Slashdot Mirror


Google Chrome 55 May Use Less Memory (blogspot.com)

Slashdot reader justthinkit writes: Google Chrome is arguably the best browser and the biggest memory hog. Presently. But the Google engineers are hard at work, optimizing the next version of Chrome. Will this be an important, or just another incremental, upgrade?
They're specifically targeting the browser's JavaScript engine, V8, and they've already "analyzed and significantly reduced the memory footprint of several websites that were identified as representative..." (For example, on the mobile New York Times site they've reduced heap memory consumption by about 66%.) Chrome 55 is scheduled for release in December. Any Chrome fans looking forward to testing its performance?

40 of 74 comments (clear)

  1. on the other hand by Chrisq · · Score: 5, Insightful

    it may not

    1. Re:on the other hand by jellomizer · · Score: 1

      Well the story more or less states that Google is putting attention on reducing the memory footprint in chrome. Now how will that affect overall performance is a big question.

      --
      If something is so important that you feel the need to post it on the internet... It probably isn't that important.
  2. Zero content article and summary? by dwsobw · · Score: 3, Insightful

    "We used the tool to identify inefficiencies with a number of internal types." That is about the most technically interesting part of the whole article. Would have been nice to have a bit more of what was changed, how, why, ...

    1. Re:Zero content article and summary? by justthinkit · · Score: 1

      The article had a lot of content, and detail. Just not the kind of detail I was looking for -- it sounds like part of itself might be reducing memory usage by 0.25MB on a page, for example. That is obviously nothing, but maybe there are other more substantial gains. Who knows.

      Also, I submitted this story to find out if this is a hot concern in general, or particular. Someone above this comment posted that they had 32GB of RAM and who cares. Fair enough. But how many admin systems that are memory starved? And how starved are they? I've seen 4GB laptops slowed to a non-responding crawl by Chrome, so to me this new version is potentially a big deal.

      --
      I come here for the love
    2. Re:Zero content article and summary? by CrashNBrn · · Score: 1

      So install Win 10, and run Linux without the hyper-v VM overhead. Or just install Linux?

    3. Re:Zero content article and summary? by darkain · · Score: 1

      It is extremely important to me! Chrome needs massive optimizations, as it has become so godawefully slow, that myself and countless other business that I've managed has had to switch off of it in the past two years. Funny enough, we're all running Opera now, which is based on the same Blink rendering engine, but we all internally jokingly call Opera "Chrome Stable", because it just gets the good stable bits after Google is done experimenting with alpha and beta code in production with live real users. These businesses I manage tech for all have intranets which heavily rely upon the web browser, and when certain machines are unknowingly shoved into A/B live tests with no notification and no way to opt out, it became way too much of a pain in the ass to manage Chrome for the businesses. JavaScript processing performance was one of those issues for sure.

    4. Re:Zero content article and summary? by fbobraga · · Score: 1

      my corporation has me living in the dark ages with 8GB of ram

      My work desktop have 4GB of RAM, you insensitive clod!

  3. Waiting to see if Chrome 55 runs better than Opera by ITRambo · · Score: 2

    I look forward to a better Chrome experience on Android 6. It can be annoying at times. I prefer Opera Mini and will until Chrome, or another browser proves to be faster and with fewer ads.

  4. A variation on Betteridge's Law? by Hognoxious · · Score: 2

    Indeed. And while it's "arguably the best browser" one might well argue otherwise.

    Does the article actually say anything?

    --
    Confucius say, "Find worm in apple - bad. Find half a worm - worse."
    1. Re:A variation on Betteridge's Law? by Yvan256 · · Score: 2

      They're optimizing Chrome for specific websites.

      You're reading it wrong. They simply used the New York Times website as an example of how their new optimizations are doing with existing websites instead of using benchmarks.

    2. Re:A variation on Betteridge's Law? by guruevi · · Score: 1

      A lot of the devices Chrome runs on only have 1 or 2GB of RAM or sometimes even less. Not desktops but mobile devices is Chrome's biggest target market.

      --
      Custom electronics and digital signage for your business: www.evcircuits.com
    3. Re:A variation on Betteridge's Law? by Alumoi · · Score: 1

      YOU are reading it wrong: "analyzed and significantly reduced the memory footprint of several websites that were identified as representative..."
      So basically it's the same shit as with gaming the tests for graphic cards.
      Google: See? Our browser loads site x, y and z 10 times faster! Never mind the other sites, we've decided these are the sites that count.

    4. Re:A variation on Betteridge's Law? by darkain · · Score: 1

      Facebook did the same thing with developing HHVM (their PHP interpreter)... Do you honestly think that their optimizations would ONLY effect the specific sites being targeted, even through it is the general underlaying architecture which all sites use that is being optimized? For instance, Facebook sent a team of developers inside of Wikimedia to help optimize HHVM specifically for the code patterns used in Wikipedia and other Wiki powered web sites. These same code path optimizations were also extremely useful for thousands upon thousands of other web sites that used similar code paths. This is all apart of the development process, pick a specific benchmark, test against it, optimize against it, then move on to the next benchmark. For Google, one of their benchmark sites is the New York Times on Mobile. I'm sure just like with HHVM (which has a well documented and publish lists of all the sites they optimize for), Google has a very wide array of different common sites that use common code paths and patterns they're optimizing for, so virtually everyone wins in the end.

    5. Re:A variation on Betteridge's Law? by Alumoi · · Score: 1

      OK, so not like the graphic card manufacturers and more like VW? Optimize against a specific banechmark and fuck the rest, right?

    6. Re:A variation on Betteridge's Law? by exomondo · · Score: 1

      Why do you say that? It's pretty clear from the article here that they took real world sites to understand real world performance issues and addressed those performance issues. What part of they explained do you think is analogous to VW?

  5. Until chrome sandboxs tops requiring root access by Viol8 · · Score: 5, Insightful

    ... on linux, then I won't be using it on Linux and would recommend others don't either. Google may think its sandbox code is perfect with no possible exploits but I don't intend to test out the veracity of their naive belief for them on my systems.

  6. Another perspective: by 110010001000 · · Score: 2

    Current versions of Google Chrome use 66% or more memory than they should. I guess no one noticed for years. But now the engineers are going to get to work.

    1. Re:Another perspective: by fintux · · Score: 1

      Still not right. It implies 3x as much as needed, or in other words, 2x more than needed.

  7. Re:Until chrome sandboxs tops requiring root acces by ArchieBunker · · Score: 1

    How is the sandbox running as root?

    --
    Only the State obtains its revenue by coercion. - Murray Rothbard
  8. Re:Until chrome sandboxs tops requiring root acces by norweeg · · Score: 4, Insightful

    Needing root access would mean needing to sudo to run chrome, correct? I don't know what you're doing, but my chrome processes run as user processes, not root processes

  9. Arguably the best browser? by Nunya666 · · Score: 1

    Remember to collect your payment from Google, and do not pass Go.

    Can we get some unbiased reporting, please?

    Arguably the best browser, my ass.

    { rant }
    1) Chrome sucks at tab management. In today's age of wide-screen monitors, tabs belong on the side of the browser, not the top. Although there are add-ins that try to work around Google's arrogance, they all suck.
    2) Chrome was created to help put Google.com in front of user's faces. Why else would Google/Chrome refuse to do DNS lookups for one-word entries in the address bar?
    3) The add-in choices are nowhere near as robust as those for Firefox. Thankfully, Pale Moon is keeping that option alive, since Mozilla is killing Firefox by becoming "just another shitty browser."
    { /rant }

    1. Re:Arguably the best browser? by Scarred+Intellect · · Score: 1

      Well, you kind of just supported the statement you were attacking: "Google Chrome is arguably the best browser...". Well, you disagreed and argued. What's the problem?

      As to your actual points, I disagree with number 1. Tabs do not BELONG anywhere except where each user wants them. Your own arrogance sucks.
      Number two, of course it was created for that.
      Number three, this Pale Moon is new to me, I'll check it out, thanks! I loved Firefox back when it was still Firebird...hate to see what it's become these days...

    2. Re:Arguably the best browser? by Yeechang+Lee · · Score: 1

      Firefox, Edge and Safari also refuse to do DNS lookups for one-word entries. This isn't Chrome-specific behavior.

      Both Firefox and Safari support one-word DNS lookups with a slash at the end, like mail/ or www/.

  10. Google Chrome 55 May Use Less Memory by Anonymous Coward · · Score: 5, Insightful

    Google Chrome 55 May Use Less Memory

    But then again Kate Upton may come by my house tonight looking for a good time. I figure the probability of both being about equal.

    1. Re:Google Chrome 55 May Use Less Memory by antdude · · Score: 1

      That is why I always ask for proofs.

      --
      Ant(Dude) @ Quality Foraged Links (AQFL.net) & The Ant Farm (antfarm.ma.cx / antfarm.home.dhs.org).
  11. Would node.js benefit from this ? by Eric.pl · · Score: 1

    It's using V8.

  12. Is this noob week? by Viol8 · · Score: 3, Informative

    You've heard of the setuid permissions bit, right? You'll find that the sandbox is owned by root with 4755 permissions. You figure out the rest.

    1. Re:Is this noob week? by Anonymous Coward · · Score: 1

      But that information is out of date. They now use namespaces and whatnot now: https://bugs.chromium.org/p/chromium/issues/detail?id=598454

    2. Re:Is this noob week? by norweeg · · Score: 1

      I know what SUID is. The chrome executable is not SUID and has 0755 permissions. So again, what the hell are you doing?

    3. Re:Is this noob week? by norweeg · · Score: 1

      I concede that point. If they're that concerned, run it in an unprivileged container. It isn't hard!

    4. Re:Is this noob week? by norweeg · · Score: 1

      Why not run it in an unprivileged container then? It isn't hard!

  13. Re:Waiting to see if Chrome 55 runs better than Op by Anonymous Coward · · Score: 1

    Waiting to see if Chrome 55 runs better than Opera

    Well it's safe to say that it won't run better than Opera 12, that's for damn sure.

  14. Re:Not adressing the real problem by guruevi · · Score: 1

    As long as you use object oriented programming, that's exactly how it will have to work. You can't load an object into memory and declare them static. Take any library (eg. jQuery), call a function and it will change the entire library object's internal structure meaning you need to either have several copies in memory or do a memcpy every time you call an object.

    --
    Custom electronics and digital signage for your business: www.evcircuits.com
  15. What a non-story by Opportunist · · Score: 2

    Wake me when we know whether it does use less memory, until then, where the fuck is the story?

    --
    We used to have a Bill of Rights. Now, with the rights gone, all we have left is the bill.
  16. Re:Not adressing the real problem by Fwipp · · Score: 1

    You could have shared objects that are just copy-on-write.

    However, I'm not convinced that loading 10x copies of a javascript library is the "real memory hog." Even if each library is 10MB, that's only 100MB used. What machine can't afford 100MB? (Mobile devices can use optimization tricks for multiple tabs; as the usage pattern on mobile involves a lot less tab-switching, so it's okay if there's a bit more latency there).

  17. Re:Until chrome sandboxs tops requiring root acces by norweeg · · Score: 1

    chrome's executable is not SUID. Viol8 did not mention SUID, just that they believed chrome to require root in their original post

  18. Re:Waiting to see if Chrome 55 runs better than Op by fbobraga · · Score: 1

    yeap! Anyone knows why Gello, the default browser for cyanogemod, cant block ADs?

  19. Re:Until chrome sandboxs tops requiring root acces by fbobraga · · Score: 1

    What a polite post!

  20. Browser memory usage by um...+Lucas · · Score: 3, Interesting

    Is it just me to wonder why browser need gigabytes of memory just to display a webpage? They receive text, format it according to CSS rules, display relatively small sized images, and, yes, execute Javascript. Still, a HUGE webpage is still a tiny amount of data.

    Considering that entire operating systems used to run comfortably on systems with 32MB of RAM in yesteryear, and could display all this media, it just astounds me that systems now require 4-8GB to provide a comfortable browsing experience.

    Even if Chromes memory footprint has shrunk a little, i'm certain it still uses an obscene amount of RAM relative to what it actually does most the time.

  21. Re:Not adressing the real problem by Fwipp · · Score: 1

    Sure, but do you think that of those 12GB, a meaningful amount of memory is javascript libraries loaded more than once?