I hope it works. The copper heat sinks i sent the link to are 1 mm and .5 mm silicone based pad. So need 1.5 mm room.
So I switch again on kubuntuthing for evalute difference between powertopās reporting stats.
First I notice, temperature of nvmeās immediate drop down, and powertop mark over 400 wakeup/second.
So I think powersave isnāt the real issue.
nvme-pci-e200
Adapter: PCI adapter
Composite: +35.9Ā°C (low = -0.1Ā°C, high = +82.8Ā°C)
(crit = +84.8Ā°C)
Sensor 1: +35.9Ā°C (low = -273.1Ā°C, high = +65261.8Ā°C)
Sensor 2: +39.9Ā°C (low = -273.1Ā°C, high = +65261.8Ā°C)
Without I post more screen, look all powertop tunables at default same as Arch live systemā¦ this is the current status.
Now would be a great thing find some possible solution for high temps of nvme on Arch and EnOS live and EnOS I installed and run with LTS kernel, current at 6.x release.
Someone can help me to digging and follow some ācommands protocolā for grab some system internal feature information to understand how component or module get ābrokenā on weird Dell Laptop, you are really welcome.
Does buntu have tlp
installed?
Thank you for hint, Iām appreciate, but on 'buntu tlp (Battery powersave manager) is not installed.
Thank you, Iāll try to buy today
Hopefully you can find something like it in your location.
Hello everyone partecipate and thanks for yours feedbak. @keybreak @Echoa @ricklinux @freebird54 @mrvictory
Just inform I found the problem, actually is the Intel RST controller issue.
I switch RST to AHCI on BIOS, after reboot EnOS load fine and nvme temperature is normal.
$ inxi -Daz
Drives:
Local Storage: total: 1.38 TiB used: 193.7 GiB (13.8%)
SMART Message: Unable to run smartctl. Root privileges required.
ID-1: /dev/nvme0n1 maj-min: 259:0 vendor: Crucial model: CT1000P5PSSD8
size: 931.51 GiB block-size: physical: 512 B logical: 512 B speed: 63.2 Gb/s
lanes: 4 type: SSD serial: rev: P7CR403 temp: 38.9 C scheme: GPT
ID-2: /dev/nvme1n1 maj-min: 259:3 vendor: SK Hynix model: BC711 NVMe 512GB
size: 476.94 GiB block-size: physical: 512 B logical: 512 B speed: 31.6 Gb/s
lanes: 4 type: SSD serial: rev: 41002131 temp: 36.9 C scheme: GPT
Soo this is cool, because Iām previous installed windoze 11 with Intel RST support, and Iām pretty sure win11 is āunableā to bott in AHCI swapped setting, lol I need to reinstall win 11 as workaround.
So I can learn about Intel RST driver for Linux? I can open a bug report on whatever developer work on this? Iām not an expert but I want help to report my experience maybe useful for others guys with similar Hardware based on 11th gen Intel core processor and RST / RAID controller performance.
Not really sure for next step nowā¦
Windoze requires reinstall to boot into AHCI, i believe
That slipped my mindā¦When given choice in BIOS - you should always choose AHCI, i even wonder why it wasnāt.
Iām surprised you could install and run it with RST enabled. I always recommend that be turnoff off for installing and use AHCI.
Boot into Windows
Open a shell as administrator
Run bcdedit /set safeboot minimal
Reboot, enter BIOS and switch to AHCI
Reboot into Windows again
Open another shell as administrator
Run bcdedit /deletevalue safeboot
Reboot
Edit: Hopefully this works properly.
Well guys, wait a little, from performance point of view, RST controller works 4x faster than AHCI I/O operation.
I really donāt want AHCI workaround, files transfer from SSD (thunderbolt port) is dropped from 4 Gbits/s to 1
I need to figure it out exactly how RST work on Linux, and understand if it possible tweak for cooling down nvmeās chips.
And need to check if 'buntuthingh load RST driver or notā¦ many more check to do, argh!
EDIT: This is my peak writing file AHCI 1 Gbit file transfer speedā¦
$ sudo inxi -Daz
Drives:
Local Storage: total: 1.85 TiB used: 1.15 TiB (62.0%)
ID-1: /dev/nvme0n1 maj-min: 259:0 vendor: Crucial model: CT1000P5PSSD8
size: 931.51 GiB block-size: physical: 512 B logical: 512 B speed: 63.2 Gb/s
lanes: 4 type: SSD serial: rev: P7CR403 temp: 53.9 C scheme: GPT
SMART: yes health: PASSED on: 12 hrs cycles: 46
read-units: 84,806 [43.4 GB] written-units: 1,410,727 [722 GB]
ID-2: /dev/nvme1n1 maj-min: 259:3 vendor: SK Hynix model: BC711 NVMe 512GB
size: 476.94 GiB block-size: physical: 512 B logical: 512 B speed: 31.6 Gb/s
lanes: 4 type: SSD serial: rev: 41002131 temp: 41.9 C scheme: GPT
SMART: yes health: PASSED on: 268d 15h cycles: 495
read-units: 11,072,275 [5.66 TB] written-units: 13,729,134 [7.02 TB]
ID-3: /dev/sda maj-min: 8:0 type: USB vendor: Crucial
model: CT525MX300SSD1 family: Micron Client SSDs size: 489.05 GiB
block-size: physical: 512 B logical: 512 B sata: 3.2 speed: 6.0 Gb/s
type: SSD serial: rev: M0CR070 temp: 40 C scheme: MBR
SMART: yes state: enabled health: PASSED on: 351d 10h cycles: 2338
written: 12.8 TiB
I need to compare with RST
Kernels ālinuxā and ālinux-rt-ltsā 5.15 both, look no differencesā¦
EDIT 2: Copy files (Dolphin) between ID-1 and ID-2, increase temp. to 58 CĀ° at 2 Gbit/s speed, it chek by bashtop
-AHCI Mode-
My goal is enable RST and keepin low temperature, how make this more clear is a discovery, yeah.
Iām keep note of all temps today
Tomorrow check others thing, I need to get a look on RST kernel.org sources
EDIT: Iām on reading this one https://www.intel.com/content/dam/www/public/us/en/documents/white-papers/rst-linux-paper.pdf
Thanks, actually useless, Iām go to find way to manage RST driver calls.(If any is available)
I donāt know for example:
If RST is open source module driver, builtin kernels, or if possible compile user binaries with OPTIONS
(Iāll check this tomorrow)
Only few documents related topic for RST host controller is online, indeed not usually for user-side feature is publicā¦ so we need to make once more, ahahaha ā¦ Iām close for get out of time nowā¦ see you tomorrow.
This is my āAHCI modeā lsmod for tomorrow reminder: https://0x0.st/Hsla.txt
next I get RST lsmod
We have had the following stuff for RST:
- nvme_load=yes kernel parameter
- vmd in MODULES in /etc/mkinitcpio.conf
Can you check if you have them?
(And Iām not at the linux box right now but writing from my headā¦ )
Its likely if its a newer system dell no longer uses that same interface then for controlling it or not use it on consumer hardware anymore I canāt be sure. Its was used for a long time and bound to see replacement eventually
I wonder if youāre running into similar issues to some of dells gaming line like the G5 where fans act weird in Linux. Thereās a number of modern dell systems that expect Windows user space components for fans to operate in a sane way
Hello to everyone, today I start by swap AHCI/RST mode and boot,EnOS (kernel linux)
So One thing over all is nvidia drm booting problem, it not load and unable to start sddm after systemd graphical target reached, just login on TTY and dump lsmod. My first sep is done.
This is āRST modeā https://0x0.st/HsG1.txt
If we get a look on compare, we seen ādifferencesā
RST lsmod
load vmd
AHCI lsmod
load jbd2 and nvidia modules.
I previous installed from AHCI mode by nvidia-inst --prime
and optimus-manager
is looking fine.
If try same, days ago, on RST mode, optimus-manager fail to setting mode switch, envycontrol
just work only for hybrid and integrated modes, nvidia not running after reboot and reach graphical target. (on RST mode)
Next for me is understanding why nvidia and vmd among conflict. (Probably vmd is RST module driver for RST?)
Here my journalctl -b0
for last RST mode booting: https://0x0.st/HsGj.txt
And is soo very weird thing cause conflict on load nvidia drm.
(here I start to understanding some bad things for this laptop hardware and Linux kernel).
Interactions with Linux modern (last two years) laptop hardware isnāt easy will Iām in mind!
Now Iām under reading yourās kindly reply guys. Sorry for me is very much overload and Iām slow to reply to all, but I reply for sure when relevant discover and hint is found.
See you soon after I read all your reply, thanks.