Announcement

Collapse
No announcement yet.

Google Chrome 51 Released As Stable

Collapse
X
 
  • Filter
  • Time
  • Show
Clear All
new posts

  • #11
    Originally posted by F1esDgSdUTYpm0iy View Post
    Why do I use Chrome and not FF? FF is just too damned slow. And it's been slow, in my experience, for years now. No matter what they do to it, it's not speeding up, in my experience. Although, and this is why I haven't completely given up on it just yet, they're slowly implementing some of Servo's tech into FF.
    It's funny, because the reasons why I switched from chrome to FF this year is because with each chrome release it has become slower with more bugs in each iteration. FF is not as well optimised as chrome, but at least it does not eat all of my 8GB RAM causing my entire machine to slow down. Loading/Rendering only the tabs that I click on is something I wish chrome could do, also scrollable tabs are useful to me too. The only reason I still use chrome is for sites with flash content and the chrome dev tools are still very useful for me.

    I feel like blindly blaming over-optimisation / new potentially unstable features like shadowdom for the issues I have been seeing creeping up in chrome, however I have not used the linux-tick-processor to see what the actual cause is. Here's one CSS issue I have been able to reproduce on chrome V49+ on Linux/OS X/Windows, but not a single response on my issue report pissed me off. I am also extremely disappointed by "did you try turning it on and off" skill level/effort in other bugs that effected me.

    I only started noticing issues around chrome V40. I still prefer chrome V40 above FF, but I value security updates too much so I just ended up not using chrome for general web "stuff".

    The CSS issue really shocked me, especially since twitter's bootstrap uses those properties. I have never seen such a massive regression in any other browser. Where newer versions of chrome can't handle 10k element updates, any other browser handles 50k element updates without breaking a sweat (including old versions of Internet Explorer)... Meanwhile chrome's thread locks and you can't even scroll up or down/enter a different url in the address bar and that one page uses more than 2GB RAM.

    Edit: Sorry about my double rant, the forum claimed it did not receive my initial post then after my second post it appeared.
    Last edited by Jabberwocky; 26 May 2016, 09:26 AM.

    Comment


    • #12
      Originally posted by Jabberwocky View Post
      Here's one CSS issue I have been able to reproduce on chrome V49+ on Linux/OS X/Windows
      If I run that Fiddle on Chrome 53, the tab in question never peaks over 220 Mb. And the stream I'm watching in the background never interrupts, stutters or whatever. 5 open tabs, one of which a Twitch.TV stream, one of which GMail, one of which your Fiddle and the system remains perfectly fine. Even when running it multiple times.

      Might it have been an issue with one very specific version?

      Comment


      • #13
        Originally posted by F1esDgSdUTYpm0iy View Post
        Might it have been an issue with one very specific version?
        I've just tested Chrome on OS X, there has been a major speedup between version 50 and 51.

        The 50k update is still noticeably slower than FF, I ran it a couple of times. On average it takes about 4-5 seconds in Chrome 51 and just over 1-2 seconds in FF45 on my Intel(R) Core(TM) i7-2620M CPU @ 2.70GHz. I'm glad to see that the RAM spike is gone on my side that tab peaked at ~230mb too.

        The screenshot was taken on Chrome 50 where it still failed badly (sorry if the link stops working, forum does not authorise attachments).

        Comment


        • #14
          Originally posted by Jabberwocky View Post

          It's funny, because the reasons why I switched from chrome to FF this year is because with each chrome release it has become slower with more bugs in each iteration. FF is not as well optimised as chrome, but at least it does not eat all of my 8GB RAM causing my entire machine to slow down. Loading/Rendering only the tabs that I click on is something I wish chrome could do, also scrollable tabs are useful to me too. The only reason I still use chrome is for sites with flash content and the chrome dev tools are still very useful for me.

          I feel like blindly blaming over-optimisation / new potentially unstable features like shadowdom for the issues I have been seeing creeping up in chrome, however I have not used the linux-tick-processor to see what the actual cause is. Here's one CSS issue I have been able to reproduce on chrome V49+ on Linux/OS X/Windows, but not a single response on my issue report pissed me off. I am also extremely disappointed by "did you try turning it on and off" skill level/effort in other bugs that effected me.

          I only started noticing issues around chrome V40. I still prefer chrome V40 above FF, but I value security updates too much so I just ended up not using chrome for general web "stuff".

          The CSS issue really shocked me, especially since twitter's bootstrap uses those properties. I have never seen such a massive regression in any other browser. Where newer versions of chrome can't handle 10k element updates, any other browser handles 50k element updates without breaking a sweat (including old versions of Internet Explorer)... Meanwhile chrome's thread locks and you can't even scroll up or down/enter a different url in the address bar and that one page uses more than 2GB RAM.

          Edit: Sorry about my double rant, the forum claimed it did not receive my initial post then after my second post it appeared.

          Chrome has switched the way they handle memory in the last few updates. They now have a reference counted memory architecture. I suspect they have not optimized it to its limits, but I also suspect it is probably missing some things in its counting, since it is a huge code base.

          Comment


          • #15
            Originally posted by Jabberwocky View Post
            I've just tested Chrome on OS X, there has been a major speedup between version 50 and 51.

            The 50k update is still noticeably slower than FF, I ran it a couple of times. On average it takes about 4-5 seconds in Chrome 51 and just over 1-2 seconds in FF45 on my Intel(R) Core(TM) i7-2620M CPU @ 2.70GHz. I'm glad to see that the RAM spike is gone on my side that tab peaked at ~230mb too.

            The screenshot was taken on Chrome 50 where it still failed badly (sorry if the link stops working, forum does not authorise attachments).

            Weird, tested it on Chrome 50 on Windows and it worked fine, although it really was slower than other browsers. I did not get any hangs, only around 10% CPU usage for a few seconds. After the update to 51 it is now much faster.

            EDIT: After testing again, I realised that that was the total CPU usage, Chrome is maxing out one thread, which is 12.5% of my CPU.

            On Linux Chrome 50, the 10k ran for around 10 seconds, 50k actually gave me the "page unresponsive" popup, I think you were right, this is pretty bad.
            Last edited by PZiggy; 26 May 2016, 10:25 AM.

            Comment


            • #16
              Originally posted by PZiggy View Post
              I think you were right, this is pretty bad.
              Try it with Unstable channel (currently on 53); I'm not getting the unresponsive notification, even on the 50k. Just runs for a while, that's it. Mouse remains responsive, other tabs remain responsive.

              Which kernel are you on?
              Code:
              uname -r -m

              Comment


              • #17
                Originally posted by F1esDgSdUTYpm0iy View Post
                Try it with Unstable channel (currently on 53); I'm not getting the unresponsive notification, even on the 50k. Just runs for a while, that's it. Mouse remains responsive, other tabs remain responsive.

                Which kernel are you on?
                I'm on 4.5.4-1-ARCH x86_64

                It is massively improved with google-chrome-dev, 50k is faster than 10k on stable.

                P.S. Are you sure about it being on 53? Mine is 52, and according to this it should be at 52.

                Comment


                • #18
                  Originally posted by PZiggy View Post
                  I'm on 4.5.4-1-ARCH x86_64
                  4.5-5.dmz.1-liquorix-amd64 x86_64 here; the Liquorix tweaks might keep the tab responsive more readily although I'm not sure on that one.

                  Originally posted by PZiggy View Post
                  It is massively improved with google-chrome-dev, 50k is faster than 10k on stable.
                  Beginning to suspect something is/was just off about Chrome v50.

                  Originally posted by PZiggy View Post
                  P.S. Are you sure about it being on 53? Mine is 52, and according to this it should be at 52.
                  You are absolutely right and my fingers were absolutely wrong:
                  Code:
                  dpkg -l | grep -i chrome
                  Code:
                  google-chrome-unstable                                      52.0.2743.6-1                              amd64        The web browser from Google

                  Comment


                  • #19
                    Originally posted by F1esDgSdUTYpm0iy View Post
                    Or maybe it's because FF simply lost its momentum?
                    ^ this. I was a die hard FF user, refusing the temptation of any Google products. Until one day I realized how slow, laggy, and crashy FF had become, especially with any sites that use Flash. Chrome "just worked", was so fast and responsive, and Flash sites were fast and smooth. I was sold. If FF can manage to dump the bloat and get back to being fast and reliable, I'd have no hesitation in switching back. But until then, I'm using Chrome.

                    Comment


                    • #20
                      Originally posted by F1esDgSdUTYpm0iy View Post
                      Or maybe it's because FF simply lost its momentum? Honestly, the fact Mozilla is writing a new renderer (Servo) from scratch in a language they themselves created (Rust) that completely replaces their existing renderer is telling me even Mozilla has given up, to some extent, on seeing a truly long future for the technology employed in FF.

                      Why do I use Chrome and not FF? FF is just too damned slow. And it's been slow, in my experience, for years now. No matter what they do to it, it's not speeding up, in my experience. Although, and this is why I haven't completely given up on it just yet, they're slowly implementing some of Servo's tech into FF.
                      No, I'm pretty sure it's because of the reason I cited. Your average user doesn't notice performance differences between FF and Chrome because they're not appreciable, especially outside of linux where Flash gets updates. I use both (I have a Chromebook and Android devices) but I still prefer Firefox for a number of reasons. On linux I see one major advantage of running Chrome--for watching netflix and amazon content. Aside from that I would still rather support mozilla and use their solid browser than megacorp Google who's been shoving their browser down the throats of internet users for a long time.

                      Comment

                      Working...
                      X