Upcoming Firefox Update Will Decrease Power Usage on macOS by Up To Three Times (zdnet.com) 84
Mozilla teased today an upcoming update for Firefox on macOS that it says will reduce power consumption by a factor of up to three. From a report: The primary beneficiaries of this upcoming update are Macbook users, who can now expect longer battery lives while using Firefox. Firefox's increased battery consumption has been a problem for Mozilla, and a black stain on the Firefox Quantum release -- a revamped, performance-centric version of the older Firefox browser. While Firefox Quantum has received praises for its increased page loading speeds, Macbook users haven't been that delighted, especially when they're mobile and have to rely on the notebook's battery as long as possible.
What are they going to do (Score:3, Interesting)
Just my 2 cents
Re:What are they going to do (Score:5, Insightful)
A process doesn't take any more CPU when split out. The culprit here was OSX being purposefully crippled at OpenGL -- switching to an Apple proprietary API helped.
Re: (Score:1, Troll)
OpenGL sucks these days. It's not designed for the modern system - modern multicore processors or modern GPUs. Apple tried to multi-thread OpenGL back in the day, but that effort didn't pan out it seems.
It's why there are new APIs that are far more friendly to modern architectures - if you're still using OpenGL, it's time to upgrade to Vulkan. Of course, I would fault Apple for not supporting Vulk
Re: (Score:2)
Neither is ActiveX or most of the other widely used drivers for 3d, They were all designed pre-GPU days (Being the GPU used on consumer level computers). However OpenGL was designed to work with older GPU's as well. Such as what was built in SGI workstations, and Sun Microsystems $3k 3D graphics card. They were also a lot of high-powered super computers that used a lot of the modern GPU technology back in the old days, which OpenGL worked quite well with.
The problem is that the Standard OpenGL community,
Re: (Score:1)
Re: (Score:3, Interesting)
spawn a half dozen fewer child processes. Firefox and Chrome are real pigs. Just my 2 cents ;)
I've tried both of them on Linux. My experience has been that Chrome is worse, quite a lot worse. Neither of them, however, can hold so much as a teeny tiny birthday cake candle to Java when it comes to chewing up resources.
Re: (Score:2)
I'm a long-time Firefox-on-mac user, and do a good chunk of that on battery (on the train home, mostly). It can indeed suck power for apparently no reason. I find that quitting and re-starting is usually beneficial - I suspect because the tabs I'm not using aren't actually loaded until I click on them, so all that background crap isn't actually running (yet). I also suspect there's something of a "CPU leak" somewhere which means a stop/restart is required from time to time.
I really like the idea that someon
Re: (Score:3)
Actually it is because some page usually using the canvas API has kicked your mac to use the "high performance" video card, even though you don't need it. The battery is cut in half just to power that stupid stuff. Try pmset -a gpuswitch 0
Re: (Score:2)
Not using Java, so Firefox is by far the biggest energy user, according the the Activity Monitor. Other tasks sometimes zoom up but only when I'm using them. Chrome is maybe 2 to 4 on the list. Now I don't know how "energy" is calculated, it's not just CPU, it's not just memory usage, it's not just writing to disk. But by whatever formula it is, Firefox comes up top despite not being top in all those categories.
OS X is a Unix (as Linux is too) (Score:4, Insightful)
spawn a half dozen fewer child processes.
OS X (being a distant BSD derivative over a Mach microkernal) is a Unix operating system.
So, just like on Linux, starting a child process is dead cheap (thanks among other, to CoW memory management, efficient process scheduling, etc).
Firefox (and Chrome for that matters) isn't using these process as a work pool (spreading compute work over the processes), it's using them to isolate the various tabs, so if one tab crashes it doesn't take the whole browser down.
So in short, having 103 processes, of which only a couple are actually running CPU (the tab in which you're currently playing a video youtube) is absolutely peanuts for any decent OS worth its salt.
(For that matters, on a modern OS like Linux with its Cgroups/Namespaces/etc. and a launch system that can leverage them nicely (e.g.: SystemD), starting 103 processes each running CPU at 100% isn't a problem neither - the first demos of Cgroups was about running make with 128 jobs of gcc in parallel and the system staying responsive. I don't have enough experience with Mac OS X, but I suspect similar performances).
Re: (Score:2)
The main reason for the proliferation of processes is actually security. It turns out that JS can leverage Spectre hardware issues to snoop on whatever's running in the same process. Thus browsers need to move to a separate process per Web site.
Default to noscript to minimize processes (Score:2)
Does a website still need to be isolated in its own process even if the site isn't running any script? If so, then one theoretical way to avoid process proliferation might be to set each domain initially no-script and then let the user choose to enable script for each site that the user visits. This would let the browser put sites not running any script, such as static or form-driven sites, into one or two processes.
Yes: corrupted data. (Score:2)
Does a website still need to be isolated in its own process even if the site isn't running any script?
Yes. Example:
Buggy implementation of some media decoder
+ corrupted or maliciously crafted invalid file.
The decoder could crash instead of gracefully reporting the error.
If it's in the main process, it could take the whole browser down.
wrong scheduler (Score:2)
But lots of processes all contending for CPU can quickly result in up to 80% of your CPU spent on context switching. It really is that bad for some use cases.
If you're spending more time switching than actually running stuff, then your scheduler isn't doing a proper job.
Yes, eons ago, it used to be the norm to switch extremely fast between processes to keep all of them responsive, at the cost of extreme performance degradation.
Nowadays, modern scheduler are a bit more clever in the way they work, they'll correctly prioritize task given enough information. That's the whole purpose of cgroups. The hundreds of processes should compete as one single unit against you
Re: (Score:2)
Most operating systems can handle hundreds of processes before there is an issue with context switching at the levels you're saying.
Re: (Score:2)
A process is just a thread with a load of OS gump associated with it, so there's no more context switching than if you put all those threads into a single process instead.
Re: (Score:2)
(For that matters, on a modern OS like Linux with its Cgroups/Namespaces/etc. and a launch system that can leverage them nicely (e.g.: SystemD),
cgroups are manipulated with simple commands. There is/was no need for systemd to use cgroups.
e.g. == "example given" (Score:2)
"e.g." means "example given".
I'm just giving one example of a software that can leverage that.
I just choose the most widely deployed one, that lots of distro happen to use, including a sizeable chunk of the major ones.
Of course, there are lots of OTHER solutions, I just don't bother making an exhaustive list of every single last way to manipulate Cgroups.
Re: (Score:2)
"e.g." means "example given".
e.g. is the abbreviation for the latin phrase "exempli gratia", meaning "for the sake of example"
Re: (Score:2)
spawn a half dozen fewer child processes. Firefox and Chrome are real pigs.
Just my 2 cents ;)
While this is an issue, it is made worse by the fact child process are kept open for apparently dormant tabs. Many people will have 20 tabs open, with maybe a third of them being active. For the remaining tabs there is really no need to assign a process to them, until they are made active via user interaction.
Re: (Score:2)
I think the problem there is that it doesn't know if those tabs are going to do any background work at some time - and people will complain if FF puts a tab to sleep if they expect it to check for emails or notifications or keep playing the porn vid they're not really watching any more. People are like that.
Frankly, the whole "omg if a tab crashes" thing was nonsense anyway, if a tab crashes on me, I'm going to close and reopen the browser anyway as you never know what its done. So 1 process or 100 processe
Re: (Score:1)
spawn a half dozen fewer child processes. Firefox and Chrome are real pigs.
In Firefox, go to "Preferences, Performance" uncheck "Use recommended performance settings" and set "Content process limit" to 1. This solved this problem for me.
Negative Power Usage (Score:1, Insightful)
If Firefox uses 100W of power, and they decrease it's power consumption by 3x that amount, it means Firefox will now be using -200W. Since that's impossible, I genuinely don't know what the headline means.
Does it mean Firefox will reduce power consumption by up to two thirds? If so, why not say that?
Re: (Score:2)
Does it mean Firefox will reduce power consumption by up to two thirds?
Your math is wrong. By decreasing something by 3 times, you divide by 3.
Dividing by 3 is the same as reducing by two thirds, so GP is correct.
Re: (Score:2)
Does it mean Firefox will reduce power consumption by up to two thirds?
Your math is wrong. By decreasing something by 3 times, you divide by 3.
Dividing by 3 is the same as reducing by two thirds, so GP is correct.
Thank you for also proving that when the end result is the same, the argument is usually rather pointless, as we've now boiled this math problem down to grammatical preference.
How about we not expend any more effort on this six-of-one-half-dozen-of-the-other "problem".
Re: (Score:2)
The target audience for this number is Mac users, so as long as its grammatically correct and presented in a nice font, they could say its eleventy twelve times better and it'd be just as meaningful to them.
Re: (Score:2)
Here's a handy chart [segmeowtationfault.com] to help you.
Re:Negative Power Usage (Score:5, Interesting)
Yeah, that headline is messed up, but it's a quote, not a slashdotism.
It would be better to say that the new version uses two thirds less power, or one third the power, as the current version. Or as a percentage, such as 66% less power, or 33% of current power.
However, brains are weird, especially at math. Some people are able to parse "three times less", a multiplication operation, as a division operation (one third) instead. The key to this swap is the "less", indicating a result that normal multiplication cannot produce.
That only works when the resulting numerator is a one, though. How would you say a new process uses two thirds the power as the original, using this kind of operator swap? "two times (three times less)"? One would have to use parentheses to isolate the operator swap, then multiply the result using a non-swapped multiply operation.
Re: (Score:2)
How would you say a new process uses two thirds the power as the original, using this kind of operator swap?
One-and-a-half times less, of course.
The point is that if you multiply the lesser quantity "L" by the "X less", it gives you the higher quantity "Q". Thus "L * 1.5 = Q".
Re: (Score:2)
That only works when the resulting numerator is a one, though. How would you say a new process uses two thirds the power as the original, using this kind of operator swap? "two times (three times less)"? One would have to use parentheses to isolate the operator swap, then multiply the result using a non-swapped multiply operation.
You would simply say it uses "one third less power". You are over-thinking things, as the type of people that visit slashdot frequently do(not a dig at anyone here). You have to look at it from a lay (and not necessarily unintelligent) person's point of view: If you have "three times" more of something, you multiply it by 3 (I have 5 apples, you have 3 times more apples than me, so 3x5= 15 apples). Therefore, extrapolating that, if you have "three times" less of something, you would divide by 3 (I have 2
Re: (Score:2)
If you have "three times" more of something, you multiply it by 3 (I have 5 apples, you have 3 times more apples than me, so 3x5= 15 apples).
WRONG! If I have 5 apples and you have 3 times as many as me, then you have 15 apples (5*3). If you have 3 times more than me, you do the multiplication then add back to the original. You have 20 apples (5*3 + 5).
Therefore, extrapolating that, if you have "three times" less of something, you would divide by 3 (I have 21 apples, you have three times less than me, so 21/3= 7 apples). Is it the correct mathematical wording? No. But in the common vernacular it makes sense.
For less, you do the multiplication (or division) then subtract from the original.
Re: (Score:2)
The problem with this interpretation is that the purpose of language is to infer the correct message - the intent of the speaker, not to misinterpret something because your misinterpretation is "more correct" by some argument.
When people say "three times more than I have" they literally *never* mean "4 times what I have" they always *mean* three times *as many*.
So, the actual correct answer here is that you infer the intended meaning that the speaker had. That's the "correct" interpretation, since that is t
Re: (Score:2)
The problem with this interpretation is that the purpose of language is to infer the correct message - the intent of the speaker, not to misinterpret something because your misinterpretation is "more correct" by some argument.
When people say "three times more than I have" they literally *never* mean "4 times what I have" they always *mean* three times *as many*.
So, the actual correct answer here is that you infer the intended meaning that the speaker had. That's the "correct" interpretation, since that is the message that was sent.
Literally never? I think this current thread shows that at least some people go against your literally never. I avoid using the terms "more" and "less" (or "fewer") when using multiples just to avoid possible confusion. She has 5 more apples than me, but she has twice as many (not two times more!). We should show understanding and patience to non-native speakers.
Re:Negative Power Usage "Twice as Cheap" (Score:2)
"Twice as Cheap" is something I was griping about when a lawyer explained it: Flange is worth $1.00, Bob is selling them for $0.99 = "Cheap", Alice is selling them for $0.98 = "TWICE as Cheap"
Re: (Score:2)
Re: (Score:2)
That's clearly some lawyer-ese explanation that's used in marketing.
Almost all people who hear that would interpret it as meaning you can afford twice as many.
Re: (Score:1)
Thank you, I was about to gripe about this.
Re: (Score:1)
If Firefox uses 100W of power, and they decrease it's power consumption by 3x that amount, it means Firefox will now be using -200W. Since that's impossible, I genuinely don't know what the headline means.
Does it mean Firefox will reduce power consumption by up to two thirds? If so, why not say that?
No, I think they mean that Firefox will from now on use 100W / (the afforementioned factor of 3) = 33.333W of power on MacOS, but yeah they could just have said that the power consumption on MacOS has been reduced by 66%, or two thirds, which is a way of saying this that even the severely mathematically challenged would understand.
Re: (Score:2)
but yeah they could just have said that the power consumption on MacOS has been reduced by 66%, or two thirds,
In common understanding, "3 times less" does mean reduced by 2/3rds.
Re: (Score:3)
Re: (Score:2)
"they decrease it's power consumption "
It is power consumption? I genuinely don't know what means. You want to reduce power consumption, learn to grammar.
Grammar is not a verb.
Re: (Score:2)
Whether or not words are verbs entirely depends on convention. That could be a dialect, arguably, even a one-person dialect. The rules are not hard and fast.
Re: (Score:2)
If Firefox uses 100W of power, and they decrease it's power consumption by 3x that amount, it means Firefox will now be using -200W.
So we should keep Firefox running all the time... and, if we’re planning on doing something really CPU-intensive, we should first fire up several more Firefox tabs.
Re: (Score:2)
It's just another way of saying "decreased by a factor of X". "Decreased by three times" and "decreased by a factor of three" are saying the exact same thing. Some people prefer the "factor of three" wording, but you could make the same argument: "factor of three what's exactly?"
Re: (Score:3)
Re: (Score:2)
When something is specific to Apple machines, especially their laptops, everyone usually does drag their feet on supporting it. Well, everyone for whom Apple isn't their only (or majority) platform. Just look at how many large open-source applications poked clueless fingers at the Retina display, dragging their feet on caring, when it was new. Now look at how all the same people are dragging their feet on supporting the touchbar.
Re: (Score:1)
from the https://www.mozilla.org/en-US/... [mozilla.org]
vs the special needs of Mac browser code?
Global power consumption of software (Score:4, Interesting)
With billions of computers running the energy consumption of processes becomes something which should be contained. Until now software developers have had the freedom to disregard the global power consumption they cause. In many cases you can still say ok it's the consumer's choice but mostly it has stayed under the radar. Battery autonomy yeah. that is not much of a feedback factor.
How much is the global energy consumption of antivirus and antimalware? Why has my Thunderbird used 30 minutes cpu time since startup? What is the energy consumption of the world watching the latest episode of er, game of thrones.
Re: (Score:2)
With billions of computers running the energy consumption of processes becomes something which should be contained.
We only recently ran an Ask-Slashdot about how often people shut down their computers. Let's just say we should focus on people's behaviours long before we start tackling the smaller stuff such as how efficient a process is operating.
Re: (Score:2)
Not sure I completely agree, most computers these days throttle the CPU and suspend things as much as possible, so they're using little more most of the time compared to being asleep.
What is killing the planet is stuff like Java and .NET, where the most bloated, inefficient code is deployed in order to save a developer a few seconds of typing. Then this code is stuck on a server and a hundred thousand users access it, and the CPU spends way too long going over layer after layer of bloat to perform useful wo
Re: (Score:2)
I'm new here. I'd be happy with a something like a table with numbers of energy consumption of various processes, and their sums.
Maybe theres' some interesting stuff in there. Maybe not.
Super-unity power generation (Score:5, Insightful)
By a factor of up to 3? (Score:3)
What does that even mean? Does it mean down to as low as 1/3rd of current consumption or something else entirely?
Why can't these articles even use basic sane arithmetic or do they assume people now are even too stupid to understand fractions?
Re: (Score:1)
Re: (Score:2)
What does that even mean? Does it mean down to as low as 1/3rd of current consumption or something else entirely?
Why can't these articles even use basic sane arithmetic or do they assume people now are even too stupid to understand fractions?
As I said in a reply to someone else, they are using a vernacular. It's safe to assume that most people would agree that "3 times more" of X is X *3, correct? Most people would also agree that the opposite of "more" is "less". Therefore if "3 times more" of X means you multiply X by 3, "3 times less" of something means you would divide X by 3. There's also context: most people realize that, when talking about a physical object or property (except for your bank account), it is usually not possible to hav
Re: (Score:2)
You know what it means, but you pretend not to. The punchline is that you appear to be inferring that it's other people that are stupid.
Re: (Score:2)
Yeah, even an 8 year old would be able to work out that if someone has 15 candies, and you say I'm going to give them three times less candy than that, then they're getting 5 candies.
Zero point energy (Score:1)
Re: (Score:2)
But only if you have a Zero Point Module.
Dear Firefox developpers: thank you (Score:2)
I mean it I truely appreciate the efforts you made to improve Firefox performances.
Now I also happen to use Linux. Is it possible to get rid of Pulseaudio unnecessary dependency ? Audio under Linux is ALSA, Pulseaudio is crap. Thank you.
Re: (Score:2)
Is it possible to get rid of Pulseaudio unnecessary dependency ?
I don't know, is it? Perhaps you could [hire someone to] make a patch and see.
Re: (Score:2)
I had no idea Firefox wasn't using ALSA until your comment. Granted I haven't had any problems with PulseAudio in nearly a decade.
Sadly it's not possible. Firefox devs said they will accept patches concerning ALSA, but no new features, so it doesn't look like they are interested even if someone handed them a bunch fully working, bug-free, sandbox-compatible ALSA code .
There's a hack out there called apulse [github.com] which implements just enough of the PulseAudio API in a thin wrapper library to allow firefox to pla
What about Android? (Score:2)
I really want to use Firefox on my Android phone too, but it's just not usable as it also has a battery drain problem on that platform last time i used it.
Does anybody know if this is still the case today?
Re: (Score:2)
https://play.google.com/store/... [google.com]
There's something wrong with the safe browsing aspect of Firefox for Android. Firefox Preview is amazing, however. It's fast, light, and now supports sync (which it didn't at first and that kept me from switching until recently).
Re: (Score:2)
FirePig (Score:1)
It's still the best choice out there for a browser, but that isn't saying much.
Who cares? Fix Mobile (Score:4, Interesting)
Who cares? There are like four people still running MacOS, and only two of them want to run Firefox. Firefox Mobile is a much bigger problem. It's crap in all incarnations, and they're apparently not even trying to make it good.
First, the interface is crap, and getting crappier. They use their own text input field (in forms etc.) where text selection doesn't work right. You can't just pop up the context menu to paste on the current line, it selects it every time you long press (if there is any.) There is a "desktop mode" switch, but no global "desktop mode" setting, and it will actually forget which pages are supposed to be in desktop mode if the browser gets taken out of memory. When it gets loaded back in the pages reload to their prior location, but then you click a link and it becomes clear they turned that setting off. There is no way to specify to always load certain sites in desktop mode. The throbber has been eliminated and replaced by a progress bar which doesn't work. It doesn't appear until well after you've clicked a link, for one thing, so there's no useful feedback that you've clicked. And speaking of clicking links, it is literally harder to select a link than it is on Chrome. I often have to zoom a page to select a link that I could have just touched in Chrome.
What's worse is that in the "Firefox Preview [theregister.co.uk]" where they allegedly show you where UI is going, not only have they fixed zero of these bugs, but they've moved the address bar to the bottom of the page, and made a drawer pop out of it. I'm constantly opening the drawer when I'm just trying to scroll, because it's where I don't see it (under my fingers and hands.) So instead of fixing bugs, they're spending their time fucking the UI up worse.
Statistically nobody gives a fuck about Firefox on MacOS compared to the number of people who could be running it on Android, but AFAIK numerically nobody is running it because it sucks. Why not fix the problems people actually complain about? Lames.
Re: (Score:1)
I've been using DuckDuckGo's browser on my phone. What's the underlying browser for that? Highly unlikely that hey developed their own. BTW, it works pretty well most of the time.
Re: (Score:2)
If you don't know, it's probably Chrome. I was using Brave (which is Chrome) but they changed the address bar so that you could no longer copy the URL out of it, so I changed browsers. Now I'm using Firefox even though I hated it, just so as not to use Chrome. (If I'm switching, why switch to another Chrome?)
Re: (Score:2)
I looked and it wasn't easy to figure it out. They have the source of their browser for Android here: https://github.com/duckduckgo/... [github.com]
I'm guessing it's Chrome or Chromium-based because it does mention Chrome, but it's hard to tell. It doesn't contain enough source code to implement its own engine.
Re: (Score:2)
I actually run Firefox everywhere. And I personally like the UI changes in Firefox Preview. Not having to go all the way to the top of the screen to enter an address is ergonomically much better, at least to me. Not sure what you're talking about with the text input fields. It works just like every other app on Android for me. No problems with the progress bar (again, in Preview) for me, either. And speaking of a small subset of a small subset of users - people that want sites to load in desktop mode on a m
Re: (Score:2)
"Not sure what you're talking about with the text input fields. It works just like every other app on Android for me."
No, it doesn't.
Find a web form, open it in both browsers. Type some text. Now copy and paste that text onto the same line, right after itself, on both browsers. Not onto the next line, the same line. And in that moment, the student will be enlightened.
I hope nobody is paying you to do QA.
Re: (Score:2)
I know one who uses Firefox in mobile. They all old fashion computers like Macs. :P
Re: (Score:2)
Just noticed another desktop mode bug... If you request desktop mode before the page loads, not only does it never load, but reload from the menu won't make it load either. You have to go to the address bar and hit enter to get the page. And you can't turn on desktop mode when no page is loaded either, so you have to start it loading, at which point you get the mobile version, and then have to select desktop mode AGAIN.
What I honestly don't get is how garbage like this ever gets out. It's clearly both poorl
And what about Linux? (Score:2)
On my Linux laptop I always have a lot of tabs open and Firefox have a number of cycle-eating "Web Content" processes running (top shows at least ten at the moment) and those processes take up about 25% of one core all together. I do not feel it a lot on this machine but the previous laptop got completely bogged down at times and more or less stopped responding unless I had re-niced the processes before it was too late.
I hope Mozilla finds a way to reduce the resource usage on Linux as well as I have a numb
Re: (Score:1)