V 6.6 | Slow Performance Since Upgrade?
-
@KlausFranbrau2 said in V 6.6 | Slow Performance Since Upgrade?:
I'm about to revert,
Don't revert in place, that is asking for trouble; new installation and profile directories are needed, and you can use sync to transfer data.
-
@KlausFranbrau2 It is not advisable to run on old versions, except for testing. Just too many potential security problems.
I don't know if I have any real information, beyond a couple of pieces: 6.1 is supposed to work fine, 6.2 is supposed to have issues, and there was a chromium update to 116 a few weeks before 6.2 was released.
-
6.1 Final is https://vivaldi.com/blog/desktop/6-1-rc-2/
-
First 6.2 snapshot was https://vivaldi.com/blog/desktop/catch-up-vivaldi-browser-snapshot-3054-3/
-
The snapshot before default enabling Portals was https://vivaldi.com/blog/desktop/another-round-of-fixes-vivaldi-browser-snapshot-3060-3/ [1]
-
The last 6.2 snapshot based on Chromium 114 was https://vivaldi.com/blog/desktop/same-old-new-window-now-opens-faster-vivaldi-browser-snapshot-3062-3/ , which is also the one that enabled Portal windows by default [2]
-
The first Chromium 116-based snapshot was https://vivaldi.com/blog/desktop/macos-geolocation-and-customizable-address-suggestions-snapshot-3070-3/ [3]
-
6.2 final was https://vivaldi.com/blog/desktop/6-2-rc-2/
Of these, the ones I tagged [1], [2], and [3] may be the most likely regression points. Given that Portals was an Experiment, it is possible to enable it in [1], and AFAIK disable it in [2] and likely for a while longer.
While determining a breakage in [2] will only determine the feature responsible; the actual code involved was under development for months prior to that release, that might help guide an investigation.
-
-
Good news, I think (fingers crossed) that the problem has been solved for me. I believe the fix was a recent update to the driver of my Intel integrated graphics card (31.0.101.5382, release date: 29th March 2024).
Try to update your drivers and see if that helps.
-
@yngve said in V 6.6 | Slow Performance Since Upgrade?:
@RasheedHolland said in V 6.6 | Slow Performance Since Upgrade?:
That there is a possible issue with memory management seems likely, based on what I have seen during shutdowns. But whether that is caused by something in Vivaldi's UI, or something more fundamental in the Chromium V8 engine (which handles the memory management) is something that is not possible to say at this time.
OK thanks for the feedback, I now understand it better why it's so hard to pinpoint this problem.
And yes, I have never seen Vivaldi hanging in memory before, it all started with 6.4 (I upgraded directly from 6.1 to 6.4) but people have been complaining about performance degradation and Vivaldi not being able to exit memory instantly since 6.2 apparently. It usually takes around 5 to 10 minutes to exit from memory. And this only happens after Vivaldi 6.4+ has been active in memory for a couple of hours and probably also if quite a lot of tabs are open.
I have tested it with Process Monitor and no third party app/process seems to be involved with this problem. It's really a shame because I can't upgrade 6.1 because of this issue. Also, can you perhaps take a look at this bug (see link), this is another reason why I can't upgrade to 6.7, this bug has been present since 6.4, should be easy to fix I suppose.
https://forum.vivaldi.net/topic/95696/6-6-bookmarks-nicknames-not-working-in-url-bar
-
@vivaldibecameSLOW said in V 6.6 | Slow Performance Since Upgrade?:
Good news, I think (fingers crossed) that the problem has been solved for me. I believe the fix was a recent update to the driver of my Intel integrated graphics card (31.0.101.5382, release date: 29th March 2024).
Try to update your drivers and see if that helps.
Wow, I would be really surprised if this was the fix for the problem. But the thing is, people who have reported this were all using different GPU's, so I don't see how this would be the issue. Also, I have a laptop with both an Intel UHD GPU and a Nvidia GeForce GPU, which would mean that if I force Vivaldi to use the Nvidia GPU, it should also be possibly solved? I mean without me having to upgrade my Intel UHD GPU drivers?
-
@RasheedHolland I'm surprised as well, but it was right after I updated the driver that Vivaldi didn't slow down after a certain time anymore. I also have a Nvidia GPU + Intel integrated graphics. If I remember correctly, I had already tried in the past to force Vivaldi to use the GPU but it didn't solve the problem (though I'm not sure if I did correctly).
Just upgrade your Intel integrated graphics' driver, it's worth a try!
-
@RasheedHolland said in V 6.6 | Slow Performance Since Upgrade?:
I don't see how this would be the issue.
There is probably no such thing as "THE" issue. There are probably a range of completely different issues with similar symptoms experienced by different people and complained of as the "same" thing.
-
@vivaldibecameSLOW said in V 6.6 | Slow Performance Since Upgrade?:
@RasheedHolland I'm surprised as well, but it was right after I updated the driver that Vivaldi didn't slow down after a certain time anymore. I also have a Nvidia GPU + Intel integrated graphics. If I remember correctly, I had already tried in the past to force Vivaldi to use the GPU but it didn't solve the problem (though I'm not sure if I did correctly).
Just upgrade your Intel integrated graphics' driver, it's worth a try!
I guess it's worth a try, but I don't want to mess things up so that I have to reinstall Windows all over again, it takes too much time. And from a technical point of view, I wonder how the GPU can play a role, and why does it only happen with Vivaldi? Shouldn't this then be fixed in Vivaldi, know what I mean?
-
@Ayespy said in V 6.6 | Slow Performance Since Upgrade?:
@RasheedHolland said in V 6.6 | Slow Performance Since Upgrade?:
I don't see how this would be the issue.
There is probably no such thing as "THE" issue. There are probably a range of completely different issues with similar symptoms experienced by different people and complained of as the "same" thing.
Well, we already spoke about this. You have seen so many problems being reported on this forum, that you simply can not imagine that there is some type of issue with Vivaldi that causes multiple people to report the same thing.
It's the combination of performance degradation (over time) combined with Vivaldi hanging in memory that leads me to believe that this issue is caused by some problem in Vivaldi that was introduced in 6.2.
Nothing has changed on my system in terms of other apps (no new versions) or hardware, except for Vivaldi itself. Which leads me to the conclusion that the problem is in Vivaldi itself, probably in the resource management area. Why it's triggered on certain machines is of course the million dollar question.
-
@RasheedHolland Vivaldi's UX is essentially an embedded webapp, so it's quite possible for some obscure deficiency in Chromium itself w/r/t rendering to be triggered by same. It'd be one of those things where, if you can find it, would be less "fixed in Vivaldi" and more "worked around in Vivaldi until Chromium fixes it".
-
Also, while I'm here, I'm hesitant to declare victory until it's been a couple of days, but so far, the slowdowns for me seem to have gone away after I updated to Vivialdi 6.7.
I've not tried any graphics driver updates because I don't have Intel graphics (I'm on a Ryzen 7), and there have been no updates to my graphics drivers in the interim. But so far, I'm hopeful.
-
@Viqsi said in V 6.6 | Slow Performance Since Upgrade?:
@RasheedHolland Vivaldi's UX is essentially an embedded webapp, so it's quite possible for some obscure deficiency in Chromium itself w/r/t rendering to be triggered by same. It'd be one of those things where, if you can find it, would be less "fixed in Vivaldi" and more "worked around in Vivaldi until Chromium fixes it".
But doesn't the same goes for Brave and Opera? Or perhaps Vivaldi somehow interacts in a ''strange way' with Chromium which causes this problem.
@Viqsi said in V 6.6 | Slow Performance Since Upgrade?:
Also, while I'm here, I'm hesitant to declare victory until it's been a couple of days, but so far, the slowdowns for me seem to have gone away after I updated to Vivialdi 6.7.
Sounds awesome. I will take 6.7 for a test drive as soon as the bug with bookmark nicknames is fixed.
I've not tried any graphics driver updates because I don't have Intel graphics (I'm on a Ryzen 7), and there have been no updates to my graphics drivers in the interim. But so far, I'm hopeful.
I have never updated my GPU drivers in 25 years of computing, I rather stay away from this type of stuff. I'm always afraid of messing things up. I did see there is a new driver for my Intel UHD GPU.
-
@RasheedHolland said in V 6.6 | Slow Performance Since Upgrade?:
But doesn't the same goes for Brave and Opera? Or perhaps Vivaldi somehow interacts in a ''strange way' with Chromium which causes this problem.
Brave and Opera, to the best of my knowledge, both have native UIs rather than "embedded webapp" UIs. That said, I've never used Brave and haven't used Opera since the Presto days, so I don't know for certain.
I have never updated my GPU drivers in 25 years of computing,
I confess I find that nigh-impossible to believe. Perhaps you haven't explicitly manually done so, but these things get automatically updated with some regularity.
-
@Viqsi I'm pretty sure Vivaldi's webapp UI is unique in the browsing universe. It is an additional layer and thereby an expanded area where not only Vivaldi code changes, but also Chromium changes can trigger errors seen nowhere else. Further, every single machine that runs Vivaldi creates a unique environment, non-replicable by devs or testers, where some incompatibility or error may arise.
One way Vivaldi tries to defend against unanticipated friction in the interface, is to have (around seventy at present, I think) volunteer internal testers (called Sopranos), using Vivaldi on every age, capability and configuration of hardware possible, as many hours a day a possible, report back on every cock-up detected. So if any of my machines trips over anything, I report to devs ASAP. Same with the other Sopranos.
I not only use the bleeding edge of Vivaldi as my daily driver some 17 hours a day on my main Windows machine, I also test on three more towers (both Windows and Linux on one of these) and an old X86 laptop here in my office, two newer laptops that don't get turned on every day, a phone, a tablet, and two more towers (one somewhat obsolete) at my remote office when I'm there. To make matters worse, I read all the English posts in the Forum daily and try to reproduce bugs reported by users.
I'm not sure other Sopranos test on as many different kinds of hardware - in fact I'm sure they don't. No one hordes computers like me.
But when the devs and testers cannot reproduce an issue, there is no way to address it. Luckily if, say, even ten percent of users have a specific issue caused by a specific bit of code, one or more of the testers or devs is going to be able to reproduce it, and it's going to get tracked down. It's for this reason that thirty to fifty issues are resolved by the team every. single. day. And I don't even recognize most of the issues in the daily changelog, so I know even obscure glitches are being detected and remedied all the time. Yay team.
-
@Ayespy said in V 6.6 | Slow Performance Since Upgrade?:
@Viqsi I'm pretty sure Vivaldi's webapp UI is unique in the browsing universe. It is an additional layer and thereby an expanded area where not only Vivaldi code changes, but also Chromium changes can trigger errors seen nowhere else.
It's not currently commonplace but I'm not sure I'd call it unique per se. A fair argument could possibly be made that it's fundamentally a modern HTML5 twist on what Netscape/Mozilla started with XUL. (Which I remember teenage me thinking of as being OMG SO HERETICAL when it was first announced... ah, the naivete of youth).
And I'm aware of the Sopranos - many thanks to all y'all, btw. Sometimes I've considered looking for ways to volunteer, but frankly I don't have the patience for it - my bughunting sessions tend to involve lots of incoherent screaming at screens and regular trips to the fridge for more ice cream.
I would hope that Vivaldi's input would be considered very welcome by Chromium developers just by virtue of being in a position to find all sorts of (from their perspective) crazy edge cases as a result...
-
I've seen enough and it's been long enough; I'm calling it. I am no longer able to replicate this slowdowns problem with Vivaldi 6.7. (And there was much rejoycing!)
Funny thing - I also noticed this:
https://forum.vivaldi.net/topic/90135/6-2-6-6-slow-auto-complete-history-suggestions
This also seems to be a mysterious slowdown-related issue that also seemingly started with 6.2 and also seems to have been solved with 6.7. I have no idea if these are connected in any way (probably not), but the parallels amused me.I'll keep an eye out in case the slowdowns perversely deside to come back after my declaring victory, but for now, I'm happy.
-
@Viqsi said in V 6.6 | Slow Performance Since Upgrade?:
Brave and Opera, to the best of my knowledge, both have native UIs rather than "embedded webapp" UIs. That said, I've never used Brave and haven't used Opera since the Presto days, so I don't know for certain.
Yes, you may be right. Brave and Opera don't have most of the GUI customizing that Vivaldi has, so perhaps they indeed use the native Chromium GUI.
@Viqsi said in V 6.6 | Slow Performance Since Upgrade?:
I confess I find that nigh-impossible to believe. Perhaps you haven't explicitly manually done so, but these things get automatically updated with some regularity.
Yes, I've never manually updated my GPU drivers. But now that I think of it, I don't remember updating them via Win Update either.
@Viqsi said in V 6.6 | Slow Performance Since Upgrade?:
I've seen enough and it's been long enough; I'm calling it. I am no longer able to replicate this slowdowns problem with Vivaldi 6.7. (And there was much rejoycing!)
Sounds cool. I remember that with Viv 6.6 I also thought it was fixed, but after 3 days or so it happened again. I assume when you say it's fixed, you mean that even after running Viv 6.7 for at least 7 days with 50+ tabs open, it won't slowdown and won't hang in memory after you close it?
-
@RasheedHolland said in V 6.6 | Slow Performance Since Upgrade?:
Sounds cool. I remember that with Viv 6.6 I also thought it was fixed, but after 3 days or so it happened again. I assume when you say it's fixed, you mean that even after running Viv 6.7 for at least 7 days with 50+ tabs open, it won't slowdown and won't hang in memory after you close it?
Correct. Normally I would get a slowdown within 10 hours of reopening. I have gotten nothing of the sort since the update to 6.7.
-
@Viqsi yes, I think this is now fixed for me too.
I know I said this before, but has now been several days and not needed to restart due to grinding to a halt.@RasheedHolland - thanks for your posts on this, was a frustrating issue, when I've normally had a great experience with Vivaldi.
-
@Viqsi said in V 6.6 | Slow Performance Since Upgrade?:
Correct. Normally I would get a slowdown within 10 hours of reopening. I have gotten nothing of the sort since the update to 6.7.
@magilex said in V 6.6 | Slow Performance Since Upgrade?:
@Viqsi yes, I think this is now fixed for me too.
I know I said this before, but has now been several days and not needed to restart due to grinding to a halt.Sounds awesome guys. I have not tested it yet, because I'm waiting for the fix for the bookmark nicknames.
@RasheedHolland - thanks for you posts on this, was a frustrating issue, when I've normally had a great experience with Vivaldi.
Yes, I know what you mean, it was the first time I had a huge problem with Vivaldi. And the developers couldn't pinpoint the problem, but I was confident they would fix it anyway by tweaking stuff related to memory management, which they probably did.