MintyBloke wrote:I recently installed LMDE and applied the updates. My laptop is 32-bit Intel Dual Core T2500 CPU (Dell Inspiron 6400, from 2006).
Initially the System Monitor showed just one CPU.
I followed the LMDE "Welcome" splash link to known problems (http://www.linuxmint.com/rel_debian.php) and took note of the item titled "Multi-core and multi-CPU support in 32-bit kernel". I performed the advised changes:
apt install linux-headers-686-pae linux-image-686-pae
After doing this the System Monitor correctly shows two CPUs.
All is generally OK, with the two CPU's workload generally balanced.
The exception, though, is some strange behaviour that occurs predictably when using Firefox v26.0 with the LastPass add-on v2.0.20.
When LastPass does anything network-related the two CPUs perform a game of "After you. No, I insist, after you". Basically, the CPUs take it in turns to run at 100 percent while the other runs around 12 percent, then they swap over and do it again. Each "turn" lasts between 1 and 10 seconds. This continues for between 10 and 150 seconds, during which Firefox is totally unresponsive. The CPU graph in System Monitor looks like a string of sausages of varying length with the plot of each CPU an almost perfect mirror image of the other (mirrored at approx the 55 percent mark).
It happens every time I do any LastPass activity that requires an update of my vault, e.g. editing a record or even just logging into a web site using credentials stored in LastPass (probably because of a "last used" update for that record).
This seems to be primarily a Firefox/LastPass problem, except that I have noticed some of the same behaviour in other situations, but these are rare and short-lived.
I've reported the problem on the LastPass support forum, but I am wondering if the underlying cause is related to the installed linux-headers-686-pae and linux-image-686-pae mentioned above.
I've been using Linux on and off for the past 15 years, but I'm not deeply technical. Any help with this would be appreciated. Thanks.
MintyBloke wrote:I appreciate the reply (although I'm not sure a quote of my entire post was really necessary), but in my experience if software is not written for multi-core CPUs it uses just one of them. This is not the behaviour I have described.
Users browsing this forum: CMAR606 and 9 guests