my cpu temp in Suse 11.2 is over than 55 C (sysem don not have any task) , but in windows 7 temp is 42 C
in fedora is 46 C
I have new fresh install of suse 11.2
my cpu is : AMD Phenom™ II X2 550 Processor
what my system working warm ?
amir33647 wrote:
>
> my cpu temp in Suse 11.2 is over than 55 C (sysem don not have any
> task) , but in windows 7 temp is 42 C
> in fedora is 46 C
> I have new fresh install of suse 11.2
> my cpu is : AMD Phenom™ II X2 550 Processor
> what my system working warm ?
- electricity consumption.
- whatever your system is doing
- poor cooling
- environment
If you’re asking why your three different systems are reporting
different temperatures, it’s anybodys guess. Are these temperatures
read with lmsensors and have you let the system achieve the same
operating conditions after each reboot?
/Per
–
Per Jessen, Zürich (32.0°C)
http://en.opensuse.org/User:Pjessen
your AMD should throttle down when the load goes down…mine does,
maybe you have not set yours up to do that…
bottom right of your screen i guess you have a icon on the task bar
that looks kinda like a power plug…if you mouse over it it might say
something like “Plugged in” or “On Battery”… and, if you right click
on that icon you will get a chance to “Configure Power Save” or “Set
CPU Policy”…
set it as you need and see if that helps any…
if not, post again…
–
Enjoy Packman’s benefits? Show your appreciation by donating at
http://se.unixheads.org/?donate
DenverD
CAVEAT: http://is.gd/bpoMD [posted via NNTP w/openSUSE 10.3]
Sorry to jump in and ask another question…
What tools are you using to monitor CPU temperature within openSUSE? I’m looking to overclock my new build and need overclocking tools, like CPU temp monitoring.
Thanks!
@topic: different temperatures are likely to be caused by different cpu-profiles (e.g. “ondemand” vs. “performance”).
@linuxvinh: overclocking itself should rather be done within the BIOS. To monitor temperatures, workload etc. I use →GKrellM, cpu-profiles can be checked and set easily by using ‘cpufreq-info’ and ‘cpufreq-set’ (both are provided by the package ‘cpufrequtils’).
I am a commercial A/C technician, so temperature measurements is something I am familiar with. In a case like this the first thing I would do is determine if, a) the sensor is accurate (do this by measuring the resistance and looking it up in a chart) and b) what is the temperature of the medium being measured (do this by using a known accurate device suited for the purpose; ie air flow, chilled water, etc.) If both these check out the system is at fault, which usually means replacing some board or calibrating the device.
Unfortunately, neither a or be is really possible in a computer, as you cannot get to the CPU core nor that I know of can you measure the resistance of the sensor in the mobo. I do not even know how the devices report temps; I assume they interpret the sensor and report it as degrees, not resistance, but I do not know. I do know I have seen posts on this subject stating that there is a great deal of variability in the accuracy of these readings from devices. I would assume newer devices are better at this, but I do not know anything about which devices are less accurate. Besides, you have to have some application to dig out this info from the devices and put it into a displayable form; again, these apps may vary in accuracy. Or the short version: how do you know the different temps are really accurate? Bottom line, you can’t. BTW, anyone who has a more detailed knowledge of devices and apps in this regard I would love to know more myself.
All that said, there is something you can easily do in suse to see about this. Opem System Monitor and look at the CPU usage for a while; if nothing is going on it should be very low. Look at all the processes running for things you don’t need and see if you can disable them (a lot of times you can’t, but usually they are also not doing anything and only take up a very little memory). If, however, your usage is consistently over 25%, say, look to see what is using all those cycles and stop it/them. Watch that temp go down. I do not use gkrellm because it gives much more info than I normally need; instead I use Ksensors, which is a graphical interface for lmsensors, which I’m told is very accurate, and I set it to display only the CPU temp in an Icon, and go red over a certain temp. When I see that temp go up and nothing much is going on I immediately go to sysmon (I keep it running all the time, too) and shut down the processes at fault, usually kio-file, which seems to good for not much but cannot be disabled; if anyone knows how, please tell. Sometimes other processes get hung up, probably for not closing properly.
BTW, 55 is not bad; mine runs around 52-55 when not doing much and climbs over 65 when really busy; of course our room temps are fairly warm to conserve money, 80-86f in the daytime and this drives internal temps up.