Quite some time ago I wrote up a post concerning a “fix” for Firefox 3.0 that involved setting “browser.cache.memory.enable” to “False” in “about.config” (Firefox’s options editing dialog). This “fix” solved the ongoing problem with Firefox 3.0.* causing unacceptably high CPU usage when a page was loaded into a foreground tab. A problem which not only slowed the browser to a crawl but caused your CPU to bake itself as well, something that no user ever wants to see.
This problem showed up in the middle of the development of Firefox 3.0 and although did not affect all users it affected a rather large portion of them, so much so that I received over 60 comments on that original post of mine (posted on both of my sites), the majority of which stated that the “fix” worked beautifully. Bug reports were posted and generally ignored and Firefox 3.0 was released without this problem being fixed however.
Then came the development of Firefox 3.1 and along with it came the same exact problem which was still able to be fixed by the above solution until one day the solution no longer worked. After some digging I found that it wasn’t the same bug causing the High CPU usage in the exact same manner as the the original but a whole new (unknown) problem altogether and that’s where my experience in the trade ran out. Too make a long story short, I took the time to trace the regression range (between which build# worked okay and which build# did not) down to two sequential hourly builds, one which did not exhibit the new problem and one that did. The related forum was posted to and a bug report was then filed but that was where things came to a seeming halt—it also virtually halted my testing as well. I couldn’t test a broken browser and no one was apparently listening.
Note: It might be of interest to note that I am not a programmer or developer, I’m what used to be called a beta tester which is not the same thing. There are those who can code and those who cannot. There are those who can troubleshoot in the field as well as the lab and those who cannot. I’m the latter in the former and a former in the latter. Thank you very much.
So here’s the 3.1 development producing build after build with not only the original problem as stated above but now with an additional problem that would make Firefox 3.1 virtually unusable for a rather large group of users once they upgraded unless both these problems were tracked down and fixed.
I didn’t abandon the Mozillazine forums though. In fact, I actually checked in on a regular basis and reading through the list of bug landings (fixed bugs) for any clue that things had been put to rights. I also installed various nightly builds of 3.1 from time to time to see if they fared any better but all to no avail and no follow up to the bug report even to tell me that I was way off base.
Okay…so it’s fixed…apparently.
I hadn’t tried the nightly builds of 3.1 for over two weeks so I really don’t know when these two issues were fixed nor the whys, how’s, or wherefores behind it but with yesterday’s nightly build of 3.1 (Minefield) with nothing but a default installation and settings, 3.1 beta2-pre fired up without a problem, loaded up my usual tabs and my CPU barely woke up for it. And so far I haven’t seen hide nor hair of those infamous problems (knock on whatever wood I can find). Now I can finally get on with the testing without fear of baking my poor single core CPU into something that resembles a hunk off a burnt steak fry.
Life is good again.
So what version will have the fix? I have been suffering with this problem and it is a real PITB. I have version 3.0.4.
How do I get to the 3 RC1 location bar? Help does not find anything.
@blake: Always glad to help.
@Jerry: Currently, FX 3.1 latest trunk builds have not exhibited this problem but I’m not sure as to what exactly the fix(s) was at the moment. The bottom line here though is that 3.1 should finally put an end to this and many other long time bugs. And I’m not sure what you mean by “…get to the 3 RC1 location bar?” Maybe if you explain a bit differently I can answer your question.
Thanks for blogging on this issue. I’ve been dealing with random lockups of firefox on my Macbook Pro laptop ever since upgrading to 3 from version 2, most especially in gmail with nary a clue where the issue was coming from. Anyway, your earlier post’s config changes fixed my lockup issues and I am glad to see the 3.1 trunk shows signs of this issue being addressed.
I have not done tests to prove, but apparently the excessive use of CPU has to do with the damn flash. I stopped using Firefox and was using Netscape and Opera, the excessive use of memory occurred in the same way. I believe that when the window is minimized the trash in Flash continue running (using CPU), even doing nothing useful. In my view, when you minimize the window such filth should be put into total inactivity (NO ONE CPU usage).
LC – There’s been a long time problem with Adobe’s Flash plugin for Firefox which is also used for Opera, Netscape and Google Chrome as it always causes high CPU usage when Flash is present in a website page. ironically enough, this high CPU usage does not often occur with the version made for Internet Explorer which is a much more tested and quality controlled version than the other non-IE version (after all, they get paid for the IE version?).