On Sat, 3 Nov 2007 20:38:16 +0100 Beso <[EMAIL PROTECTED]> wrote: >> This is a laptop with a tiny fan and 53º C while processor is idle is >> understandable since cooler is on passive mode >> most of time. But 73º C looks way too much for me. > this for me seems the thermal zone, and not the core temperature... No, they are core temperatures:
#ls /sys/bus/pci/drivers/k8temp 0000:00:18.3 bind new_id unbind /\ Also if you read k8temp doc in kernel src: "Mapping of /sys files is as follows: temp1_input - temperature of Core 0 and "place" 0 temp2_input - temperature of Core 0 and "place" 1 temp3_input - temperature of Core 1 and "place" 0 temp4_input - temperature of Core 1 and "place" 1" So I'm sure /sys/devices/pci0000:00/0000:00:18.3/temp1_input is a core temperature. > anyway, install a monitoring daemon (kima for kde is very light/good and > integrates into the kicker) or any other like ksensors or others (i don't > know the gnome ones since i completely dislike gnome but there are a lot of > superkaramba widgets to do this) and see there your effective temp. > there you should see: > the core temp. mine turion 64 single core go from 30-35° idle to 60-65° C > when at max speed and max load. > the thermal goes from 58° when the fan starts to 75° when the processor goes > in passive mode decreasing speed for avoiding overheating problems. at 105 > it shuts down. They are just a frontend to lm_sensors which, in place, for k8temp sensor gets values from sysfs. Values provided by lm_sensors are basically the same as /sys/devices/pci0000:00/0000:00:18.3/tempX_input > now, i've heard that dualcores can have a temp of 5 to 10° higher than the > single cores, but i don't know if this is really true. > anyway, how do you know that that 2 devices are effectively the processor?! # lspci -v -s 00:18.3 00:18.3 Host bridge: Advanced Micro Devices [AMD] K8 [Athlon64/Opteron] Miscellaneous Control /\ > your problem is that acpi doesn't get the processor states. if it doesn't > provide you with trip points in /proc/acpi/thermal/thrmx/trip_points then > you have an acpi problem. this may be caused by bios or by wrong kernel > config. > make sure that your thermal is compiled and started and that your bios is ok > reading and doing this: > http://gentoo-wiki.com/HOWTO_Fix_Common_ACPI_Problems > the steps needed are the ones till you recompile the dsdt and see if the > compiler raises errors. if it does then you might have a bios problem. try > to fix it according the infos there and recompile. I have processor states working and thermal_acpi is built into kernel. I've already fixed some dsdt problems but they were not related to thermal management, maybe this laptop ACPI doesn't have thermal management at all. > if the bios is ok or the fixed one doesn't make the processor work with the > latest kernel then search on the kernel.org the acpi bugzilla or the mailing > list and report there your problem. they could help you fix it or they could > patch the actual acpi to make this work. > if you have an acer notebook, this might be the problem, since i've > encountered many acer laptops that don't read the thermal correctly. The strange thing is that this laptop ACPI do have an OSI specification for Linux, it should be fairly compatible with it. > -- > dott. ing. beso > -- Angelo Arrifano AKA MiKNiX CSE Student at UBI, Portugal Gentoo Linux AMD64 Arch Tester [EMAIL PROTECTED] http://miknix.homelinux.com PGP Pubkey online - - People of privilege will always risk their complete destruction rather than surrender any material part of their advantage. -- John Kenneth Galbraith -- [EMAIL PROTECTED] mailing list
