I’m not entirely sure what update started it, but over the last couple weeks my laptop started running hotter. I have a widget showing my CPU usage / temp / etc. on my desktop and my laptop always hovered around 45-deg C. Now it hovers around 50-deg C. This doesn’t seem like much but it’s enough for it to run the fans all the time now. I’ve. My use hasn’t changed (mainly browsing web, etc.). CPU usage is low, < 10% and I’m using only the iGPU.
My hardware is a 2020 HP Envy 15 x360 convertible. i7, 32gb, 2tb NVMe, Nvidia MX250. Nothing has been changed in quite a while on my system besides regular updates.
I don’t know where to begin to even troubleshoot. Although my laptop is about 4 years old, I did replace the fans about 1-year ago and I do blow air in there to keep it clean. It was a solid 45-deg C for “all of time” until just recently. Is there any settings I can check? I run KDE and x11 if that matters. Thanks.
I have the on-screen KDE widgets for CPU / etc. usage. Just the default ones available when you “edit” your desktop. I only mentioned that because they show low CPU usage, and of course top or btop seems to agree with it. My first thought was I had some processes running int he background using CPU so I keep the right 10% of my desktop visible. Usually usage is in the 2-5% range, and that’s normal but the temps have gone up. So I assume something has changed in how Tumbleweed handles the fan or something similar.
What I’m hoping is someone can help me understand why the temps would go up 5-deg C. I’m pretty sure it’s software related since nothing has changed on the hardware, nor have I changed any settings. I assume it’s a kernel, package or maybe even Plasma update that changes default behavior. My hope was that someone reading this may have inside knowledge and be able say “We changed x/y/z that may cause this. Here try these tools and see what the output is”.
Otherwise, I don’t know what to do other than live with it. Which I can. But something changed and I would like to understand it.
First off. It would be rare that folks participating in here have the SAME configuration as your system. Software, hardware, settings, and so on are highly variable. It would be a coincidence that someone can offer a direct comparison
Secondly, the default KDE System Monitor (I guess that’s what you’re referring to) is, what I consider, highly generic and doesn’t necessarily offer a lot of detail. Plus, you must watch the system for quite a while, while executing the typical applications and such.
Thirdly, the running fans could be caused by high CPU usage, high graphics usage, or occasionally excessive drive usage (or a combination of).
A while back, yes, there was an issue with the Baloo indexing system that many folks were affected by (Baloo is a file indexing and search framework for KDE Plasma) and reported on. (We’ve disabled it on all our machines).
Considering your post seems (so far) specific, it may be related to something for your system and possibly to hardware specs (or not).
What if you boot a previous kernel? Any diff?
What if you boot a previous snapshot (if using BTRFS snapshots). Any difference?
It is possible that the thermal grease aged badly or the CPU heat spreader is no longer seated properly. Also blowing air into the cooling fans of a laptop may make it worse as you push the dirt in. It seems you already had opened the case once to exchange the fans. I would open it again and check seating of the heat spreader, thermal grease and blocking of cooling funnels.
I hope I’m not just sensitive but I’m feeling a bit like a pariah now reading the responses. I’m just looking for people who maybe know more than me, not people who have my exact hardware. Maybe someone who has an idea how to even troubleshoot (I’m stabbing in the dark) My hail-mary hope was that being opensuse.org perhaps might even have a dev in here with some kind of knowledge that some aspects of power management have changed recently and could really get me on the road to either troubleshooting or submitting a bug report.
With that said, I have left btop running in a terminal to see what’s going on and there is nothing in the background making the CPU spike (and the desktop widget shows this). It’s pretty low 1-2-3% then it’ll jump to 8% for a split second, and back down. Never really spiking, pretty typical “me browsing the web” behavior. Most time on my laptop I’m just browsing the web, so I was looking to see if some tab was spiking the CPU and now I am confident there isn’t any process causing the CPU to run high. Basically for the last year since I’ve had Tumbleweed my laptop has ran 45-deg C, and now the last week or two it has been running 50-deg C, which also apparently is the temp at which the fans kick on so it’s a bit louder than usual.
I switched my power profile to ‘balanced’ instead of ‘performance’ to see if maybe that would change the temps but it didn’t. I wanted to run powertop to see if that will give me clues so unplugged the laptop and left it in ‘balanced’ mode. I did also run sensors to verify that the posted temp (in the KDE desktop widget) is the same and it shows temps at 50-deg C as well. Ironically powertop is showing a power usage of 255w, and I’m down to 90% in about 10-minutes and it’s showing 2:45 left in the battery app. Somethings up because my battery is 60-ish watt-hours so at 255w it’d drain fast. I’m just lost. I know all of the battery stuff is “approximate” but I also put a new battery in this year and typically was getting 4+ hours on a charge as recently as a week ago Saturday (running zoom no less for a meeting and my CPU does run around 20%-25% while hosting meetings). My suspicion is something changed in the power profiles, or at the kernel level, or some system I am not smart enough to even know about. And, hence me coming here. Maybe someone more knowledgeable in Tumbleweed development (or kernel dev) might know there were changes to something relevant and offer some kind of way to either troubleshoot and if necessary confirm there’s a bug and I can submit it.
I guess the next thing could be to try and change the thermal paste as mentioned - I’ve heard of that before but never really had a computer exhibit that behavior. Usually the fans get noisy after a few years and I’ll swap those out and all is good. I’ll probably give it a few weeks to “sort itself out” as Tumbleweed tends to do, and if it doesn’t then I’ll order some paste and give that a whack. I do understand 50-deg C is well within spec, but it’s apparently where the fans kick on and that sucks.
EDIT: It’s been a few more minutes and I’m now down to 76% on battery (according to btop and the battery widget in Plasm), so something eating my battery, generating heat (my lap says so) and has my CPU running 5-deg C hotter, but the CPU doesn’t appear to be loaded up. I can say a week ago Saturday my laptop ran for > 3-hours on battery hosting a zoom meeting and I still had 30-40% left when it was done. So, the change has been an update in the last 8-days.
I’m using nvtop to verify that it’s using my iGPU, but you may be on to something. I have noticed recently things aren’t working right with my dGPU. I use prime-run to start some programs and it doesn’t work any longer. It opens the app, but doesn’t use the dGPU, just the iGPU so there may be something there generating the heat. Since I rarely use it, and since my MX 250 is only marginally better than the iGPU I just ignored it. That broke recently as well, about a week ago I tried to open 0 AD and it actually locked up my system where I had to hard power down. It always ran fine starting with prime-run.
How would I disable it, is there a systemctl service I can disable? I’m fine never using it since like I said a MX 250 isn’t exactly worth writing home about.
I switched back to Wayland, and it’s running right around 45-deg C and I had zoom running and a few other apps running on different desktops (it ran at 45-deg C on x11 before). Yesterdays update did include a kernel update, so I’m not sure if it’s wayland or the update. I’ll probably have to switch back to x11 next time the drivers start flaking out in Wayland and I’ll know then I guess.
Anyway, I’ve been on battery for 20-25 minutes and still at 95%. Ironically powertop shows the “the battery reports a discharge rate of 272w” which is higher than yesterday but still doesn’t make sense to me since the battery is 60-something watt/hour and that rate would drain it in 15 minutes.
Basically guys and gals, I’m not an idiot, but I’m also not a Linux or openSUSE guru so I know a tool or two but now even how to interpret its output! So, it would be nice to get schooled by those more experienced in proper “troubleshooting” methods for this. I did find some stuff on the Arch wiki but it wasn’t really what I was experiencing (and I don’t hand-build my drivers! Thank god the openSUSE team sets that all up for me)
Hello team!
I’d like to top up this topic. I use OpenSUSE for long time, and it is really the problem for me - recently, maybe, at April/May or so, CPU fans on my 2 laptops started to work all time. I have two laptops - one pretty old msi bravo 15 a4ddr, and the second is pretty much new Dell XPS 15 9520. Both have 64Gb RAM, and fast 1TB NVME. I can provide more details. One has AMD Radeon intergrated and dedicated, another has intel and dedicated NVidia.
OK, you can told me that i need to clean up fans and so on, but i’m glad that i have enough time to check different things - and i simply installed two other distros.
I use Dell laptop for my work - it has only qemu-kvm installed, and it runs 3 VMs, so i just copied qcow2 disk images and did “virsh dump-xml/define” and that’s all. For second, MSI, i use firefox and visual studio code, and some python. So i’ve copied folders like .mozilla, .vscode and installed required python modules.
For OpenSUSE, i always hear fan sound, always, even when i do nothing - i sleep and i can hear it. Running pretty loud.
For Fedora 40, i hear it only when i use it, so if i lock screen - it is silent, all night long, any time when i do not use it, it’s silent. Same for Kubuntu.
I used these different flavours for some long time - about 3 weeks each
I do not like any other distro, as i really enjoy using OpenSUSE, my first was Emerald.
So any suggestion will be highly appreciated! I’d be happy to provide any required information, but i just do not know what to look. Baloo is disabled, top/atop/htop do not show any suspicious, lm_sensors are installed.
If someone can help to investigate - i will be happy!
P.S. I use KDE, i’ve tested both wayland and X11, it doesn’t matter at all. I do not use official Nvidia drivers, only open-source.
P.P.S. It can have a lot of different reasons, but for my specific task - running 3 identical VMs, it is extremely suspicious. Exaclty the same VMs, nothing else running - no browser, no messangers, only qemu/kvm.
For @gigzbyte it would be good to get the same info from Fedora 40. Let turbostat running for at least 1 minute before breaking it off. Please share the output (incl. the command issued) using preformatted text (ctrl-e of the</> button).
analyzing CPU 1:
driver: intel_pstate
CPUs which run at the same hardware frequency: 1
CPUs which need to have their frequency coordinated by software: 1
maximum transition latency: Cannot determine or is not supported.
hardware limits: 400 MHz - 4.90 GHz
available cpufreq governors: performance powersave
current policy: frequency should be within 400 MHz and 4.90 GHz.
The governor "powersave" may decide which speed to use
within this range.
current CPU frequency: Unable to call hardware
current CPU frequency: 4.84 GHz (asserted by call to kernel)
boost state support:
Supported: yes
Active: yes
Note that as mentioned I switched over to Wayland and did my daily zypper dup and the temps are back around 45-deg C and fan activity is normal. I haven’t gone back to X11 yet because Wayland has been good to me, even working in Zoom which traditionally the video was always choppy in Wayland which is why I went to X11 since it was a smoother experience.
The dumps you provided look normal only strange thing to me is “current CPU frequency: 4.84 GHz”. For me, running also the powersave governor I see the frequency is down to 900 MHz when idle. If you dump it multiple times, is it always ~4.8 GHz?
My turbostats are somehow in complete idle mode way lower:
Interesting, I just did another turbostat and the results are very similar. My PkgWatt seems 5 times higher than yours. That seems odd, this is a 10th gen Intel i7 (10510U) so I’m kinda jealous that when mines using 4 watts yours us using ~ 1/2 a watt.
I ran cpupower frequency-info again and current frequency was 4.86 Ghz. I am plugged in and on “performance” mode, so I switched over to “balanced” and ran it again and it was down at 800mhz. I reran turbostat but the output was generally the same.
I still haven’t gone back to x11 to see if it’s still sucking my battery down. Wayland got real flaky for me for a while, but this time I came back it’s running nice and smooth.
You switched to “balanced” but I am running always on “powersave” as I found that has no negative downsides for me.
When stressing the cpu (stress-ng --cpu 8 --iomix 2 --vm 1 --vm-bytes 1G --timeout 200s) I see the expected behavior, about 50s at ~115 Watt and after that it remains on~95 Watt until the test is over after which it falls back to < 1 Watt for idle.
If you still like to lower the idle power I would look also in the BIOS.
