*** Official Clevo X170KM-G/Sager NP9672M Owner's Lounge ***

Hi all, just thought i post my fire strike score on a 11600k running at 4.9 all core 91C

 

You're images are not showing normally best to have a direct link to the 3dmark result.

 

thanks was wondering why it was showing up

 

https://www.3dmark.com/3dm/68351726?

 

Nice scores bud. Even nice temps too for the GPU!

 

I would like to get 5 Ghz but its just not having it, thermal shut down kicks in boo!!

 

Maybe try undervolting the CPU might help with temps.

 

I already had an undervolt running -50, i think i just need a little more tuning on voltage and maybe a repaste on the cpu but thats for another day. Then i'll start on the GPU side on thing as i want to get into the top 100 of the 3080 moblie gaming laptop with 8 core cpus as i think i'm only 800 point away

 

Check that your VRM is not the one suffering and that all the pads are making good contact.

 

thanks ill check them when i do a repaste

 

Intel didn't offer the KB chips from 11th gen in laptops. Only for small form-factor desktop designs as etc their own NUC
Intel's Enthusiast-Grade Core i9-11900KB Tiger Lake CPU

 

Yeah but I suspect intel treated the 11th gen maybe a bit differently as more of a stop gap.

 

11th gen BGA was on 10nm to try compete with AMD for notebooks (where the $$$ is). 11th gen LGA got nothing.

 

Yeah but only is Alder lake really properly scaling on that process now which gives them more options.

 

Yes, 11900KB, the past generation was not offered for laptops, but the current one will be as per the "Alder Lake-S BGA" roadmap confirmed in Laptop. Probably to compete with AMD as seems they plan to launch a 65W 16c/32t for laptop. It can have a different name tho, but certain 8P+8E configuration AKA 'BGA1964 Alder Lake -S series'

 

Update to my Linux adventure, it would seem right now its a tad be toooooo early for the X170KM-G, i mean when you have to use switches in kernel to get most things working then you still have issues with SDCard Reader and TB4, not only that lack of software for keyboard, fan, led's makes it kinda pointless.

 

Isn't that standard Linux?

 

Yes it is Standard linux, however i have it now booting Fedora fine, just with the following for booting first acpi_osi command tells the system to forget everything which is where the booting issue is and the second one forces it to run the Linux setup in the DSDT.

Code:

acpi_osi=! acpi_osi=Linux

The only issues really left is SD-Card reader which i think blocking a driver from loading and then using realteks driver should solve, then the thunderbolt 4 DMAR issue when its enabled in BIOS which slows down booting.

Sound has been fixed with Tuexdo's realtek pin fix, just got ITE and Keyboard to compile into Fedora as although OpenSuse uses .rpm files its not compatible with some dependencies which isn't hard work.

Everything else works fine though. Once i have gotten Fedora working 100% i will then move over to debian as i am so used to using

Code:

apt

over using

Code:

dnf

Edit: Gone over to Ubuntu 21.10 Fedora was bugging me to much with the CLI differences to much always been a Debian CLI user.

Edit: So I have LED working Hotkeys working and Fan control working still working in TB4 and SDCard.

 

I lost the link for the windows 11 drivers, or oem CD
Can someone please link it

 

https://download.schenker-tech.de/packages/win11/xmg-ultra-17_e21_driver-package_v1.08.01_win11.zip

 

Thank You

 

Right new Update about Linux, I decided to use Elementary OS 6 Odin and in the process of making a script to set it up fully.

Thunderbolt 4 is still an issue however sound is working on stock kernel as is Fan/CPU control also the LED is working.

Once I have it all figured out and tested what I can and fixed what I can will release it, although most of the work was thanks to Tuxedo Computers, so will also be asking permission from them before posting the script.

Man I have missed Linux. Going to start daily driving it soon and if I can live without Windows fully its bye bye Microsoft.

UPDATE: SDCard is working now, so only Fingerprint reader and thunderbolt4 issues left.

Sent from my SNE-LX1 using Tapatalk

 

Enabling OC in the bios causes my computer to automatically shutdown 40-80 seconds into a cinebench stress test (without actually adjusting any settings in the OC, all as default), temps don't seem super high so not sure what is going on. Anyone else have this problem? If I disable OC in the bios it runs the stress test fine, no unexpected shutdown.

i9-11900k
Bios: 1.07.07
EC: 1.07.03

 

What is your ram

 

When you are ready, I would like to try that script

 

128 GB (4x32 sticks) (running at 2933 MHz)

 

Are you running in performance mode or entertanment ?

 

Performance (although it keeps switching on me to entertainment after it powers off by itself).

 

And no under volt?

 

Everything is set to default.

 

Thats strange are you running 3.30 for control center ? My issue has become dorment since i locked all cores at 4.5 and -60 undervolted. I would play around with clocks maybe dropping it down slightly. Could be a compatibilty issue are you running on windows 11 ?

 

lets just be honest the CLevo's bios is really bad. I would expect this kind of bios in a lower end notebook but the good news is Prema is on the way

 

I am using 3.55 control center.

I can try it, Which settings did you change from default exactly so I know what to change to see if it those same settings increases my stability?

 

It may be a ram issue

 

Where did you find 3.55

 

Clevo's website: https://www.clevo.com.tw/clevo_down.asp?lang=en

 

How soon

 

Try rolling back control center to 3.30

 

I would agree

 

Not soon eneough lol !!!

 

I don't think anyone has an ETA for it.

 

Good things come to those who wait

 

Yes, I am running windows 11. When it shuts down sometimes its giving a BSOD CLOCK_WATCHDOG_TIMEOUT, but also sometimes no BSOD appears on the screen (although it does freeze for a few seconds before powering off just like when BSOD happens, so not sure if they are one in the same just a different delay in displaying of the BSOD error). I did notice that most of the time there is an unexpected power off during cinebench, the overclock and performance settings in CC are reset back to default.

I've tried with 2x32gb sticks, both behind keyboard or behind panel cover, same issue in both configurations, although behind the keyboard 2 sticks clocks in at like 2667hz instead of 3200hz like the 2 sticks in the back cover configuration.

Ram is 22-22-22-52 1.2v modules (2x teamwin / 2x ripjaw) (Tried both pairs and same thing)

I tried CC 3.30, 3.35 (OC Update), and 3.55, all do the same.

I have 2 CPUs, a i9-11900k Delidded with pure copper IHS (temps stay very low on this, never above 78-80* no matter the settings, although XTU shows Motherboard VR Throttling on the delidded one a lot). On the unmodified i9-11900k temps are much higher (a lot more throttling) but same thing, BSOD during cinebench test. The delidded cpu is about .7mm higher then the factory one due to the copper IHS, so not sure if that makes any difference, or would cause Motherboard VR Throttling somehow. I do have Fujipoly Extreme Thermal Pads on various parts the heatsync.

I also tried downgrading bios from 1.07.07 to 1.07.06 and no benefits.

I will try downgrading back to Windows 10 to see if it solves the issue, will update once I test it.

If that doesn't work I can't really think of anything else to try other than maybe completely different ram sticks? I guess I'll stick to no OC mode for now if Windows 10 downgrade doesn't solve it.

 

Who is your reseller ? In my opinion using Windows 11 your just asking for trouble at this stage. 10 is so much better from all the work they did. I would suggest only using from your resellers suggested
Undervolting and down clocking is your friend in getting better temps

 

Did you reset BIOS to default after BIOS update? You need to reset it also to default, when you turn off OC. I have no uncontrolled shutdown with OC on with 1.07.07. With all other BIOS I had the same issue.

 

Software overclocks will reset on a watchdog failure, that's how they try and get round overclocks that are unstable so you are not trapped in a loop.

 

Just an update, I downgraded to Windows 10 and it's a lot better. I still get uncontrolled poweroff if I undervolt more then -80mv and run cinebench, but I'm able to get 200w with -80mv and steady at 4800 mhz with minimal throttling on my stock i9-11900k. (Cinebench R23 Score 14,970) - Side note, in the Clevo OC panel, If I click Save and undervolt is -150mv or more it just immediately powers off within 2-3 seconds of clicking save without even running a stress test, so I think my hardware just has less threshold for undervolting then others.

On the delidded i9-11900k + copper IHS one I have a different problem, CPU stays very cold but because the CPU is 0.7mm higher due to the beefier copper IHS, the laptop's one piece heatsync doesn't sit right on the rest of the components and I get Motherboard VR Throttling until poweroff (I assume the heatsync isn't able to cool the Voltage Regulator on the motherboard when it's elevated 0.7mm at the CPU).

Thanks for the help everyone, seems Windows 11 just isn't ready like you said.

 

With a 0.7mm increase in HSF height you more then like have no contact between the Heatsink and the CPU VRM's, so you will have two options get a thicker thermal pad for the CPU Block which will then offset the GPU as they are both connected or lap the 0.7mm off the intel hsf. This is why your constantly getting the VR Throttle.

Its not the case that Heatsink/Vapour chamber can't cool the VRM for the CPU its the case its not making decent contact between the CPU block and the VRM's

@Meaker@Sager @Mr. Fox My logical is right?

Due to the z-height + offset its going to be a pad contact issue for the reported VRM Throttling right.

 

This is most likely the issue especially if you can see the VRMs throttling in HWinfo.

You could get some thicker pads and do a progressive compression test or just get some aux thermal strips to lay on top of your pre-existing ones. Or you could optimize and replace them flat out with thicker fujipoly for proper contact and optimal heat transfer.

 

Loving the sig update to 9900ks and 3070.