630446 - Having the bookmarks sidebar open affects speed of web page rendering

Status: NEW

(Core :: Graphics, defect)

Description

[Boris Zbarsky [:bzbarsky]]

See bug 600410 comment 25 through bug 600410 comment 27.

Comment 1

[Timothy Nikkel (:tnikkel)]

Testing on Linux (no acceleration, and I can't enable it) it gets faster for me with the bookmarks bar, most likely because of the decreased window size.

Comment 2

[Timothy Nikkel (:tnikkel)]

Perhaps hardware acceleration is playing a role here, testing with it on and off would determine if that was the case.

Comment 3

[David Jackson]

(In reply to comment #2)
> Perhaps hardware acceleration is playing a role here, testing with it on and
> off would determine if that was the case.

I see the sidebar slowdown on both Win7 and OS X.  Since hardware acceleration only seems to be working on Win7, this makes me think that the sidebar effect is at least partly independent of hardware acceleration.  [The Speed Reading times are ~50X slower on OS X than on Win 7, making me believe that GPU-acceleration is not yet working for OS X; this is in contrast to Kraken scores which are actually ~30% faster on my OS X machine].

Comment 4

[David Jackson]

(In reply to comment #1)
> Testing on Linux (no acceleration, and I can't enable it) it gets faster for me
> with the bookmarks bar, most likely because of the decreased window size.

Just the opposite of what I see on Win 7 and OS X.  Hmmm.

Comment 5

[Timothy Nikkel (:tnikkel)]

You can tell for sure by going to about:support and looking at the graphics section. There is also a preference for hardware acceleration that can be checked/unchecked.

Comment 6

[David Jackson]

(In reply to comment #5)
> You can tell for sure by going to about:support and looking at the graphics
> section. There is also a preference for hardware acceleration that can be
> checked/unchecked.

Just to be clear, the hardware acceleration box is checked for me on both my Win7 and OS X machines, it just only seems to give a strong boost to the IE9 Test Drive tests on Win7.  On OS X, Minefield, Opera, Safari/Webkit and Chrome 10 all are quite slow in the Test Drive speed tests.  Again, this is against a background of the OS X machine beating the scores of the Win 7 machine on some of the javascript benchmarks.  Graphics is slow in OS X and gets slower when the sidebar bookmarks are active.  

On Win7, the deceleration due to the sidebar is perhaps more immediately noticeable because Minefield goes from rocketing along without sidebar to crawling with sidebar bookmarks active.

Comment 7

[Timothy Nikkel (:tnikkel)]

Making the hardware acceleration checkbox be checked only allows Firefox to try to use hardware accleration, there is no guarantee it will use it. The only way to know for sure is to go to about:support. I'd also like to know if you see the slowdown with hardware acceleration turned off.

Comment 8

[David Jackson]

As you surmised, about:support shows acceleration true only for Win7, with acceleration false for OS X even though the preference for "use hardware acceleration" is checked for both.  

HOWEVER, I updated Minefield and the sidebar slowdown NO LONGER HAPPENS!  

I have been trying various combinations of +/- sidebar and +/- acceleration for the last half hour and no longer see sidebar caused slowdown with either acceleration setting.  I will keep checking, but if I can't see the slowdown again, then I guess it would be fixed for me.

Comment 9

[David Jackson]

As you surmised, about:support shows acceleration true only for Win7, with acceleration false for OS X even though the preference for "use hardware acceleration" is checked for both.  

HOWEVER, I updated Minefield and the sidebar slowdown NO LONGER HAPPENS!  

I have been trying various combinations of +/- sidebar and +/- acceleration for the last half hour and no longer see sidebar caused slowdown with either acceleration setting.  I will keep checking, but if I can't see the slowdown again, then I guess it would be fixed for me.

Comment 10

[Gregory Pappas [:gregp]]

David, are you still experiencing this issue?

Comment 11

[BugBot [:suhaib / :marco/ :calixte]]

Clear a needinfo that is pending on an inactive user.

Inactive users most likely will not respond; if the missing information is essential and cannot be collected another way, the bug maybe should be closed as INCOMPLETE.

For more information, please visit BugBot documentation.