated hardware?

[GT220]

Mozilla is dropping support for CPUs without SSE2, that's Pentium 3 and Athlon XP or older, those are over 15 years old and makes developing difficult because of the lack of SSE2 instruction set support.

[eydee]

I'm curious though: Do recent Firefox versions yield acceptable performance on such dated hardware?

If you have only 1 tab open, and the web page doesn't have giant images, flash/HTML5 animations/videos, it will run fine, provided you have at least 1GB of RAM. However, a 3GHz P4 can only run Youtube videos at 240p at an acceptable frame rate. And that's under windows, using some basic acceleration. Pre-P4 should have even more trouble.

[franglais125]

Here you don't want to use old versions, not because they're old but for security reasons and because some of those overscripted websites won't work on older browsers.

Not ideal, but you can still run Firefox ESR 52, which is supported for another year.

[rene]

Not happy about all this obsoleting hardware that might still be okay and doing the job. And yes, there is HW from that area that is not power hungry like crazy. But if you want to operate a few things from the old HW and you have any network connection you'll want to have an up-to-date web client SW. Here you don't want to use old versions, not because they're old but for security reasons and because some of those overscripted websites won't work on older browsers.
Besides I wonder what the actual focus of Mozilla is now.

yeah, ironically just the week when I was bitten with my overoptmized portable Linux SSD wanting to run on old core2duo hardware myself ;-) https://www.youtube.com/watch?v=457zniNGVfU

[fedora-user]

Natively or via XWayland?

via XWayland

[Holograph]

I'm curious though: Do recent Firefox versions yield acceptable performance on such dated hardware?

I have a Core i5 4570S (4 core 4 thread), 16GB DDR3, Samsung 850 Pro SSD, Radeon 290x. Running Fedora 25, kernel 4.10.10, Firefox 52, and mainly Cinnamon on Xorg (not Gnome on Wayland most of the time, so I am generally using FF natively and not via XWayland). Addons: uBlock, Disconnect, NoScript, and a password manager that I won't name for security reasons.

I use the same addons on Windows and FF performs... well, not great, but usable without frustration. It's still usable in Linux too but there is definitely more of a performance issue on Linux, at least with my setup. I don't know if it would be better if I tried a Nvidia card or not.

In Linux, having Youtube open without actually playing a video, FF still uses >10% CPU without any other tabs doing anything. Playing a Youtube video uses between about 50% to just over 100% (100% being one core being fully used) CPU. Regardless of whether I use H264ify or not (I didn't check if it was using VP9 without H264ify, just figured I'd give it a try - I also didn't check that H264ify was really properly forcing H.264).

It was pretty slow when I was recently running Gentoo as well, and I was using march=core-avx2 and O2. I want to say it was a bit faster on Gentoo but there were other differences on that setup than just the compilation of Firefox so I can't draw too many conclusions from it.

Chrome/Chromium do perform somewhat better much of the time, though there is still much improvement to be made in the user experience with that as well.



Edit: Oh, and as for this default CFLAGS update, about time. Even if it only helps a little bit on new PCs, frankly it's ridiculous that all new PCs should be held back in performance because a few people want to run old machines. Those people should have to compile their own software for old PCs; it shouldn't be a burden on the newer PC users who deserve to get better performance. You want to use a P4? You should be the one forced to run a source-based distro like Gentoo (or a specialized distro like Damn Small or Puppy), not people with reasonably modern CPUs who rightly expect their software to run with higher performance.

But I expect the distros to ignore this and keep catering to the lowest denominator with their generic (non-)optimizations.

(Though if Firefox performed well in the first place, maybe I wouldn't feel so frustrated about that!)

[Hibbelharry]

Still no Wayland support, no Servo engine, no image pasting, no <input type="date", "datetime", "time", "week" or "month".

Input type date/time fields are available at least in the aurora channel, hidden behind some about:config flags for now. There are some active bugreports with last developer activity just days ago, so it won't be that long anymore.

Mozilla won't suddenly switch to the servo engine, but they will integrate parts from servo into gecko. Bigger steps are going to happen with FF 57.

[stevenc]

frankly it's ridiculous that all new PCs should be held back in performance because a few people want to run old machines.

If SSE2 and portable C codepaths exist in the same binary, choosing between them at run-time might take... 4 cycles on each call. On your 3GHz+ machine I don't think that's really much overhead. This change will not fix your performance issue. If developers only care to test on fast machines, they just add bloat elsewhere to slow it down again.

Or maybe developers are too lazy to maintain both implementations - which are otherwise great to keep for side-by-side regression testing - it will just hurt later when they need that portable C implementation to support some trendy new non-x86 platform, but already they stopped maintaining it.

Those people should have to compile their own software for old PCs; it shouldn't be a burden on the newer PC users who deserve to get better performance.

So if you're just poor, and you can't afford anything better, you deserve to stay that way and not burden the rest of society. F*** you.

(I'll accept that you maybe didn't think of this, or will have trouble believing these economic situations exist: where electricity and Internet access are affordable, but recent (<<10 years) hardware is really not. It's not a strawman argument. I've put new Ubuntu Firefox on dozens of pre-P4, pre-owned machines for people who cannot afford a newer replacement.)

[Holograph]

So if you're just poor, and you can't afford anything better, you deserve to stay that way and not burden the rest of society. F*** you.

Speaking of strawman arguments...

I said that fewer distros should cater to such people, not that all support should be dropped from all software for older hardware. There's no reason that a Pentium 4 user and a Core (even Nehalem) architecture user need to be supported by the same distro.

I can understand disagreeing with that opinion as stated, but I think your wording's a bit strong.


Also, yes, I know the difference wouldn't be major, and I'm also aware of branch prediction when it comes to optimizations that are selected at run-time, and used frequently, such as in a loop. But, frankly, with such terrible performance to begin with, everything helps.