Crash fixes before Christmas – Vivaldi Browser snapshot 2526.3
-
@pettern Windows Task Manager nothing out of the ordinary. When scrolling, one vivaldi-process may jump up to 10 (I had to open reddit and scroll quickly down to achieve such high use). Every other process is at 1 or lower.
Using browser task amanger, the only thing jumping up in high CPU use is GPU Process, and also only when using the browser.
Thinking about it, I wouldn't mind giving Vivaldi more of the CPU's ressource^^
But I've not noticed any higher CPU use over time. Scrolling and other animations, like moving boxes in garlandtools.org/db, just stutter more. So do YT videos over time. Each "slowness" is contained in its tab and is linked to time and activity in the corresponding tab. So reloading the tab would "reset" the stutter for that tab only. It's also just the "display" of the tab which stutters, not the tab's content. As if the framerate would go lower and lower, but the site itself is responding normally. I just don't receive enough frames as before. Killing the GPU process resets the stuttering as well.
/edit Just had a very high case of GPU Process after opening a YT vid in 1440p60. In Windows Task Manager Vivaldi also consumed +80% of CPU. Ending the GPU Process in browser task manager resetted it, but nonetheless when continuing to play the video there's a few audio stutters here and there.
-
Strange problem: Sometimes when closing tabs to the right, the won't close no matter how many tries. If I close one manually, then it works for the rest. The problem is I can't quite find a consistent behaviour, sometimes it works, sometimes it doesn't.
-
Just in case miracle happens:
With few latest builds I started to notice increased CPU usage with time while browser doing nothing.
Browser restart solves it for some time. But it have chance to lose some of cookies and passwords. So I have to choose - either progressive lags or lost data.
Not a good behaviour for such important software as browser. Please fix these bugs. -
@xerkrosis said in Crash fixes before Christmas – Vivaldi Browser snapshot 2526.3:
I've also disabled Hardware Acceleration there to see how different both browser's performance is
I doubt that it's a correct method to measure performance.
It is ok to optimize software for 99% of people who do not disable acceleration.
So problem is only when with acceleration enabled newer builds works slower than older ones.
Better performance for both options is a bonus, but not necessity. -
22 days since last snapshot and we're in '22 (or 2022)…
Meanwhile stable got two minor updates. -
@gif2d Huh, i fear we will get a mutiny here if the Master Chief of Snapshots does not release a new Snapshot in a few days.
-
@doctorg Mutiny of 2 counts?
Nevermind, meanwhile, I'm bored waiting.
-
@doctorg said in Crash fixes before Christmas – Vivaldi Browser snapshot 2526.3:
mutiny ... release a new Snapshot
Oh look, cool, another new update just arrived!
-
@guigirl said in Crash fixes before Christmas – Vivaldi Browser snapshot 2526.3:
Oh look, cool, another new update just arrived!
@guigirl Hah!, heresy!!!!!
-
@guigirl Reported as unnecessarily smug and gloating.
-
A Snapshot with Chromium 98 core is not ready yet.
But more work is in progress with 5.1 Daily. Stay tuned. -
@pathduck It's a fair cop, guv.
I blame it all on... society.
-
@gif2d said in Crash fixes before Christmas – Vivaldi Browser snapshot 2526.3:
Meanwhile stable got two minor updates.
Those were backports of already fixed bugs in Snapshot, no big deal.
-
@guigirl: This is a real mutiny in progress as I can see…
-
@npro OK, I will wait and see…
-
@gif2d well yes, the current release model is not optimal, if you are still using Snapshot as main right now you are not protected with the security patches Stable has received a week ago and if you use Stable you are using a stagnated version without those 4 weeks of further bugfixes, (like the amount of list items in the recycle bin which was fixed in Snapshot 6 days later but was never backported, which truly sucked).
IMO Snapshot should come 2 weeks earlier in a more raw (closer to a real testing version) state but with all security patches applied, and Stable should receive far far more backports in those 8 weeks (closer to a real stable version) before the next version.
-
-
Ppafflick locked this topic on
-
Ppafflick unpinned this topic on