Focus on performance?
-
Although I cannot confirm an excessive slowness, I think that regarding the functionality that Vivaldi has, which allows a more optimized and faster work, the milliseconds that may take longer to open a page is irrelevant.
At least it's my personal opinion, in speed, I can't find a notable difference compared to the other browsers I have installed -
One reason I use Vivaldi is because I found it used much less computer resources than Edge or Chrome, and I just liked its many features which make browsing easier. I use a very modest computer for internet browsing. It has an old i5-2400 with 4 Gbs of Ram. Even with 50+ tabs open each new one only takes a fraction of a second to open. With this amount of use on Edge or Firefox, I am looking at noticeably worse and slower performance. So I find it strange that is your issue. Of course I have never tried using it with 200 tabs open, but I bet it would still out-perform other browsers. I suspect something else is going on. Have you opened your Task Manager to see what else is competing with your computer resources? Maybe antivirus is doing a search and using up your hard drive at the same time?
-
@jamesbeardmore Then we have totaly different view on what is meant by perfectly acceptable I had Atom (before I throw away that mistake of HW engineering) with Fedora/Opensuse XFCE which took maybe 100MB of RAM and it was unusable for me. I will repeat, amount of RAM doesn't have any impact on speed of UI.
@Gwen-Dragon Yeah, thats one thing which can be done, but I don't want to do that if I want to have instant reaction of browser UI. I don't care abut RAM consumption, that's no problem. For example Chromium Edge doesn't have such feature and it doesn't need.
-
@jamesbeardmore said in Focus on performance?:
I just checked my current running Vivaldi instance, and it occupies just under 5% of that..
Where are you looking. In Windows task manager Vivaldi is using 589K in 14 processes on my system with just one tab open in each of two windows.
-
@pesala I generally don't use Windows, so I probably got my stats from "top" when I wrote that. It would have listed the Vivaldi binary plus all RAM it had allocated for web pages and extensions, etc. all together. Wow your 589K seems a rather small amount of RAM; I would expect the Vivaldi.exe binary to be at least a couple of MB before it has attempted to render any web pages, which nowadays normally weigh more than 589K themselves!
@enc0re I define "perfectly-acceptable" as "no noticeable delay between user input and the system's reaction to that input". A psychologist friend of mine said that this equates to about 2 seconds according to some studies - i.e. if there is a 1.5 second wait for a page to load or a dialog to appear, the average user will not be conscious of that wait, and will consider the computer to be "responsive". I deliberately highlighted usage of Vivaldi on my elderly Atom hardware specifically because it is incredibly slow. I don't consider the hardware itself "perfectly adequate" for most tasks anymore; however Vivaldi's performance on it is. And that's what I said in the first place. It's a roundabout way of saying that if you have better hardware and your Vivaldi is sluggish, then you possibly have a problem with your installation because it should be much faster. If it can run fast even on an Atom, it can run fast on your machine.
@jumpsq When you say Vivaldi is more sluggish than Presto, do you really mean "more sluggish than Opera" (i.e. current versions of Opera)? As far as I'm aware, it's been over 6 years since a Presto-powered browser (i.e. Opera 12) was released. A lot of security vulnerabilities and new standards will have appeared since then!
-
@jamesbeardmore Please tell me about the security flaws in Presto. People come up with that once in a while, but I've never encountered the glimpse of a proof. The problem is compatibility.
-
@jumpsq said in Focus on performance?:
@jamesbeardmore Please tell me about the security flaws in Presto. People come up with that once in a while, but I've never encountered the glimpse of a proof. The problem is compatibility.
For openers, no updates have ever been supplied to the Presto engine to mitigate the entire class of side-channel exploits (Meltdown, Spectre, etc). Likewise, no internal provisions against currency mining have ever been patched into Presto. Any 'universal' browser exploits appearing since roughly 2015 or so remain unpatched in Presto. Whether any of these are serious threats against a Presto browser user depends on other security layers/patches in place on a user's system and his browsing habits. The one saving grace of Presto is that it has receded to an extremely minor role in the browsing world, so it represents a very unproductive potential target for concerted attacks.
-
@Blackbird
I disagree, but that is rather off-topic. Disable JS and you should mostly be fine.However, you did not give a single argument as of why a much newer browser should perform worse. Spectre mitigation alone cannot be an answer to that. So what are your posts about? Of course I compare the browser I'd like to migrate to (because I see advantages) with the one I've used for over a decade.
-
@jumpsq And when you disable JavaScript (which disabling indeed is an acknowledged protection mechanism), you will find you've inherently disabled much of the functionality of many, many websites. Please understand, I still use Opera 12 (Presto) for certain kinds of website examinations, analytics, and such (with and without JS disabled), but I no longer use it for casual browsing.
In my above response to you, I was merely pointing out examples regarding your request for Presto security flaws or risks, not why a newer browser might perform worse. Presto Opera was an excellent single-process browser... but browsers today are largely multi-processed to prevent a multitude of problems with a hiccup in one tab crashing the whole browser, among other things. Presto Opera was designed mainly in 2012 with its last tweaked release early in 2013. In electron-years, that represents a very old design since a lot of things have happened in web protocols (both general and security-wise) since then. One of the compatibility issues Presto Opera has is that it no longer offers all the security protocols offered (or in some cases, demanded) by websites in setting up browser handshaking via https.
-
On occasion over the (is it 5 already?) years I have noticed the odd version being a resource hog but this was ususually fixed quickly by Vivaldi devs.
Vivaldi now usually runs between 3% & 8% of system resources on my i5 laptop w/16 GB RAM (latest snapshot).
I usually do not use the # of tabs you mention, but when I do I hibernate areas where I am not focused and I often use 2 separate windows with dozens of tabs each.
Also I am finding it works fine on my older hardware... 10 year old stuff (I've gotta purge). -
I came to the conclusion that unloaded tabs take up a lot of resources (with the Lazy Load Restored Tabs option turned on). Just do an experiment:
- Open about 30 tabs.
- Combine them in a group.
- Close the browser and start it again. (For the reliability of the experiment. Now we know for sure that 29 tabs are not loaded into memory).
- Close the tab group.
It takes me about 3 seconds to close a group of tabs. But they are not loaded into memory! How so?
-
It seems to me that in this place there is simply an abyss of unoptimized code. For example, when a group of tabs closes, the title bar of the browser window changes the text, as if switching to each tab before closing it. And, sometimes noticing out of the corner of your eye how slowly the extension indicators are updated when switching tabs, we can conclude that when you close a tab group, at least the following events occur:
- Change the window title for closing tab title.
- Update all indicators on extensions.
- ...
- And only here the tab closes.
But it's still not about closing a group of tabs, but about the fact that tabs that are not loaded into memory have a very strong and negative effect on browser performance.
-
@neonturbo This is difficult to understand. My system is not nearly as powerful as yours, and Vivaldi is very snappy here. It's not even unacceptably slow on some very dated hardware I run here for testing purposes (both Windows and Linux). The only place it's problematic is in the Linux Mint boot on one of my machines where Linux has installed (and won't let me change) an open-source graphics driver that prevents hardware acceleration. Vivaldi in Windows on that box is smooth as butter - but in Linux it's like walking through quicksand.
That said, there is discussion backstage right now on this very point - some systems generating performance problems (which unfortunately our devs cannot yet reproduce and run to ground) while the vast majority of systems sail blithely on, not a care in the world. But the very fact that there is attention on this exact point is, I think, a good thing. Devs and testers are going to be soliciting more info on system specifics from some affected users, and see if this puzzle can't be solved, I think.
-
@neonturbo , I really can't say, at least on my humble laptop it works fluid and fast. Even more so with the last two updates to v3. I can only explain your negative experiences due to external causes (system configuration, some soft or services that interferes, extension, etc.).
-
Hello! Any news on this topic? Well, the browser works very slowly on a fairly modern PC.
-
@obiwan1
Hi, the image is not very helpful, we need more information.
Vivaldi and OS version, system specs and what is slow.
It looks you open new tabs, I can open 20 tabs in 10 seconds or more with Ctrl+T.
It open speed dial tabs as fast as I can click on an 6-7 year old laptop.OpenSUSE Tumbleweed 20201012
KDE-Plasma-Version: 5.19.5
i7-3520M CPU @ 2.90GHz 8 GiB
HD Graphics 4000
Vivaldi 3.5.2080.3 / Chrome 86.0.4240.101 SnapshotCheers, mib
EDIT: Same on Windows 10. -
@obiwan1 It works very quickly here.
I'm not sure what you are looking for, or how/why Vivaldi should be built differently for your particular system specs. Then, too, I have no idea what your system specs are, whether you have a million bookmarks, have it set to save history "forever," if there are any GPU issues, if you use extensions that slow things, or what. So what is new on "this topic" may or may not even apply to you locally.
-
@Ayespy said in Focus on performance?:
whether you have a million bookmarks, have it set to save history "forever,"
Just as a reference point for comparison, my workstation is 4 years old (my motherboard only has half the memory slots in use and thereare much better processors available for it than what I have). I have over 130,000 bookmarks, my history is set to forever, and I don't notice any significant performance issues. I try to keep my extensions to a minimum, but I've got 18 installed of which a 12-13 are regularly enabled.
-
@BoneTone
Man, you really try to max out Vivaldi resources.
I guess 50% of your 130.000 bookmarks are outdated, I don´t even know what I have done last week (who care where I was surfing last year) and I try to keep my extensions to a minimum too, but I've got 4 installed of which a 2 are regularly enabled.
It seams you have no problems with Vivaldi performance so go further.Have a nice week, mib
-
The tech specs of the system do not matter.
The video was recorded on Core i5-4200M, 2.5 Ghz, 16 Gb RAM, GeForce GT 740M, Windows 8.1, Vivaldi 3.5.2080.3.
But absolutely the same slow performance on Core i7-3770, 3.4 Ghz, 32 Gb RAM, GeForce GTX 660, Gentoo Linux, Xfce 4.12, Vivaldi 3.5.2080.3 - on clean and not clean profiles.That is, neither installed extensions, nor bookmarks, nor something else - do not affect performance, in any case, there is no noticeable effect.
Here: https://forum.vivaldi.net/post/354768 I wrote my observations. I think it's not the system, not the hardware, but the lack of optimization.
As an example, Vivaldi can immediately close the selected set of tabs and recalculate the tab sizes 1 time at the end of this action, instead of closing the tabs one at a time and recalculating the sizes after closing each individual tab.
Here are the related bug reports:
VB-38227
VB-45931
VB-48892There is also a topic in the Russian-language forum: https://forum.vivaldi.net/topic/37992/тормозит-создание-вкладок/
For example, here's a video from another user: https://forum.vivaldi.net/post/298990
He writes that the more time the browser is open, the more it slows down, even if you close all the tabs. And he made a script that reproduces the problem: in the loop automatically opens a new tab and immediately closes it. On a clean profile, after 10 minutes the browser starts to load the CPU by 100%, here is a screencast in this post: https://forum.vivaldi.net/post/299121.https://forum.vivaldi.net/post/299161 - "For 1000 iterations, the delay reached 1000 ms."
I want to add that at the time of the discussion in that topic, this behavior was also reproduced for me. And although now I have lost that script, according to external signs, nothing has changed in the latest versions of Vivaldi.
I also want to add that I noticed such a decrease in performance in Vivaldi after about 1.4-1.5 versions (more precisely, I can't say - their distributions were removed from the server). Right now, on a Linux system, I have version 1.3 installed and everything is fine with it.