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

Thank you for your fast reply!

This issue seems to be one that's specific to Universal Audio TB audio interfaces on Windows systems.
I suppose it's because their Windows drivers development isn't their main interest, their products catering mostly to Mac users.
The buffer size doesn't seem to make a difference. I'm usually running a 256 samples buffer, but I've tried the whole range from 64 to 2048.

I didn't perform any TB driver update. Thanks for the link. I'll look into it.
This is what I'm currently using:


LatencyMon also shows no issues on a 60 minute run, while still working on the laptop. The only underruns I noticed were reported by the DAW, when using demanding plugin chains with the low 256 samples buffer.

I've thought of upgrading to the XMG latest BIOS. However, I'm not sure if this could impact the warranty of the system.
Any suggestions on this?


Cheers!

 

One question:

How should be set up the clock of all the core?

Now i have all of them at 4.8ghz.. but maybe i can push the first 2 at 5.1ghz? or more?

PL1 should be at 200w? or 170w?

Wha is the best set up for the i9 11000k ?

 

You cant cool 200w. The max with my i9-11900kf is around 160-170. Than thermal throttling
Run Intel xtu and check temperatures.

 

Setting 170w i completed one run of Cinebech r23 with a score of 14917 without thermal throttling but the temperature was about 95-98c.

All the core now are at 4.8ghz, but can i set up the fisrt 2 core at 5/5.1ghz or is useless?

 

Here my tests from today with i9-11900kf and Cinebench R23

3.85Ghz @ 130w with 75°C | score 12457
4.14Ghz @ 160w with 88°C | score 13299
4.36Ghz @ 160w with 89°C -80mV | score 14040
4.42Ghz @ 170w with 93°C -80mV | score 14240
4.47Ghz @ 170w with 93°C -100mV | score 14403
4.50Ghz @ 175w with 95°C -100mV | score 14453
4.55Ghz @ 175w with 96°C -120mV | score 14577

For 4.8Ghz my system need >200w.

 

The trick with these crispy chips is to lower the temps as much as possible, which reduces the power draw due to less thermal leakage.

 

This is where seeking out a binned chip would come into play that is not too leaky and naturally pulls less at stock so you can extract maximum performance from the X170KM-G. I'd also dig under the hood and start running cooling scenarios (flat vs elevated vs bottom off for starters) and redo the thermal compound application.

I've tested several 11900k's (as noted the wide variance at stock pull and heat under load in a previous post). The worst offender pulled ~210w (which would be 215-220 in a laptop due to extra heat) at stock settings running CB23. The best pulled 165-170w at stock settings. That also turned out to be an SP95 (previously an SP94 before re-flashing the bios to update the microcode) when I plugged it into my friend's Asus board and the chip I'm using now in my desktop while I wait to pick up a competitively priced X170KM-G .

I have some Corsair 3800 sticks sitting on the shelf I want to test too along with a binned 11900k once @Prema finalizes the BIOS on the KM. In the US, it looks like HIDevolution will be the only Prema partner carrying it?

You can set it to 200w, but seeing as even setting it to 170w your system is still pulling 200-210w already and hitting 95-98c at stock, it will be an exercise in futility. 11900k's at stock auto boost to 5.1 single core naturally, but you could set some of the cores to 5.1 and re-run CB23 and see where you hit a thermal wall if XTU on the X170KM-G allows it. ABT mode on desktops will set your 11900k to 5.1 all core, 5.3 boost but that is well beyond the scope of the X170KM-G as the pull is usually 280-290w+ up to some staggering pulls of 330w+ depending on chip quality.

 

It looks like your 11900kf might not be as good of a sample as @marco89ct 's sample as he can do a 4.8 run all core (still pulling 200-210w), but you could start to check your cooling as a start. You'd be surprised what you find if you get inside there and take a look. Only once have I encountered a laptop that a re-paste didn't improve thermals (it matched them) and sometimes you find other areas you can improve.

@jc_denton is someone who takes it to the next level in extracting maximum performance from these true DTRs. Click on his links in his sig and go take a look at some of his magic he extracted from the P870TM1.

I'd imagine you would have a field day with an X170KM JC.

 

@electrosoft
I dont know what i can do better from the cooling side. Fans are clean, stock thermal paste exchanged grizzly kryonaut.

@Alex Zander
The reseller send me a kit of 2x32GB patriot ram and i send the 4x16gb back

 

It could just be the chip itself as a hot and heavy sample there's only so much you can overcome. You could go next level and do some pairing tests to make sure the fit is as tight as possible and potentially switch to nanogrease extreme which handles non optimal pairings better and see about a shim assuming those are potential issues but at its core some chips just run hotter and heavier than others. The fact your system hits a thermal wall at 4.55ghz @175w vs 4.8 @ ~205w for the other chip leans that way it is the chip it self.

With the issues running 4 slots, that's a good call swapping to 2x16.

 

This is what we call a real laptop.

 

can you send me the style note bios for this

 

This guy is a TOTAL JOKE and a NOOB , doesnt read his manual , nor does he feature anything worth .... didnt know the laptop has 3 nvme , let alone the 4th pci-e 4.0 , lol ,.... watch video and see as well mentions that the only laptops before with desktop cpus WHERE AMD RYZENS , as intel desktop cpu in the laptop is completely new.... anyone have input on this guys video to the machine ....

@Donald@zTecpc
@Mr. Fox
@win32asmguy
@Prema

 

He may be a nice guy and all that and I like for the X170KM chassis to get some exposure but:

up to 64 GB RAM - really?
missing the 4th SSD slot?
mentioning the X170KM as some kind of first laptop with intel desktop CPU's?

not impressed...

 

sadly i am testing this on windows 11 pro currently as a lot of clients are wanting it on release date of the os.. stupid i agree but hey they want to pay for it and get it loaded .....trying to find flaws , all of the apps wont work exp the sound blaster atlas works for a few boots then gone , the audio driver updates and cant get back

 

have same issues with this software

 

i got to say i am blown away being i am sure origin sponsored this video and review that NO One looked at the content or the information before a post almost if the guy is simply reading a script , at least linus does mention most the specials on the clevos and as does many others this guy has 3.2m followers that are getting COMPLETE incorrect information!!!

 

any one know were to get a extra plastic cage for the power supplys ... the plasitc case? to hold two

 

Where are you from? I have a cage and I dont need it.

And is send you the bios today, when I'm at home.

 

gota say this guy always keeps it simple and to the point and covers all options pretty well compared to many videos and reviews I have read and seen ,....

 

portage indiana shoot me a pm , thanks man

 

I did some more tests today but 4.66Ghz is limit with the cooling and my i9-11900kf.

4,65Ghz @ 180w 97°C -150mV score 14851 (thermal throttling from time to time)
4,66Ghz @ 183w 98°C -150mV score 14865 (thermal trottling nearly all the time)

Offset of -150mV is also max. -155mV and more => crash

Would be nice, if anyone else with an i9-11900k or kf can do such a test.

I dont think, @marco89ct your i9 run all the time with 200w. Your score is to bad for 4.8ghz on all cores, when you compare it to my score with "only" 4.66ghz. Did you run Intel XTU in background and thermal throttling while benchmark with cinebench r23?

 

Ah beta testing a microsoft OS.

 

ehhh , idk if best choise lol , some customers want new new new lol before new new new is fully ummm whats the words , STABLE lol , but yes so far havent seen to many issues ,out side the sound difference for that model compairing it to a p870tm1-g or a p870dm3-g ..... sound blaster atlas doesnt like to work once updates run course , and clevo has 0 support for this model currently as i see .... so idk , running stable with i9 11900k , 128gb 2933 ram ,and samsung 980 pcei-e 4.0 2tb , 3x4tb sabers ..... rtx 3080
no issues so far decent temps too during games ,

aka warzone
ac
diablo
protools
fl stduio
adobe packages
ect...

no temp issues yet

 

Well I'd call it more like less unstable as it releases

 

2 weeks since the last crash... guess i can stop holding my breath and crossing my fingers, and let NVidia and Msft update themselves.
I hhaven't flashed Bios or anything. Everything's stock Clevo, no overclocking.
I just spent the afternoon with BofA deciding my annual cell phone upgrade was unusual activity, and finding a nice bottle of scotch to soothe the pain.
Hitting "Buy" caused the blue screen, on a 45 watt charger...
Long deep sigh, I'll read y'all's activity for the past 2 weeks in the morning.

 

How dare you spend your own money

 

Here is where an 11900k would land at stock clocks with zero throttling for reference:

 

@electrosoft

Thanks for sharing! That shows what I was thinking. If his 11900k dont run in thermal throttling, his score should be nearly the same. But his score is ~500points lower and thats a lot.

 

I am continuing the tests and i got in Cinebench r23

4,8Ghz @ 200w 98°C -100mV score 15030 point (thermal throttling only once)

So i decide to keep this configuration

4,8Ghz 2 core and 4,7Ghz 6 core @ 170w 98°C -130mV score 14860 point (no thermal throttling)

I think this last is the best setting for my laptop and during half hour of battlefield 5 in 1080p ultra setting (CPU is very under stress) the temperature were stable about 85-93°C.

In any case, the best practise is to keep low the temperature with additional cooling fan under the laptop (which i don't have) and keep the room temperature at 25°C or lower.

So in the end i have this setting:

CPU
4,8Ghz 2 core and 4,7Ghz 6 core @ 170w 97°C -130mV
Cinebecnh score 14860 point (no thermal throttling)

GPU
Clock +160mhz
Memory clock + 1200mhz
Unigine Heaven Benchmark 4 i got 4587 point (min fps 35.8 max fps 402.1 resolution 1080p 8xAA fullscreen, Quality Ultra, Tessellation Extreme)

This GPU point are in line with a Nvidia 3080FE desktop version without OC as i saw on You tube.

 

What is it set for 8 cores?

As for GPU, run Timespy to get a more accurate score as the 3080 laptop version is ~3070 desktop performance.

 

3080 laptop version is ~3070 desktop performance for all the laptop with a gpu 3080 or for our Clevo X170KM with gpu 3080?

Because as far i saw with Unigine Heaven Benchmark 4 in stock condition my 3080 laptop was similar to 3070 desktop, but after all the overclocking i got +22% of performance.

Without Overclock
Unigine Heaven Benchmark 4 i got 3962 point which are similar to the result of 3070 desktop without OC

With Overclock
Unigine Heaven Benchmark 4 i got 4587 point which are similar to the result of 3080 desktop without OC

 

Yes, the ~165w 3080 in your X170KM-G has desktop equivalent performance of ~3070 at stock.
Install and run Timespy to give you a more accurate reflection of your GPU performance and check your GPU OC.

 

Wow, y'all been busy. Crossshot got my hopes up, then immediately dashed them.

NVidia update crashed the machine as it finished.
Somehow my Balanced profile got set to hibernate after three hours. ?
¯\_(ツ)_/¯
If i get another extended period without crashes, i'll sit on my hands and wait for Prema.
If not i'll try the Xmg stuff.

 

I'm in the next 8 days in vacation without computer. After this, I will continue the tests. I will also test Thermal Grizzley Conductonaut for CPU and GPU. And next week the new RAM will be delivered.

 

In order to get the maximum performans from CPU and GPU at same time (without remuving power to the CPU when the GPU is under stress) i tryed to disable the NVIDIA platform controller and framework.

It work but has the collateral effect to lose the power boost of 15W in the GPU.

Is there the opportunity to have the GPU running at 165w without NVIDIA platform controller and framework?

 

I don't think those 2 components have anything to do with power draw, they're mainly needed for for things like GeForce Experience and Shadowplay - stuff like that. I don't have either of those components installed and GPU goes to 165w nicely whenever it needs it during gaming.

Are you installing drivers using NVcleanstall? https://www.techpowerup.com/download/techpowerup-nvcleanstall/

 

This is random but I noticed your signature says" de-lidded i9-11900k" from HIDevolution, but HIDevolution don't do de-lidding for that processor (I know because I got the same CPU right at launch, thinking they would de-lid it but they came back saying they actually couldn't do it). Is your information correct? If so how did you get them to do it?

Are you using NVcleanstall? https://www.techpowerup.com/download/techpowerup-nvcleanstall/ to install drivers? IMHO the default Nvidia driver installer with GeForce Experience is often a dumpster fire crapshoot, especially for laptops and probably more so for DTRs like this laptop. Just an opinion though.

 

It was an option when i ordered, they said it was delidded, i haven't actually checked. might be a matter of how big a pita it was to do the first few, and they stopped.
i did a clean install when i rolled back to oem, i just do express normally.

 

You have to be careful of the passives on the top, you could kill it if not careful:

 

Hi, i did it and it works, but the GPU now goes till 150w maximun power usage.

Seams that disabling the Nvidia Platform Controller And Framework i lose the boost of 15w in the GPU. Am i right?

Is it possible to push the 15w GPU boost while the Nvidia Platform Controller And Framework is disabled?

 

What clocks are people seeing while gaming? JT's review of the KM-G with 11900K show it clocking very low (3.5GHz-ish) but still scoring well in gaming benchmarks, but in CPU stress testing he shows clocks >4GHz. Are people running fixed clocks here or letting it boost?

The 10900K is still available at my reseller (but who knows for how long), but it looks like the 11900K wins out in CPU bound scenarios anyway. It just doesn't seem to clock very well so the IPC gains must be good.

 

Playing to battlefield 5 at 1080p in multiplayer my i9 11900k has all clock at 4.7/4.8ghz (CPU 55-75% load costant) and temperature between 85/95 degree (undervorlt -130mv).
If i play to battlefield 5 at 4k the clock lower to 3.5ghz and temperature goes to 70 degree due to dynamic bost of nvidia which give more power to the GPU in order to reach 165w.

But during the benchmark (CPU 100% loaded) with this configuration it goes thermal throttling after 1 run of Cinebench R25.

 

4.7/4.8GHz sounds promising but that dynamic boost behavior sounds awful. Does that ever tank your fps?

 

I'm actually trying to understand if there are collateral effects in that behavior of all the laptops with Nvidia GPU.

Up to now i tested several games and i reach this conclusion:

It is possible to disable the dynamic boost (disabling the Nvidia Platform Controller And Framework) BUT the collateral effect is to lose the GPU 15w boost that in GPU bound game push the performance (practically in 4k resolution).

In Battlefield 5 at 1080p ultra in a specific part of campaign (just after loading my checkpoint) i got 183fps with and without dynamic boost.
But at 4k resolution ultra setting in the same part of campaign i got 80fps with dynamic boost and 76-78 fps without. This because the GPU lose the 15w boost.

In Red dead redemption 2 at 1080p ultra setting without dlss in the benchmark of the game i got 100fps avarage with dynamic boost and 98 fps avarage without dynamic boost BUT the min and max fps were better without dynamic boost.

In conclusion the better solution would be disable the dynamic boost BUT and somehow enabling the 15w boost in the GPU, and at moment i don't know if it is possible.

Therefore i'm asking here since yesterday if someone know about that.

Also in the GPU benchmark Unigine Heaven 4.0 i got 4587 point with dynamic boost and 4040 without. In any case seams that have the dynamic boost worth it.

 

Hi all a quick update for everyone although i can't say much yet as its on going. I have been working with PCS with this issue of 4 Sticks of Ram for the following issues.

Failing to POST
POST and Instant BSOD
POST into Windows and Shutdown on Load of GPU/CPU and RAM

The issue we have come to agree on is the BIOS as that's the only thing that has changed between BIOS updates and the fact that both sets of my sticks PASS memtest.

However we also hit an issue with 2 Sticks where the system shutdown, If your CCC has updated to 3.48 go back to 3.36 seems to solve that issue, i will be in contact with PCS tomorrow to see what's what and what's going on to verify fully where the issue is within the BIOS.

I hope this helps others.

 

This is correct. Disabling dynamic boost makes you loose that extra 15 watts to the GPU. I don't know of a way to keep that boost while having dynamic boost disabled. Perhaps you could try overclocking your GPU using MSI Afterburner to offset the performance regression.

Do the GPU core clocks drop when dynamic boost is disabled? If so, try my overclocking suggestion and see if that forces a higher power draw and higher clocks.

 

An OC'd 10core 10900K is the best chip for gaming you can have, assuming you have a graphics card that can feed it fast enough.

Graphic cards should run relatively cool in this machine, have you guys tried undervolting with afterburner to circumvent the power issues?

 

Heaven 4.0 benchmark doesn't tax the Cpu. And Dynamic boost 2.0 is a disaster meant to save all the new modern thin and slimy from overheating (unified heatasink). Lets say 200w cooling capacity intended for 45-56W Cpu and 165W GPU. 56+165 = This means the cooling cant cool 21W. Hence cripple the CPU power to be able to let the Gpu run at the adverticed 165W. Darn disgusting!

@jc_denton the nowadays tech is all made for thin and slim.

 

Correct me if I’m wrong, but for the X170

dynamic boost off: >150W CPU, 150W GPU
dynamic boost ON: 50-60W CPU?, 165W GPU when it diverts power to the GPU?

that’s a huge loss of power.

What about games that push both CPU and GPU? I see the laptop does well in Cyberpunk so the algorithm can’t be too dumb. I feel like I’m missing something here, on paper it looks absolutely terrible, but the reviews show it does really well?