Firefox and chrome chewing CPU since updates

[Nataraj]

On 06/30/2012 10:32 PM, Paul Smith wrote:
> So, I was away all week on vacation (during that time my system was
> halted and unplugged) and when I got back I had 56 or so updates waiting
> for me (Ubuntu 12.04 64bit) including a kernel update and various other
> things.
>
> I applied these and rebooted.  Now whenever I start a browser, either
> the default Firefox or Chrome (I use the main/stable version from the
> dl.google.com repository), it sucks up TONS of CPU.
>
> Firefox is usually chewing at least 60% CPU and sometimes jumps to 120%
> or more (I have an Intel Core2 Quad).  top(1) says my system load is
> constantly at 1.3 or above.  This is with no extra tabs open and
> about:blank on the single tab I do have.
>
> Chrome is even worse, since it's multi-processing: when I start it there
> are 3 to 4 chrome processes running each taking between 25% and 45% of a
> CPU, and my system load goes up to 2.2 or higher, constantly.
>
> The browsers seem to WORK just fine; I can surf the web and no problems
> with internet connections, etc.  Just my system is running like crazy,
> not doing anything.
>
> This was never a problem before these updates.  Has anyone else seen
> this, or have any idea what's going on?  It's very odd that it's both
> Firefox and Chrome, I admit, but if I stop them the load goes down and
> none of my other apps (which, admittedly, I don't use too many different
> apps--mostly Emacs, rxvt, Evolution, etc.) seem to show this problem.
>
> Has anyone else seen this behavior?  I've tried strace on these but see
> nothing interesting: chrome seems to be hanging on a futex (but not fast
> enough for the kind of load I'm seeing) and Firefox seems to be failing
> to read from a socket (but again, not fast enough for the load).
>
> This is extremely frustrating.
>
>

Yes.  I saw the same behavior with 10.04.   Chromium, firefox and
thunderbird were all affected.  I rebooted a second time and the
situation seems to have improved.  I'm unsure yet whether there is still
a problem.  It seems that it was either the kernel or the microcode
update that caused the problem.

Nataraj