Re: reading on-die temperature sensor of AMD A10 5700 processor

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

 





On Wed, Aug 14, 2013 at 1:30 AM, Guenter Roeck <linux@xxxxxxxxxxxx> wrote:
On 08/13/2013 10:17 PM, Kapil Dev wrote:

On Aug 14, 2013, at 12:56 AM, Guenter Roeck <linux@xxxxxxxxxxxx> wrote:

On 08/13/2013 08:09 PM, Dev, Kapil wrote:


On Tue, Aug 13, 2013 at 11:02 PM, Guenter Roeck <linux@xxxxxxxxxxxx <mailto:linux@xxxxxxxxxxxx>> wrote:

    On 08/13/2013 07:41 PM, Dev, Kapil wrote:

        Thanks again Guenter!




        On Tue, Aug 13, 2013 at 5:51 PM, Guenter Roeck <linux@xxxxxxxxxxxx <mailto:linux@xxxxxxxxxxxx> <mailto:linux@xxxxxxxxxxxx <mailto:linux@xxxxxxxxxxxx>>> wrote:

             On Tue, Aug 13, 2013 at 05:11:31PM -0400, Dev, Kapil wrote:
              > Thanks for the response Guenter!
              >
              > I put the system on load and I noticed their value hardly changes. To my
              > surprise, I ran a benchmark for 3 minutes and none of temp*  values changed
              > during execution. The highest reported temperature is always 48C.
              >
              > I was expecting that the internal temperature would have "sensor = thermal
              > diode" or "sensor = core" as a keyword. I thought "sensor=thermistor" means
              > the corresponding sensor is on the motherboard and not the internal
              > die-sensor.
              >
             Sure, but who knows if they put a sensor below the CPU for some reason ...

             I am not familiar with the A10 CPUs. Are those similar to K10 ?
             If so, maybe the k10temp driver works or could be extended to support it.
             Can you give it a try ?

             If it does not work out of the box, can you send me the output of "lspci -nn" ?


        Kapil: I believe K10 corresponds to A8 seried, and A10 has Bulldozer architecture (http://www.cpu-world.com/__CPUs/Bulldozer/TYPE-A10-__Series.html <http://www.cpu-world.com/CPUs/Bulldozer/TYPE-A10-Series.html>). I am not sure how different they are though. I tried K10, but it did not work out of the box. I am looking into k10temp.c driver now. Also,I am attaching the output of "lspci -nn" command herewith; you might have to open it using wordpad for proper formatting..


    Your lspci output includes

    00:18.3 Host bridge [0600]: Advanced Micro Devices [AMD] Device [1022:1403]

    which is supported by the latest k10temp driver. Support was added early last year.
    What is your kernel version, and can you switch to a more recent version ?


Kapil: I have 2.6.39-020639-generic kernel. Is it too old kernel for k10temp driver to work? Which stable kernel should I upgrade to?


You need 3.4 or later.

Kapil: I was trying to avoid updating the kernel because my current kernel is setup for running different type of benchmarks etc. Seems like I can't avoid updating the kernel; I will do it..

Alternative would be to back-port the k10temp driver from a later kernel. Then you can keep running the old one.


Kapil: Not sure how to do that. I am already taking the k10temp driver from here: http://khali.linux-fr.org/devel/lm-sensors/drivers/k10temp/ and compiling and using it.
 

Just for my understanding: Currently, when I run "sensors" command, I am getting three values (apparently wrong), temp1-3. So, what do you think is happening that is causing these values wrong? And, how is it going to be fixed by updating the  kernel? Although we don't know yet, but, I am hoping that by updating the kernel and loading the k10temp kernel, I would be able to see readings from thermal-diode (core-temperature) along with the current readings from thermal-resistors.

Not wrong; those are probably different chassis and/or board temperatures. Does the BIOS report
any matching temperatures ?

In BIOS, under PC health, I see only one temperature, named CPU Tctl (control temperature). Its value is in 30s (between 29-36 or so). I believe this value denotes the distance from a specific maximum temperature (Tmax). Finding that Tmax would be another challenge; but,  I not at that stage yet.

Thanks,
Kapil

 
Guenter


_______________________________________________
lm-sensors mailing list
lm-sensors@xxxxxxxxxxxxxx
http://lists.lm-sensors.org/mailman/listinfo/lm-sensors

[Index of Archives]     [Linux Kernel]     [Linux Hardware Monitoring]     [Linux USB Devel]     [Linux Audio Users]     [Linux Kernel]     [Linux SCSI]     [Yosemite Backpacking]

  Powered by Linux