High CPU temp after resume from suspend....
I have googled on this issue for the past couple of days and have found no solution (Slackware 13.37 on a Thinkpad T60p with the generic-smp kernel).
A fresh boot gives me temps of about 48C at the highest. "Normal" usage is web browsing with Firefox, a couple terminals open, and that's about it.
I have run through powertop, using the suggestions, and it doesn't seem to affect CPU temp at all.
Before and after suspend cpufreq-info says "ondemand". Changing this to anything else seems to make no difference.
Before suspending (after fresh bootup):
If i understood you well suspending increases CPU_1 (second core) temperature? maybe process that wakes up computer heats up CPU, or just CPU heats up when fans are turned off?
Run your computer 24/7 with distributed computing
I can’t test that for you on my ThinkPad T60 because I gave it to my girlfriend two weeks ago so I tested that on ThinkPad T61.
It runs Slackware Linux, Window Maker, and 15 dockable applications (15 instances of wminfo).
After the boot the fan speed was equal 0 and the machine reported 48–49 C for both cores of my CPU.
The fan started to work about 49 C.
Five minutes later the machine reported 47–48 C.
So I started four xterms:
– running Midnight Commander,
– running ydpdict (a console shell for Collins Dictionary),
– running mocp via SSH on a remote machine,
– running root’s console with Midnight Commander.
Five minutes later the temperature was still in the 47–48 C range.
So I started Firefox running two tabs: both displaying LinuxQuestions.org website.
Five minutes later the temperature was in the 48–50 C range.
The cpufreq-info said “ondemand” for both cores.
So I suspended the machine using the command: “echo -n mem > /sys/power/state”.
Five minutes later I woke up the machine.
At the beginning it showed 43 C and the fan speed was equal 0 – then the temperature started to increase.
About 49 C the fan started to work.
Five minutes after I woke up the machine the temperature was in the 48–50 C range.
Maybe some process increases the temperature of your CPU after resuming from suspend. I use Ethernet connection to the router, simple window manager, 15 simple applications, and four simple xterms running four simple programs (plus Firefox running two tabs with the simple website). If you use Wi-Fi, KDE, and open in Firefox some websites with animations your results may vary from mine.
Thanks for the replies,
If there is "some process" that heats up the CPU, or one of its cores, that process is not showing up in top(1).
I have played with CPU "stepping" in the BIOS. I have followed detailed instructions to activate the suggested "fixes" via powertop. I have tested the relevant cpufreq governors. I have blown out the dust bunnies and re-applied the thermal pastes. I have tried "/etc/rc.d/rc.acpid restart". I even went so far as to backup, and upgrade my system to 14.0(32-bit, an amazing bargain at $52ishUSD, thank you very much!), and have since actually seen a slight decrease in post-resume-from-suspend temps with 3.2.29 but only after running "/etc/rc.d/rc.local restart" which contains my powertop "fixes". How or why they seem to get forgotten after resuming from suspend is far above my pay bracket....Also, FWIW, I'm now using the "conservative" cpufreq governor since it seems to make the temperature drop back more quickly than does "ondemand" (though, obviously this is highly subjective, non-scientific, and shouldn't be trusted by anyone reading this ;-)
I do know this is not a Linux-specific problem, though, as many of the google posts related to this issue have been in regards to Windows.
I guess it's just one of those things. Since it's not overheating, I think I might let it go for now and just enjoy my "vintage" lappy. Though, if anyone has a eureka moment, or a fix that worked for them, (aside from making the fan run full-speed all the time) I'd be very grateful!
I understand your frustration but it doesn’t solve your problem.
Looking at your temperatures report I see that the GPU temperature is in your case about 15°C higher than the CPU temperature. In a normal situation the GPU temperature should be about 10°C lower than the CPU temperature (I own a few different ThinkPads). So it seems something uses your GPU a lot. Maybe that’s some clue...
If you put in your /etc/rc.d/rc.local something that decreases the temperature and after the resume from the suspend the system “forgets” that, call /etc/rc.d/rc.local from your suspend script. For example:
Don’t care whether we believe you or not. Trust yourself. If the “conservative” mode works better for you than the “ondemand” one use the former one.
T60p is a nice machine, had one at work for a while until the company forced me to replace it. Unfortunately it was known they tend to run hot. How handy are you ? Does not take much for dust to build upon the fan and the CPU. I would
remove the keyboard and clean out any dust that built up. I have done that a few times and it helps out a lot.
It seems that according to that post STDOUBT did that and much more.
|All times are GMT -5. The time now is 02:53 AM.|