Opened 14 years ago
Closed 14 years ago
#182 closed defect (invalid)
Arora page loading performance
Reported by: | rudi | Owned by: | |
---|---|---|---|
Priority: | major | Milestone: | Qt Enhanced |
Component: | General | Version: | 4.6.2 |
Severity: | low | Keywords: | |
Cc: |
Description
With I compare my current 4.6.3 build with an older 4.6.2 I notice a performance degration in Arora. Basically, when the page load progress reaches about 60%, the CPU load maxes out and stays there for a while. After that, the page is shown correctly. With 4.6.2 this CPU load peak is there as well, but smaller and shorter.
It seems to be especially visible when a page is loaded for the first time after Arora has been started. Also it might depend on the page. I saw it with http://www.commtalk.org/ and http://www.s-t.de/ .
Could someone of you try to compare Arora's loading performance running on top of 4.6.3 and 4.6.2 ? Especially the sequence: start the program - goto one of the mentioned URLs.
Change History (5)
comment:1 by , 14 years ago
comment:2 by , 14 years ago
Hmm, I just wanted to ask here, because I don't have a clean 4.6.2 at the moment...
BTW, I downloaded the latest Arora snapshot (0.11.0) from http://code.google.com/p/arora/downloads/list
The maintainer has accepted my patches. So it now builds "out of the box" by simply running qmake and make. It also comes with an OS/2 icon.
comment:3 by , 14 years ago
Congrats, great news!
I have 4.6.2 but in my virtual environment it doesn't make sense to test something for speed.
comment:4 by , 14 years ago
I did a new installation with Qt binaries loaded from a different location (more like is used to be with 4.6.2) and it turns out that the performance increased quite a bit. I suspect that the probelm was related to plugin loading (i.e auto-detection of image formats that require a plugin) which didn't work in 4.6.2... So I tend to think that the behavior is normal.
comment:5 by , 14 years ago
Resolution: | → invalid |
---|---|
Status: | new → closed |
It can be WebKit itself. It has been updated in 4.6.3.
Anyway, one of the ways to find regressions like that is that you check out out a clean 4.6.2 copy and update it revision by revision (or using the "split in halves" technique) to see which one is guilty. It's slow but that way you will find the exact changeset.