Greetings, Lately (the past week or two?), my laptop has been running much hotter than usual, as experienced by my lap and reported by conky and sensors, even after a reboot and before I start X11, and when just idling. Previously, it had been running around 35 to 40 degrees Celsius. Now, it's more like 55 to 60. If I shut the lid and walk away for 10 or 15 minutes, it cools down. But when I open the lid and don't do anything else, the temperature rises again in a few minutes. This happens even when top reports 100% idle, and uptime shows the load average at 0.1±. After multiple lockups attributed to the nvidia 550 drivers a few months ago, I use nvidia-535xx-dkms. I run pacman -Syu daily. Sometimes, problems like this go away with the next kernel or driver(s) update, but that hasn't happened. I would include logs and version numbers, but I'm not sure where to start. :-) I checked the BBS, but I didn't immediately't see anything related. Any ideas? Thanks, Dan
On 7/26/24 8:38 AM, 2QdxY4RzWzUUiLuE@potatochowder.com wrote:
I would include logs and version numbers, but I'm not sure where to start. :-)
I checked the BBS, but I didn't immediately't see anything related.
Any ideas?
A guess, First - what model laptop/CPU? I ask because if I recall correctly, one of the changes the 6.10 kernel brought it was "improved" scaling of CPU core speed. This would have been from the release notes either from The Register or Phronix. My only thought is you may have a CPU/chipset, or or passing kernel parameters on boot that isn't playing nice with the new CPU scaling preventing your core processor speed from idling down. (top would still show 100% idle -- but your CPU may be idling a 3.5GHz) You can check if your cores are idling down with: grep '^processor\|^cpu MHz' /proc/cpuinfo e.g. $ grep '^processor\|^cpu MHz' /proc/cpuinfo processor : 0 cpu MHz : 797.850 processor : 1 cpu MHz : 798.104 processor : 2 cpu MHz : 797.105 processor : 3 cpu MHz : 798.114 If your cores are running at full-speed, this may be an area worth further investigating. -- David C. Rankin, J.D.,P.E.
On 2024-07-26 at 17:59:15 -0500, "David C. Rankin" <drankinatty@gmail.com> wrote:
On 7/26/24 8:38 AM, 2QdxY4RzWzUUiLuE@potatochowder.com wrote:
I would include logs and version numbers, but I'm not sure where to start. :-)
I checked the BBS, but I didn't immediately't see anything related.
Any ideas?
A guess,
First - what model laptop/CPU?
It's a Sager/Clevo NP8852P. $ grep Intel /proc/cpuinfo|sed 2q vendor_id : GenuineIntel model name : 12th Gen Intel(R) Core(TM) i7-12700H
I ask because if I recall correctly, one of the changes the 6.10 kernel brought it was "improved" scaling of CPU core speed. This would have been from the release notes either from The Register or Phronix. My only thought is you may have a CPU/chipset, or or passing kernel parameters on boot that isn't playing nice with the new CPU scaling preventing your core processor speed from idling down.
(top would still show 100% idle -- but your CPU may be idling a 3.5GHz)
You can check if your cores are idling down with:
grep '^processor\|^cpu MHz' /proc/cpuinfo
processor : 0 cpu MHz : 400.000 processor : 1 cpu MHz : 400.000 processor : 2 cpu MHz : 400.000 processor : 3 cpu MHz : 400.000 processor : 4 cpu MHz : 851.190 processor : 5 cpu MHz : 400.000 processor : 6 cpu MHz : 400.000 processor : 7 cpu MHz : 400.000 processor : 8 cpu MHz : 1083.960 processor : 9 cpu MHz : 400.000 processor : 10 cpu MHz : 1006.537 processor : 11 cpu MHz : 400.000 processor : 12 cpu MHz : 942.949 processor : 13 cpu MHz : 968.000 processor : 14 cpu MHz : 1361.782 processor : 15 cpu MHz : 1400.012 processor : 16 cpu MHz : 400.000 processor : 17 cpu MHz : 400.000 processor : 18 cpu MHz : 400.000 processor : 19 cpu MHz : 1384.733 And I just re-opened the lid to check for activity on my post.
If your cores are running at full-speed, this may be an area worth further investigating.
Bingo. My issue isn't resolved, but I at least I have something to go on. Thanks, David. Regards, Dan
Did you check which service were using how much resource? 27 Tem 2024 Cmt 07:10 tarihinde <2QdxY4RzWzUUiLuE@potatochowder.com> şunu yazdı:
On 2024-07-26 at 17:59:15 -0500, "David C. Rankin" <drankinatty@gmail.com> wrote:
On 7/26/24 8:38 AM, 2QdxY4RzWzUUiLuE@potatochowder.com wrote:
I would include logs and version numbers, but I'm not sure where to start. :-)
I checked the BBS, but I didn't immediately't see anything related.
Any ideas?
A guess,
First - what model laptop/CPU?
It's a Sager/Clevo NP8852P.
$ grep Intel /proc/cpuinfo|sed 2q vendor_id : GenuineIntel model name : 12th Gen Intel(R) Core(TM) i7-12700H
I ask because if I recall correctly, one of the changes the 6.10 kernel brought it was "improved" scaling of CPU core speed. This would have been from the release notes either from The Register or Phronix. My only thought is you may have a CPU/chipset, or or passing kernel parameters on boot that isn't playing nice with the new CPU scaling preventing your core processor speed from idling down.
(top would still show 100% idle -- but your CPU may be idling a 3.5GHz)
You can check if your cores are idling down with:
grep '^processor\|^cpu MHz' /proc/cpuinfo
processor : 0 cpu MHz : 400.000 processor : 1 cpu MHz : 400.000 processor : 2 cpu MHz : 400.000 processor : 3 cpu MHz : 400.000 processor : 4 cpu MHz : 851.190 processor : 5 cpu MHz : 400.000 processor : 6 cpu MHz : 400.000 processor : 7 cpu MHz : 400.000 processor : 8 cpu MHz : 1083.960 processor : 9 cpu MHz : 400.000 processor : 10 cpu MHz : 1006.537 processor : 11 cpu MHz : 400.000 processor : 12 cpu MHz : 942.949 processor : 13 cpu MHz : 968.000 processor : 14 cpu MHz : 1361.782 processor : 15 cpu MHz : 1400.012 processor : 16 cpu MHz : 400.000 processor : 17 cpu MHz : 400.000 processor : 18 cpu MHz : 400.000 processor : 19 cpu MHz : 1384.733
And I just re-opened the lid to check for activity on my post.
If your cores are running at full-speed, this may be an area worth further investigating.
Bingo.
My issue isn't resolved, but I at least I have something to go on.
Thanks, David.
Regards, Dan
On Sat, 2024-07-27 at 08:41 +0300, İsmail Arılık wrote:
Did you check which service were using how much resource?
My understanding is, that the machine is idle, quasi no resources are used at all. FWIW I'm on an tower PC with an Intel Model 6.191.5 "13th Gen Intel Core i3-13100". Nowadays you can forget all that "sudo tee /sys/devices/system/cpu/cpu*/cpufreq/scaling_governor"-foo that did the job for AMD and Intel machines a few years ago. Top might show %CPU 0.x %MEM 0.y at top of the list, but the fan{,s} runs berserk. On all my old PCs power management worked perfectly, but on my current PC it's broken and it always was broken. Most of the times it seems to be ok, but not that seldom something is fishy.
On 2024-07-27 at 08:56:15 +0200, Ralf Mardorf <ralf-mardorf@riseup.net> wrote:
On Sat, 2024-07-27 at 08:41 +0300, İsmail Arılık wrote:
Did you check which service were using how much resource?
My understanding is, that the machine is idle, quasi no resources are used at all. FWIW I'm on an tower PC with an Intel Model 6.191.5 "13th Gen Intel Core i3-13100".
Yes, the machine is idle. Top and conky both report little or no CPU activity.
On all my old PCs power management worked perfectly, but on my current PC it's broken and it always was broken. Most of the times it seems to be ok, but not that seldom something is fishy.
I think I once read somewhere that about 1/3 of everything in a laptop or a cellphone is part of power management. No wonder something seems fishy all the time. :-) Thanks, Dan -- My Hovercraft is Full of Eels, https://www.omniglot.com/language/phrases/hovercraft.htm. Dan Sommers, http://www.tombstonezero.net/dan.
On 7/27/24 12:41 AM, İsmail Arılık wrote:
grep '^processor\|^cpu MHz' /proc/cpuinfo
processor : 0 cpu MHz : 400.000 processor : 1 cpu MHz : 400.000 processor : 2 cpu MHz : 400.000 processor : 3 cpu MHz : 400.000 processor : 4 cpu MHz : 851.190 processor : 5 cpu MHz : 400.000 processor : 6 cpu MHz : 400.000 processor : 7 cpu MHz : 400.000 processor : 8 cpu MHz : 1083.960 processor : 9 cpu MHz : 400.000 processor : 10 cpu MHz : 1006.537 processor : 11 cpu MHz : 400.000 processor : 12 cpu MHz : 942.949 processor : 13 cpu MHz : 968.000 processor : 14 cpu MHz : 1361.782 processor : 15 cpu MHz : 1400.012 processor : 16 cpu MHz : 400.000 processor : 17 cpu MHz : 400.000 processor : 18 cpu MHz : 400.000 processor : 19 cpu MHz : 1384.733
And I just re-opened the lid to check for activity on my post.
Honestly, those do not look bad. You have 12 cores at dead-idle (400MHz) and 8 cores at 1400MHz or less. I'm no 12th Gen Intel(R) Core(TM) expert, but that seems reasonable for just opening up the lid and the various processes coming back to life. A 24-core laptop -- what's the world coming to :) Good luck with your issue. Maybe others have more than a guess -- but that's all I've got for you. -- David C. Rankin, J.D.,P.E.
Hi, David wrote:
You have 12 cores at dead-idle (400MHz) and 8 cores at 1400MHz or less.
A slightly longer pipeline will summarise a bit. awk '/^cpu MHz/ {print int($4 / 10) * 10}' /proc/cpuinfo | sort -V | uniq -c Running something similar on the OP's output gives an example: 12 400 1 850 1 940 1 960 1 1000 1 1080 1 1360 1 1380 1 1400
Maybe others have more than a guess
Perhaps https://wiki.archlinux.org/title/Powertop would be useful. Though it takes a bit of study to understand the tabs of readings, and can lead to searching for answers elsewhere for what's shown rather than giving an obvious answer. -- Cheers, Ralph.
On 2024-07-27 at 10:14:17 +0100, Ralph Corderoy <ralph@inputplus.co.uk> wrote:
Perhaps https://wiki.archlinux.org/title/Powertop would be useful. Though it takes a bit of study to understand the tabs of readings, and can lead to searching for answers elsewhere for what's shown rather than giving an obvious answer.
On a previous laptop, I ran tlp. But that was only to get power to the USB ports in suspend mode (so that I could charge my phone overnight). I'll dig into Powertop. Thanks, Dan
On 2024-07-27 at 10:14:17 +0100, Ralph Corderoy <ralph@inputplus.co.uk> wrote:
Perhaps https://wiki.archlinux.org/title/Powertop would be useful. Though it takes a bit of study to understand the tabs of readings, and can lead to searching for answers elsewhere for what's shown rather than giving an obvious answer.
Great tip. Powertop showed a number (eight, to be precise) of recommendations. For a while, I ran through a reboot, apply one or more changes, wait, repeat cycle with no luck. No, I didn't repeat that cycle 255 times. :-) But powertop --auto-tune worked wonders, and very quickly. This morning, there was another kernel update, to 6.10.2-arch1-1, and now my laptop is back to its older, cooler ways. Powertop shows that all tunable settings are "Good" (although there are still a bunch of Untunable Software Issues). Problem resolved, but not really solved. Thanks David, Ralf, and Ralph for your insights and your assistance. Regards, Dan
Hi Dan,
But powertop --auto-tune worked wonders, and very quickly.
Others should be careful as this can stop things working, e.g. the USB mouse and keyboard can auto-suspend and not resume. A re-plug fixes. -- Cheers, Ralph.
On Tue, 2024-07-30 at 11:49 +0100, Ralph Corderoy wrote:
But powertop --auto-tune worked wonders, and very quickly.
Others should be careful as this can stop things working, e.g. the USB mouse and keyboard can auto-suspend and not resume. A re-plug fixes.
Off-topic Hi, FWIW in my limited experiences all the expensive and elCheapo USB keyboards I was and/or I'm still using are working excellent with an USB-A female to PS/2 male adapter. My expensive dishwasher-safe keyboard was even delivered with an adapter included. Regards, Ralf
On 7/30/24 9:52 AM, Ralf Mardorf wrote:
My expensive dishwasher-safe keyboard
Note to all -- check before ever attempting and never use in a high-heat dishwasher or with caustic dishwashing powder (which is increasingly hard to avoid since boxed powder has gone away in favor of pod-type soaps) Cheap PCB, ribbon-cables and non-stainless hardware in non-dishwasher-safe keyboards will NOT survive the ordeal. Words from experience :) If your keyboard doesn't come with a key-puller -- don't even think about it! -- David C. Rankin, J.D.,P.E.
On 2024-07-27 at 02:53:14 -0500, "David C. Rankin" <drankinatty@gmail.com> wrote:
On 7/27/24 12:41 AM, İsmail Arılık wrote:
grep '^processor\|^cpu MHz' /proc/cpuinfo
processor : 0 cpu MHz : 400.000 processor : 1 cpu MHz : 400.000 processor : 2 cpu MHz : 400.000 processor : 3 cpu MHz : 400.000 processor : 4 cpu MHz : 851.190 processor : 5 cpu MHz : 400.000 processor : 6 cpu MHz : 400.000 processor : 7 cpu MHz : 400.000 processor : 8 cpu MHz : 1083.960 processor : 9 cpu MHz : 400.000 processor : 10 cpu MHz : 1006.537 processor : 11 cpu MHz : 400.000 processor : 12 cpu MHz : 942.949 processor : 13 cpu MHz : 968.000 processor : 14 cpu MHz : 1361.782 processor : 15 cpu MHz : 1400.012 processor : 16 cpu MHz : 400.000 processor : 17 cpu MHz : 400.000 processor : 18 cpu MHz : 400.000 processor : 19 cpu MHz : 1384.733
And I just re-opened the lid to check for activity on my post.
Honestly, those do not look bad. You have 12 cores at dead-idle (400MHz) and 8 cores at 1400MHz or less. I'm no 12th Gen Intel(R) Core(TM) expert, but that seems reasonable for just opening up the lid and the various processes coming back to life.
Okay, maybe not too bad.
A 24-core laptop -- what's the world coming to :)
No, just 20, but I agree with your sentiment. :-) 6 performance cores, each with hyperthreading, so 2 cores each. 8 efficiency cores, no hyperthreading, for a total of 20. I try not to replace my computer too often, but the pattern has always been to buy something bigger than I really need.
Good luck with your issue. Maybe others have more than a guess -- but that's all I've got for you.
Thanks. Regards, Dan
participants (5)
-
2QdxY4RzWzUUiLuE＠potatochowder.com
-
David C. Rankin
-
İsmail Arılık
-
Ralf Mardorf
-
Ralph Corderoy