*** Official Clevo N950TP6 / N957TP6 / N950TD / N957TD Owner's Lounge ***

The uncore deals with other stuff too.

 

I know, i know, but even when running prime95 running, it only gets up to 2Watts of power consumption.
I have also tested with "Nvidia Optimus GPU State Viewer", which Shows me instantly when the GPU turns on. Sometimes it does it simply by Opening the browser or something. Only for a short Moment, On, off, then on again and finally off. so 3-4x it "flickers".
And as soon as the GPU was active, the Uncore is consuming 10.8Watts for 10 seconds, which ist A LOT compared to the idle "whole package consumption" of 1-2Watts Idling.
I only want to know if it's normal to have a extremely high uncore consumption as soon as the GPU is active, but idling or if my CPU (i5-8600) is defective??

EDIT: The Short turning on of the GPU seems to be a widespread Problem with Optimus. I could calm it down a litte bit by re installing the Drivers.
But i still dont know if the high uncore Power consumption is normal?


Here an overview:
Idling CPU, no load, GPU off: 0.6Watt
Full Load CPU, GPU off: 2Watts
Idling CPU, no Load, GPU on, no Load: 10.8Watts
Normal CPU Load, GPU Full Load (Gaming): 3.5Watts

EDIT 2:
When i click on "Check for Updates" in Windows 10 Settings my dGPU gets triggered, i can reproduce it everytime….

 

I would question if the reading is accurate at that point and measure the consumption at the wall. Be sure to exclude things like your browser from using the dGPU in the nvidia control panel.

 

I don't have a wall plug meter, sadly :/

But if i look on my "BatteryBar"-Tool which is measuring the current from the battery on battery power, it seems accurate. 15-16Watt without dGPU, 27Watt with idling dGPU.
But even with wall plug meter, i can never know if the dGPU itself is consuming those 10Watts?


And yes, i have checked the nvidia Control Panel, if have set "Global" to iGPU and even checked every Programm like my browser and told explicetly to use the iGPU.
I even used the Windows-Tool in the Display-Settings where you can also assign apps to GPUs and added most of System processes.

But now it's ok for the Moment Only Windows Update is triggering it...

 

Oh the idle dgpu will eat power yes, I thought you meant the uncore on the CPU. Gddr5 buses are hungry.

 

Ok, yes i can only measure "complete System power consumption" at the battery with my tool or the values reported by the CPU, i don't really have other possibillities to invest the true CPU-Power.

Thanks for the Help anyway!

Would only be nice to compare with other users of this Laptop...

Your Notebook has no Optimus, Right? So your dGPUs are Always active? What is your uncore power consumption?

 

Never bothered to measure to be honest since my igp is not even powered by the motherboard.

 

Would anyone here with this laptop mind providing a dump of their vbios for me to download?

 

That might be tricky since it's in the system bios.

 

got my laptop repaired and replaced motherboard ,seems to really be more powerfull than the 8750h with again 1060 ,the es cpu is cooler while 8750h is a hot hell on gaming ...

 

The es CPU?

 

QN8H chip costed me same as 8400 cpu while being an i7

 

You just have to be a bit careful of what errata may not be fixed.

 

QN8H is a pre release version not a zero stepping therefore the error will be minor to none as QNLV,QNLM and QNLG, would not reccomend a cpu with more power draw than the draw of factory at any reason cause dangers may present ,yet qn8h draw is similliar to 8400 , good of the factory is the fact that they did not remove microcodes of these cpu on bios otherwise this would not have been possible

 

Sounds like you did your research on the stepping codes They do tend to use high quality wafers for those last ones.

 

Has anyone tried installing hackintosh yet?

 

Not that I've seen on this one, the first step is to research the major chips in use.

 

do you mind telling how are the temperatures?

Also i am using Rlecviewer as my fan controller. It's really going well i can recommend you guys.

 

Hello Forum,

i recently purchased a N957TP6 with 144hz display and got it to work with an i5 8400. Since i have 2666 ram ( https://geizhals.de/crucial-ballist...bls2k8g4s26bfsd-a1592501.html?hloc=at&hloc=de) and it only got detected as 2400 CL16 i was looking for a bios update.

So i went for the newest apex 15 bios, i used Windows Method and flashed the EC first and restarted, everything fine. But i somehow didnt use the Wmeset before using the afuwinx64.exe via the Commandline (used N95TP68M.09 as parameter). Now i have the XMG logo an its showing it flashed the right bios but initialising is way slower than before. I can still boot Windows but my sound doesnt work, my keyboard seems being emulated (takes about 30 secs to react to presses) and my wlan was messed up (i fixed this by using the repair option in intel setup). I can even play demanding games and if i use usb Keyboard/Mouse its all fine.

After i found out i forgot the meset, i tried it out and it gives me black screen even after 10 minutes the Keyboard went from rgb to blue (fan was full on so it kinda works). I tried this over windows and with the apex shell for efi flash ( https://www.mysn.de/driver/5_Sonstiges/Tools/Shell_for_EFI_Flash.zip). I also tried disabling UEFI and removing the CMOS battery.

Somewhere in this thread i read about a possible problem with the nvram reset, i even tried pulling the DC adapter after the laptop turns off and wait 10-20 secs but then it just boots normal the next time. Any advice would be nice, i´m running out of options to test

Can someone explain what ME is (Intel Management Engine) ? Maybe a FW update of ME can fix this,i found https://www.mysn.de/driver/5_Sonstiges/Tools/ME-Update-Tool/ so it should be 11 in my case?

Here are some screeens from the bios before and after.
https://workupload.com/archive/DJH4reR

 

That ram requires XMP support to run at full speed, it's not an SPD profile.

To clarify can describe exactly the boot behavior of the machine in its current state? (Your time line is a little confusing).

 

This is describing my "normal" boot:
Now i have the XMG logo an its showing it flashed the right bios but initialising is way slower than before. I can still boot Windows but my sound doesnt work, my keyboard seems being emulated (takes about 30 secs to react to presses) and my wlan was messed up (i fixed this by using the repair option in intel setup). I can even play demanding games and if i use usb Keyboard/Mouse its all fine.

This describes my tries of getting the MESET to work:
After i found out i forgot the meset, i tried it out and it gives me black screen even after 10 minutes the Keyboard went from rgb to blue (fan was full on so it kinda works). I tried this over windows and with the apex shell for efi flash ( https://www.mysn.de/driver/5_Sonstiges/Tools/Shell_for_EFI_Flash.zip). I also tried disabling UEFI and removing the CMOS battery.

I will post a video later if i find the time, thanks for the fast response!

 

Recently I noticed that my 1060 clock speed is much lower after I sent it back for USB port repair. 1886 and now is 1228 at its best. Do you guys have any idea what's causing this?

 

Check your windows and nvidia power settings, look in MSI afterburner at what is going on.

 

I made a vid showing my clevo boot/behave:

https://workupload.com/file/wbMPHeEy

So i cannnot use 2666 mhz because the Notebook does not support XMP profiles?

@ryannguyen69 This could be normal (when games dont need to utilize full power these chips have a power saving mode resulting in a lower voltage/clock), try using GPU Z and let Afterburner display the clock, it should rise as high as possible on default. If not check your curve in Afterburner by pressing CTRL + F, you can see which Voltage/Clock you are running on the fly when you are testing and have the curve editor open. Maybe you pressed L which locks a certain point. You can also put the highest clock (not the highes possible, the highges in your whole curve) on 0,8V depending on you model you can reach a very good speed with a very low power consumption (1050TI 1455mhz, 1600 1600mhz, 1080TI 1650 Mhz). Theese are my tested values and worked stable in Games (even hunt showdown). Hope that helps.

 

Have you tried a re-install? Does the bios menu react normally to keyboard presses on the internal?

 

I not sure what was causing that. I found that the gpu is locked at 0.7v. After removed MSI AB and driver than reinstalled both and reapplied the curve. I have managed to get it run correctly. Thanks anyways, guys.

 

The keyboard is working fine when in bios. What do you mean by reinstall ? Windows or the BIOS update? I have not messed around with it since i did not want to mess up further and wait for some response.

 

After a "normal" day, it got back to the problem. In heaven, it reported the intended clock speed but the actual is much lower. Also, something cross my eyes is the 1060 is in "idle".I set to max power in the Nvidia control panel and the ultimate performance power profile in windows.

 

Sounds like the gpu is refusing to enter the highest P state, nvidia inspector can be used to try and force it. Either a driver or firmware issue most likely.

 

I did look though Nvidia inspector, nothing seems to alert me. Accidentally, when I tried to install the latest version of CC, I found that if my computer doesn't have it, the clock speed of GPU is completely fine.

 

It would have been a re-install of windows I was talking about by the way.

If it works in the bios normally then the firmware and hardware is likely fine.

My questions are usually like that, ruling out certain aspects to focus in on what's causing the issue.

 

Anyone having overheating problem? I have repasted my gpu and cpu with idc 4 and configured fan curve via Rlecviewer but no use. In games like Rainbow Six at Ultra or Origin even at lowest. I am having this issue. Heres a video from my tests. I can't understand it

 

What is overheating exactly? what are your temps?
I have no issues at all and perfect temperatures, even under full load, very happy!
Not even repasted GPU.
Did you undervolt?

 

Well only in game and rendering im getting these temperatures. Its almost mostly at 85 to 95 as you can see in the video i recorded and added from youtube. I tried undervolting at -0.130 but nothing has changed.

 

Try Pressing FN+1, then you should enter maximum fan mode.
-0.130v should do more than Nothing, i think.
Maybe turn off the Turbo mode for the cpu with throttlestop. this is what i do in most games because we are gpu limited anyway.
Then i only get 65°C

What are your temps and Settings in rainbow six? So i can compare.

 

Im using rlecviewer so i dont think that fn + 1 gonna work. I set it to 100% speed when it reaches 80. I can cool down cpu with those settings you saying but gpu is still same. Now im testing rainbow at Ultra preset with aa off its getting between 80 and 85 and running 1823 Mhz. Cpu is between 75 85 without any underclock or else

 

Mine is also running at 1228Mhz... Can't adjust

 

Guys i was reading the review at https://www.notebookcheck.com/Test-Schenker-XMG-Apex-15-Clevo-N950TP6-Laptop.296458.0.html here i noticed their i7 8700 was can and run at 4.60 Ghz. Mein cpu can almost run at max 4.30 ghz why is this happening?

 

If you disable C3 and C6, you lose out on the maximum multiplier when a single core is active. See... CPU needs to have at least the C3 C state enabled. Probably used added different bios settings.

 

Try to remove control center, it completely solves the problem for me.

 

Hi Meaker, tyvm for trying to help me out here. I finally fixed the errors and im using the laptop without problems since.

The reason i got this problem in the first place was that i read an instruction (in this thread i believe) where it say you should use the flashme instead of flash.nsh. THIS IS COMPLETELY WRONG AND I HOPE S1 DELETES THIS POST.

You can only use it if you are in the disable ME mode. What i did was flashing the flashme.nsh (it will call a nvram reset at the end), disabling the ME after and patching to the newest ME FW and it worked perfectly. The reason i didnt figue it out was that i still got the new boot logo and version entries which looked like the bios update was complete. But it seems i was only trying to flash the Management engine without it being disabled which led to that bug.

Back to my ram problem:
I read out values with speccy and CPUID which state that the 2666mhz are only available via XMPP (screens attached). It seems like this board does not support xmpp profiles? Since i want afap ram installed i either buy 2400 cl14 or is there anything im missing here?
Right now i´m using this one https://geizhals.de/crucial-ballist...bls2k8g4s26bfsd-a1592501.html?hloc=at&hloc=de

Heres the ram i want to get:
https://geizhals.de/kingston-hyperx-impact-so-dimm-8gb-hx424s14ib2-8-a1620462.html?hloc=at&hloc=de
The funny thing is that it isnt labeled as dual ranked but when looking at the manufactoreres detail page it show clearly that modules are at both sides. https://www.kingston.com/dataSheets/HX424S14IB2K2_16.pdf

@mcalago
I made a post at Page 33 which describes how far you can undervolt your gpu. Right now i am running @0,8V and over 1700 mhz and +300mhz ram without any crashes so far. In order to do that you need a) newest chipset driver (from intel not from your supplier) b) newest intel gpu drivers (i used the dch one, the intel driver was the main issue for bluescreens "nvlddmkm" error, be sure your control center looks like in my picture) and c) the latest nvidia driver. I am running the newest version of msi afterburner (not beta) and the newest statistics tuner. This way your gpu should be way cooler, i didnt reach more than around 80 like this (valley benchmark). Note that every GPU is different but i think 1600mhz shoul dnot be a problem for most cards. My Cpu is @-170mv and -130mv for the intel gpu (i5 8400). When watching a replay in Hots sometimes the cpu fan turns off oO.Hope this helps, here are some screens https://workupload.com/archive/gEKEZ92

 

I just installed exactly the HyperX ram from your link, works in 2666mhz mode without problems. I checked this with hwinfo. I've got the Clevo n957tp6 from Dubaro and all I did was a fresh install of Windows 10 and install the most important drivers from the Clevo homepage. Didn't change any settings whatsoever, it just worked plug and play.

The board seems to support 2666 mhz ram.

The 1060 is running at around 1700mhz constantly (MSI afterburner indicates that it hits power limit).

 

Cheers Houzu, thank you for making an account to reply here!

So it seems there is no support for xmpp.

What voltage you run at 1600 ? I just played around and i even got 1900mhz @ 0,9v looking stable, most of the time i dont need the extra power so i have not made a demanding profile yet :=

Can you confirm if the ram is dual ranked ? I think since it has coolers it should be.

You should be able to check this with cpuid at the SPD tab under ranks

 

Any heatspreader or sticker says nothing about the actual ram itself.

 

Yeah you are right.

 

Sorry i was wrong, i've got the same ram except it is CL15.
I installed those two https://www.conrad.de/de/kingston-l...-2-x-8-gb-ddr4-ram-2666-mhz-cl15-1836423.html

CPUID says it's single rank.

But i'm confused, CPUID just shows 1333mhz as bandwidth while checking through cmd shows 2667mhz.
Is this normal since 2*1333=2666? Or am i just running at 1333mhz and 2666mhz doesn't work?

If it really is 2666mhz then i'm running it without problems.

Screens: https://imgur.com/a/HAvk5oe

 

Yep, DDR = Double data rate. So your clock is 1333MHz and it gets double pumped to 2666MHz.

 

You mixed tech. See... Single rank memory vs. ram in single channel mode. Two different things.

 

I have a problem with my gpu. It's hitting power limit and can't maintain boost clock, doesn't matter what i do on MSI afterburner curve or if i have control center installed or not. The temps are in mid to high 70s. Clock drops after a few minutes in games (Witcher 3, 7 Days to die, Miscreated). Same happens in heaven benchmark. In this Video the guy maintains 1823 mhz and all he did was uninstall Control center, he just uses Afterburner for monitoring .
Where can i download latest BIOS?
I'm downloading all the drivers from clevo again and hope reinstalling them and doing a bios update will solve the problem. If it doesn't i'll do a clean reinstall of windows.
I'm afraid that i installed some false drivers from clevo website because i selected the wrong laptop model and that messed everything up.
Where can i find bios updates? http://www.clevo-systems.com/page/support-drivers is not available and in the manual it says bios updates should be available at https://www.clevo.com.tw/clevo_partners.asp?lang=en but it seems only clevo partners can log in.

edit: don't know why but in cs:go the gpu boosts up to 1848 and stays there. Maybe it was related to the Software i was testing with?

 

That sounds like normal operation, pascal is designed to boost under 100% load up until a limit is hit, either power, temperature, or max boost voltage. On mobile, power is most common. The yt vid you linked shows <100% gpu utilisation and only occasional power limit flag, so the limit is probably coming from frame cap or vsync to desktop refresh.

Some game loads are very light and will not load the gpu much at all (csgo) giving it more boost clocks, other heavier loads will hit the power limit harder and drop clocks well below base clock (furmark) while others won't (3dmark) despite also showing the "same" 100% load and gpu power draw

If you're hitting power limit and its not running too hot, there's literally no harm in maxing the core offset as high as is stable. It's not going to go past the power limit that's already being hit. It's effectively undervolting, freeing up power budget wasted on unnecessary voltage and putting it into higher clocks instead