Hi, I’ve got an HP 17-y011cy, a laptop with an AMD A12-9700P. I just got it less than 3 months ago, but it’s a refurb that only had a 30day warranty. Within that 30day window, I thought that my slowdowns and temporary lockups were caused by me or the OS, and I re-installed multiple times, but now I’ve got thermal monitor and system load widgets running.
They show I’m hitting hi temps frequently, and the slowdowns happen in the 150s (Fahrenheit) and lockups in the 160s. The temperatures occur even when there’s little load on the cpu, like when I’m just playing solitaire. Curiously, the GPU temp usually leads the charge up the thermometer, and I’ve never done anything to stress that, just basic compositing effects in KDE like wobbly windows.
Since I’m stuck with the hardware, I wonder if there’s any way to use software to slow down the APU, to keep it cooler. I see a cpufreq panel applet for xfce, bit nothing obvious for KDE. Do I need to switch desktop environments to get this feature?
PS: I’ve read online about other HP customers complaining about the same issue with my model of laptop, even though they use Win10.
PPS: I tried this as superuser, and you can see it didn’t work:
**purplebox:/home/gef #** cpupower frequency-set -u 2.10 GHz
Setting cpu: 0
Error setting new values. Common errors:
- Do you have proper administration rights? (super-user?)
- Is the governor you requested available and modprobed?
- Trying to set an invalid policy?
- Trying to set a specific frequency, but userspace governor is not available,
for example because of hardware which cannot be set to a specific frequency
or because the userspace governor isn't loaded?
cpupower frequency-info says max base frequency is 2.50GHz and max boost is 3.60 GHz; I thought pulling the max down to 2.1 (thus completely eliminating any boosting) might be a good place to start checking if I can prevent the overheating this way, but I’m npot sure how to enable such control. Does that only work with Intel?
Hi, I’ve got an HP 17-y011cy, a laptop with an AMD A12-9700P. I just got it less than 3 months ago, but it’s a refurb that only had a 30day warranty. Within that 30day window, I thought that my slowdowns and temporary lockups were caused by me or the OS, and I re-installed multiple times, but now I’ve got thermal monitor and system load widgets running.
A refurb…hopefully this involved blowing dust out from the vents etc. Other things to check include heat-sink seating, or the thermal paste may need redoing perhaps.
BTW, more details about the graphics chipset and the video driver is in use might be useful here
/usr/sbin/hwinfo -gfxcard
inxi -G
Do you know about (or considered using) thermald? AFAIU, it works with Intel CPUs but has support for some AMD CPUs as well. Others feel free to clarify/correct/expand here.
I didn’t know about thermald; sounds like just what I’m looking for. I’ll hope there’s a youtube guide for the heat sink and thermal paste access - I miss laptops that were easy to open!
hwinfo says my APU is Carrizo; internet says it’s Bristol Ridge:
If you go that route I suggest you first make sure what thermal interface the laptop uses.
I had an old HP Pavilion laptop with AMD CPU and NVIDIA chipset/GPU (yes, that old), that started heating up a lot, to the point it couldn’t be comfortably held in your lap.
It used a thermal pad that had burnt out. The thing is a few millimeters tick, and I couldn’t use thermal paste because the cooler wouldn’t press strongly enough on the cpu. I almost fried the thing in a few minutes.
I took me a long time to find another thermal pad, but after that the laptop worked for a few years more.
Sorry, missed that. Yes, fans working - I hear 'em get louder when the temp goes up and feel the hot air blowing from the vent. I’ve blown canned air through but seen no dust come out - as I said laptop is less than 3mo old and problem has been intermittently present since the beginning. -GEG
Okay, I installed it, ran sytemctl to enable and start, ran the command itself and got output that it’s saemonizing, and then service --status-all shows no listing for thermald.service. Ironically, right after I entered the command to list services, I had a brief hangup to tell me it wasn’t working even before I scanned the output. How do I make it go?
Thanks for the tip. I’ve a strong suspicion that my laptop was never actually used, that it sat in a box for a year in some low-volume store before being returned to HP and then offered to newegg as a refurb. I doubt I’ll find any such thing as a worn out pad - based on what I’ve seen from other dissatisfied customers, some of these things are just lemons.
Hi, I’ve got an HP 17-y011cy, a laptop with an AMD A12-9700P. I just got
it less than 3 months ago, but it’s a refurb that only had a 30day
warranty. Within that 30day window, I thought that my slowdowns and
temporary lockups were caused by me or the OS, and I re-installed
multiple times, but now I’ve got thermal monitor and system load widgets
running.
They show I’m hitting hi temps frequently, and the slowdowns happen in
the 150s (Fahrenheit) and lockups in the 160s. The temperatures occur
even when there’s little load on the cpu, like when I’m just playing
solitaire. Curiously, the GPU temp usually leads the charge up the
thermometer, and I’ve never done anything to stress that, just basic
compositing effects in KDE like wobbly windows.
Since I’m stuck with the hardware, I wonder if there’s any way to use
software to slow down the APU, to keep it cooler. I see a cpufreq panel
applet for xfce, bit nothing obvious for KDE. Do I need to switch
desktop environments to get this feature?
zypper in https://download.opensuse.org/repositories/home:/malcolmlewis:/TESTING/openSUSE_Tumbleweed/noarch/amd_gpu_power_profile-0.0.1-1.4.noarch.rpm
Test manually, you should see temperature drop… if all good enable
the systemd service.
–
Cheers Malcolm °¿° SUSE Knowledge Partner (Linux Counter #276890)
SLES 15 | GNOME Shell 3.26.2 | 4.12.14-23-default
If you find this post helpful and are logged into the web interface,
please show your appreciation and click on the star below… Thanks!
I’m not sure how to enable in systemd, but I’m not sure I should, either.
I set all 3 modes to low, fired up Tetris (kblocks), and a few minutes later the blocks stopped falling. So I checked my thermal monitor widget and it showed 150-something for the CPU and 1607°F for the GPU.
Now, I know that after I install anything, I’ll get some action from snapper, too, as it deletes the oldest snapshot, but between installing the script and the lock-up, maybe 10min, my CPU and RAM utilization were under 20%.
And thanks for the systemctl tutorial. I used it to check status on thermald, and evidently it’s not running because it doesn’t like my CPU model, but it never threw an error to stdout.
Can’t seem to select battery mode even though it’s indicated in /etc/sysconfig. I rebooted on battery and still show performance mode:
**purplebox:/home/gef #** amd_gpu_power_profile
Usage: /usr/sbin/amd_gpu_power_profile <OPERATION_MODE> <PERFORMANCE_LEVEL>
Valid modes:
battery
balanced
performance (current)
Valid performance levels:
auto
low (current)
high
**purplebox:/home/gef #** systemctl status amd_gpu_power_profile
● amd_gpu_power_profile.service - GPU Power Profile service for radeon and amdgpu kernel modules
Loaded: loaded (/usr/lib/systemd/system/amd_gpu_power_profile.service; enabled; vendor preset: disabled)
Active: inactive (dead) since Thu 2018-07-26 10:46:54 PDT; 42s ago
Main PID: 1399 (code=exited, status=0/SUCCESS)
Jul 26 10:46:54 purplebox systemd[1]: Starting GPU Power Profile service for radeon and amdgpu kernel modules...
Jul 26 10:46:54 purplebox systemd[1]: Started GPU Power Profile service for radeon and amdgpu kernel modules.
The highest I have seen in the thermal monitor widget is less than 170°F, which is to say 76°C, well under the critical temperatures listed here. But as I stated above, I get slow-downs/freezes in the 150’s - around 70°C. Is that normal?
The sensors the widget reads are acpi/Thermal_Zone/0_acpitz and 1_acpitz (reports higher of pair), and lmsensors/amdgpu-pci-0008/temp1 and vddgfx and vddnb (reports highest of this group). Usually they’re with a couple degrees of each other, but under modest CPU load as immediately upon bootup, the acpitz sensors lead by 20°F, and at other times for no apparent reason the gpu sensors lead the climb, and these are the ones where the lockup occurs. This may be at say 167°F for highest gpu sensor and 150-something for higher acpitz sensor.
Thanks again, but I can’t seem to change mode. As a daemon, it loads and dies - see prev post. Evidently loading manages to change the perfomance level to low, but mode stays on performance. Even if I reboot on battery power, I’m not in battery mode. -GEF
