Thermal error while compiling

[albatro74]

Hi, I'm trying to emerge packages for my funtoo installation but after a while the system crash due to bad thermal conditions. I suspect it has to do with kernel configuration. I've got an AMD A8 mobile and chose the piledriver stage3. Could someone help me with this?

[cardinal]

Is the processor fan running at full speed when the shutdown occurs?
Is the laptop case air inlets/outlets and heat sinks free of dust and dirt?
Emerge wgetpaste and post the links to the pastes in your reply:

lsmod | wgetpaste 

emerge -vv --info | wgetpaste

dmesg | wgetpaste

[albatro74]

https://bpaste.net/show/15c5efd794ba

https://bpaste.net/show/dbef3n6e820e

[albatro74]

Any idea about the thermal error? Could anyone reply to this subject?

[cardinal]

Decrease the CPU load and frequency will lower the temperature if you can't increase the airflow over the CPU's heat sink.

Post output of this command:

emerge --info | grep MAKEOPTS

[albatro74]

MAKEOPTS="-j5"

 

I've been working quite extensively on the subject lately. First of all here's the output of sensors

 

# sensors

acpitz-virtual-0

Adapter: Virtual device

temp1:        +79.0?C  (crit = +100.0?C)

 

k10temp-pci-00c3

Adapter: PCI adapter

temp1:        +79.0?C  (high = +70.0?C)

                       (crit = +100.0?C, hyst = +99.0?C)

 

radeon-pci-0008

Adapter: PCI adapter

temp1:        +74.0?C  (crit = +120.0?C, hyst = +90.0?C)

 

radeon-pci-0100

Adapter: PCI adapter

temp1:            N/A  (crit = +120.0?C, hyst = +90.0?C)

 

 

while acpitool returns the following:

 

# acpitool -e

  Kernel version : 4.6.2-gentoo-ge   -    ACPI version : 20160108

  -----------------------------------------------------------

  Battery #1     : present

    Remaining capacity : unknown, 100.0%

    Design capacity    : 2500 mA

    Last full capacity : 1889 mA, 75.56% of design capacity

    Capacity loss      : 24.44%

    Present rate       : 0 mA

    Charging state     : Full

    Battery type       : Li-ion 

    Model number       : AL12A32

    Serial number      : 5271

 

  AC adapter     : online 

  Fan            : <not available>

 

  CPU type               : AMD A8-4555M APU with Radeon HD Graphics 

  CPU speed              : 0x6001116 MHz 

  Cache size             : 1597.096 KB

  Bogomips               : 3194.19 

  Bogomips               : 3193.34 

  Bogomips               : 3193.36 

  Bogomips               : 3193.41 

  Function Show_CPU_Info : could not read directory /proc/acpi/processor/

  Make sure your kernel has ACPI processor support enabled.

 

  Thermal info   : <not available>

 

   Device S-state  Status   Sysfs node

  ---------------------------------------

  1. PB4  S5 *disabled  pci:0000:00:04.0

  2. PB5  S5 *disabled  pci:0000:00:05.0

  3. PB6  S5 *disabled

  4. PB7  S5 *disabled

  5. SPB0  S4 *disabled

  6. SPB1  S4 *disabled

  7. SPB2  S4 *disabled

  8. SPB3  S4 *disabled

  9. GEC  S4 *disabled

  10. OHC1  S3 *enabled   pci:0000:00:12.0

  11. OHC2  S3 *enabled   pci:0000:00:13.0

  12. OHC3  S3 *disabled

  13. OHC4  S3 *disabled

  14. XHC0  S3 *enabled   pci:0000:00:10.0

  15. XHC1  S4 *disabled

  16. P2P  S5 *disabled  pci:0000:00:14.4

 

so if I have understood correctly fan, processor and thermal modules (actually they are built in) have some serious problems.

Is there any chance that BIOS interfere with ACPI? Any other thought?

 

[albatro74]

Decrease the CPU load and frequency will lower the temperature if you can't increase the airflow over the CPU's heat sink.

Post output of this command:

emerge --info | grep MAKEOPTS

This is surely not a problem of air flow: Windows does not show any sort of problem and whenever I use it I never have such behavior. I also opened the case to check whether the fan could be the issue but eventually I didn't find anything about that. It definitely seems that there's something wrong with kernel and ACPI. Fan work but not vigin me any acceptable result: I'm trying to compile libreoffice and have to quit the compilation every 30 minutes to avoid the temperature rise over 80?C. Plus cpufreq seems to work properly...

[albatro74]

Any idea?