Closed Bug 310222 Opened 19 years ago Closed 17 years ago

Bad performance with Flash

Categories

(Camino Graveyard :: Plug-ins, defect)

PowerPC
macOS
defect
Not set
major

Tracking

(Not tracked)

RESOLVED WORKSFORME

People

(Reporter: kurtbw, Assigned: mikepinkerton)

References

()

Details

(Keywords: perf)

Attachments

(2 files)

User-Agent:       Mozilla/5.0 (Macintosh; U; PPC Mac OS X Mach-O; en-US; rv:1.9a1) Gecko/20050925 Camino/1.0+
Build Identifier: Mozilla/5.0 (Macintosh; U; PPC Mac OS X Mach-O; en-US; rv:1.9a1) Gecko/20050925 Camino/1.0+

Site non-functional with Camino, Safari 1.3.9 works fine.  Camino loads page
contents eventually, but force-quit is required.

Reproducible: Always

Steps to Reproduce:
1.Go to http://www.museum.tv/home.php
2.Watch Camino hang, requiring force-quit.
3.

Actual Results:  
Camino hung, requiring a force-quit.

Expected Results:  
Camino should have loaded the page and allowed me to browse, as Safari does.
Computer running Mac OS 10.3.9.
WFM on the branch, 2005092704 (v1.0a1+).  Flash starts sucking CPU after a
while, but that's it; no hang. 
Version: unspecified → Trunk
bug was filed with trunk, can someone try that?
WFM on the trunk, 20050928 Camino/1.0+.
Mac OS 10.3.9.

But flash starts sucking CPU 70 ~ 90%.
Reporter, does this still happen with for you with a current trunk nightly?  How
fast is your Mac and how much RAM does it have?
On my 1,5 ghz G4 with 512 MB RAM and mac os x 10.4.2
Camino 2005102404 (1.0a1+)
the browser on that page uses an average of 70% of CPU. The browser becomes much unresponsive, any interaction requires some secs (scrolling down, going back).
It certainly seems a lot slower now than it did when I first looked at it (again, branch builds rather than trunk).  Still no hang, nor trouble scrolling, but I did see the beachball once.

Fx 1.5b2+ also takes 70-80% of CPU on that site.

Reporter, does it consistently hang for you?
Sorry for the delay, Smokey.

Page is unresponsive as before.  Camino doesn't crash the machine, but any sort of browsing on the page is d**n near impossible.  I'll just re-echo Fabio's comment.  Camino becomes downright unresponsive, while Safari will let me use the site in question.  Admittedly this site is pretty Flash-intensive.

iMac G4/800, 768 MB RAM, Mac OS 10.3.9.  Camino version used for this test was the "Nightly Build", downloaded straight from the Camino main page -- 2005102404 (v1.0a1+).  I think this is the same as Fabio's.
(In reply to comment #7)
> Fx 1.5b2+ also takes 70-80% of CPU on that site.

We have a bug for ffox flash slowdowns: #285275.
Is this more suitable in the Core product? Flash seems a general problem on mac.

This is also related to (or the same as) bug #303135.
Blocks: 225397
Status: UNCONFIRMED → NEW
Component: General → Plug-ins
Ever confirmed: true
Summary: Site hangs Camino on opening page → Bad performance with Flash
(In reply to comment #8)
> Page is unresponsive as before.  Camino doesn't crash the machine, 

You originally reported the hang with a trunk build, but you tested today with a branch build :-)

> I'll just re-echo Fabio's comment.  Camino becomes downright unresponsive

Your Mac is slower than mine (PB G4 1.33 GHz, 512 MB, 10.3.9), so it makes sense that it's less responsive for you than for me, but Fabio's is faster than mine, which doesn't make sense at all, unless things are worse in 10.4.x.
Severity: critical → major
Keywords: perf
(In reply to comment #10)
> You originally reported the hang with a trunk build, but you tested today with
> a branch build :-)
> 

The one day I grab a trunk build... sigh.  Anyway, I grabbed a trunk build (Mozilla/5.0 (Macintosh; U; PPC Mac OS X Mach-O; en-US; rv:1.9a1) Gecko/20051025 Camino/1.0+) and retested.  No crash (because I knew how to get out of this bug), huge performance hit.  

I guess this means there are issues with both branch and trunk.
(In reply to comment #10)
> Your Mac is slower than mine (PB G4 1.33 GHz, 512 MB, 10.3.9), so it makes
> sense that it's less responsive for you than for me, but Fabio's is faster than
> mine, which doesn't make sense at all, unless things are worse in 10.4.x.

It depends on what do you mean by "no trouble scrolling". The browser scrolls right to me too, but the delay between the click on the scrolling bar and the actual scrolling is about 3 secs. And the more the page is flash-intensive, the more this happens. The browser never hangs in my experience, but with some very intensive Flash pages the delay for any action can go up to 20 secs (and this could be perceived as an hang).
Maybe the Flash plugin version is involved. I have 8.0 r22.
(In reply to comment #12)
> It depends on what do you mean by "no trouble scrolling". The browser scrolls
> right to me too, but the delay between the click on the scrolling bar and the
> actual scrolling is about 3 secs.

Ok, I definitely see that now.  I'm not sure I did before, but that is noticeable.

> Maybe the Flash plugin version is involved. I have 8.0 r22.

I still have the older Flash 7r24 or whatever, but now that I can reproduce the scrolling, I don't think the Flash version makes much difference.
(In reply to comment #13)
 
> I still have the older Flash 7r24 or whatever, but now that I can reproduce the
> scrolling, I don't think the Flash version makes much difference.
> 

It does make a difference; how much, I'm not sure.  From the Macromedia website:

"Flash Player 8 contains numerous optimizations to its rendering and code execution pathways, including a new garbage collector which reduces memory consumption up to 50%."

Surf here:  http://www.macromedia.com/software/flashplayer/productinfo/features/  ; launch the "Interactive Features Demonstration."  Hit the "Rendering Performance" demo when it launches in a new window.  Compare and contrast FP 7 after the FP 8 demo has run for a little bit.

On my machine (the same one used to report this bug initially), the FP 7 demo might as well hang the machine.  If I wait about 20 seconds or so, I can get out of it without having to force-quit.  The FP 8 side also gives a performance hit, but it's not nearly as bad as with FP 7.  I have to wait between 5 - 10 seconds.

I'm not sure what FP I had when I originally reported this bug.  I now have FP v 8.0 r22, which is what I'll use for any future reporting/testing.
Here's some Flash that will just bring the browser to its knees:
http://www.zefrank.com/string_spin/spin_1.html
Draw some stuff, then click the button to spin it. Looks like it's just drawing more and more lines every time.
*** Bug 317490 has been marked as a duplicate of this bug. ***
*** Bug 327228 has been marked as a duplicate of this bug. ***
Another site with lousy Flash performance:

http://www.powerlineblog.com/powerlinenews/video2/

I'm not asking you to agree with their politics; just to check Camino performance with this version of Flash.

Tested with official Camino nightly build 2006080201 (v1.0+) running under Mac OS 10.3.9 (with all current security updates) on an iMac G4/800 this day.  Player loads fine, but playback with Camino stutters, and machine becomes pretty unresponsive.  No crash, though. (yay!)
Retested site in comment #1 with Camino nightly 2006091501 (v1.0+), Mac OS X 10.3.9, QT 7.1.3, Flash player 9,0,16.

Camino's performance with Flash is still lousy on this site -- concluding site's Flash animations are processor-intensive.  After opening this site in a tab, I was able to close the tab, though it took a second or two.  No crashing or force-quitting needed.
same for me....camino 1.0.3 and previous are unusable with flash sites..pages are almost impossible to scroll...and I'm using 1.42ghz mac mini osx 10.4.8  with 1 gig ram...

tried with firefox, problem disappeared...
Tested above URL with Camino TRUNK nightly build 2006102501 (v1.2+) under Mac OS 10.3.9, Flash 9,0,16.  Flash performance *vastly* improved on this build.  Previous Flash performance issues noted on this site appear addressed, with this following caveat.   Camino seems to have trouble rendering the website in question.  The website is usable, but I am not sure if the fault is Camino's, or the website designer.  Not sure if I should break this out as a separate bug.

Screenshot attached.  Will test with branch build in the next week or two.
A new Flash player version was announced, so I installed it, and tried this bug again.

Tested again on trunk build 2006111420 (v1.2+) under Mac OS 10.3.9, Flash Player 9,0,28.  Original URL that caused bug now performs usably with iMac G4/800.  Rendering anomaly has been fixed.  Site mentioned in comment #15 also quite usable on same class machine, same software.

Have not yet gotten around to testing this on branch.
*** Bug 359968 has been marked as a duplicate of this bug. ***
Closing WFM based on comment 22 and my own trunk testing. Flash is way better in trunk.
Status: NEW → RESOLVED
Closed: 17 years ago
Resolution: --- → WORKSFORME
You need to log in before you can comment on or make changes to this bug.

Attachment

General

Created:
Updated:
Size: