New LG Gram 17

Does anybody know if the new one is shipping yet, BTW?

The CPU on this one is terrible and cannot even handle web browsing without severe throttling.

It runs slower than the Surface Pro with the equivalent processor, very disappointing.

Hopefully the new one does better.

 

I realized this morning that throttlestop has been updated to 8.74 and there is now an option to disable undervolting before the computer goes to sleep on the FIVR page.
This is the first thing I'll try as it was added to solve undervolting issues that looks exactly the same as what I am getting.

 

Define throttling. I don't see those types of issues doing basic things like web browsing. Even gaming with my eGPU is perfrectly fine with AAA games.

 

I'm loving the 17" 16:10 screen and would love to get my hands on one for productivity gains. Any rumors about an AMD flavored one? That would be awesome, especially after seeing the Asus g14's performance.

 

No, it wouldn't be able to handle the 35-45W TDP in the Ryzen 4800H/4900H that's in the G14. The other strike against that for me would be the 99.99% assurance of no Thunderbolt 3 with an AMD cpu. Very very few if any companies are going to put the engineering dollars into the extra metal and licensing to get TB3 with an AMD CPU.

You might want to take a look at the XPS 17 when that drops.

 

I guess we'll have to wait for intel to go <14nm for some more processing power.-.-

 

If memories serves me well... LG had never use AMD with their earlier laptops, so even in this Ryzen age, i don't think it'll impress the conservative koreans as well.

 

For big "ScreenScape" and slim build, lookout for the IdeaPad 3 from Lenovo coming to Europe soon...
https://psref.lenovo.com/Product/IdeaPad/IdeaPad_3_17ARE05

 

Nice, although I wish they could bump the resolution to 1080p. Seems like it's all 1600x900.

 

Alright, I should've posted this here a long time ago, but do any of you see tons of hardware interrupts?

I've reinstalled Windows fresh from USB like 6 times in the past couple weeks. Every single time it boots ok, interrupts are around 4k-5k/second (even though I think that's actually still mega-high; not really sure what the "typical" amount is supposed to be? but I read somewhere anything over 1000/s shouldn't be normal. Points towards a driver problem). As soon as I install some software and do the LG Update Center etc though, I'm getting usually around 10k-40k per 1 second, sometimes even way more:


https://imgur.com/a/66jWD4M

Sorted by CPU Time, it's consistently at or very near the top of the list. Something is eating the CPU alive slowly lol - it just takes a small amount of CPU resources, but does it constantly all the time while idle or doing light browsing etc. so you wouldn't even notice.

I'm pretty sure in my case there's very well a virus involved, maybe something to do with the LG Device Manager/Platform Manager drivers, the former of which had a known security vulnerability last year. I only found that out when googling after noticing some weird stuff going on with a bunch of services/processes elevated to the SYSTEM level, and a weird file with a weird extension deep in my AppData folder containing a lot of javascript that looked like it was manipulating Windows Defender. And when I saw that I about screamed because MsMpEng.exe is usually right under Interrupts as the highest on my procexp list for CPU Time...but I'm not sure if that's just normal for Defender to eat up a lot of resources, you all might see the same thing and not have a virus or anything...idk.

Anyway it's been driving me crazy as I've noticed it for months now. There might be nothing too wrong at all, just the throttling issues I've read in the rest of the thread, but idk those interrupts don't look normal, and given LG's craptacular performance as far as putting up-to-date drivers (or ANY drivers) on their site, I just suspect extremely that they bungled up something in their driver implementation job, as it's literally documented they did, at the very least, with the Device Manager vulnerability - you can read the original post by the guy who discovered it here, or google its CVE number (CVE-2019-8372: Local Privilege Elevation in LG Kernel Driver).
http://www.jackson-t.ca/lg-driver-lpe.html



tl;dr: sorry for long post but if you can, could anyone here download Process Explorer and post the # of your CSwitch Delta (changes in the number of CPU Context Switches) for the "Interrupts" process? Can also try checking if it's at the top of CPU Time after a couple hours of use.
(You'll have to add those columns in procexp btw, cpu time+cswitches aren't enabled by default)

 

If you have the Intel RST driver installed try uninstalling that.

 

Thanks, it feels like forever since I posted that but, a lot has happened lol. Seems the thread's dead-ish now which sucks, because man have I torn into this PC.

First, real quick side note about intel RST... I was just now looking through their Optane page, and apparently our i7-8565u and its board/chipset (CNL PCH-LP (U) Premium SKU...?) - at least that's what I'm seeing on my R.AAS7U1 (2019 model) under Platform Information in the advanced bios settings - are capable of supporting Optane?





I don't actually know much at all about optane but I've heard it's the bees knees.
What's more is I think maybe the RST driver causes so much grief ONLY because, LG might've misconfigured our BIOS settings.

That would also explain my interrupts problem. I think the ACPI settings are just not optimized right, and thus no matter RST driver installed or not, (I always get them with or without the driver, even on a clean install of windows; ~5k context switches/sec) the OS is constantly pinging the driver interface or whatever and one of them responds by sending a ton of interrupts.

If that's indeed the case, it would make sense because I have a theory that a lot of the settings under these crazy tricked-out unlocked bios options were left at weird defaults, basically I suspect LG might've definitely gotten lazy and dropped the ball on tuning everything correctly before shipping these laptops.
There's evidence of that if you look at certain other stuff, like a lot of Thunderbolt settings "disabled", and a bunch of the other various intel platform settings.
The best example might be in the Intel Advanced Menu > PCH-IO config > Serialio Configuration. In there I see I2C0, I2C3, and SPI1 being the only ones "enabled", and if you go to each of those controllers' settings near the bottom, it looks like intel suggests I2C0 to be the Touchpad by default for that serialio port, but in its Settings it's "Disabled". wtf?

I'm only slowly learning what all these options do through hectic google searching but my best guess is that LG must be using those 3 as just dummy controllers, because in the help pane for each controller's enabled/disabled setting, it says certain ones depend on each other - I2C0 depends on I2C1,2,and 3; and apparently UART0 "cannot be disabled when its child device is enabled, which it uses the CNVi bluetooth device (bth0) as an example.

It looks to me almost like they wanted to keep UART0 in play for some reason, which necessitated keeping its other codependent controllers on, and then they're just routing the touchpad/fingerprint/other input devices assigned to those controllers to the CPU some other way? And then just blocked whatever i/o they would've been sending?



..............Which, brings us to the final thing I wanted to put out there and get all your opinions on. Is there some kind of inherent security vulnerability that LG left wide open by leaving one of these bios options at their default?

Because it turns out I was right, in a way..... I DID have a virus, or at least some kind of malware type thing, and holy CRAP is it an absolute nightmare. I've been messing around and digging into the settings of every computer I got my hands on since I was basically a year old, and I've never wasted so much time & frustration on troubleshooting something "weird" going on in my machine, as I have in these past almost 2 (!) weeks -__________-

Besides the interrupts, I noticed some odd "signs" of something weird in Windows, but they were random and non-correlated and I couldn't pin down a single root source that pointed towards legit malware pretty much the whole time. It was just little anomalies like the Modified Date of C:\Windows\system32 files not matching each other; opening .exe and .mui files in notepad would reveal calls to .rdata and EFI resources that just looked like they shouldn't be there in the Pen & Tablet service's directory, to name one example (in program files\common files).... stuff like that. I was seriously starting to wonder if I'd gone off the deep end and looking for something that wasn't there but I was convinced was an attacker remoting into my PC somehow.

Well it actually was. I talked to some security people once I narrowed down the processes & services that were clearly amiss, and they told me it's apparently an old malware technique, and notoriously hard to either detect OR get rid of: using the WMI (windows management instrumentation) to remotely take over a regular win10 PC and set it up as if it were an Active Directory client (like you'd do with the computers on a corporate WAN if you were the company's IT guy), and then lock down Group Policy to prevent the user from making changes to their own system themselves. From there I guess you're able to do whatever you want, collect keystrokes, manipulate their network connections and make the system remotely download more goodies from your own servers, etc etc. Looking in the event logs, there were a lot of log categories turned off/missing entirely and permissions set to restrict my own access to them (as an Administrator on my own machine), clueing me in to the fact that they were almost definitely hiding their own tracks as they went along too.

Anyway.......... I still haven't figured out what the hell I can even do about it lol, because worst of all this thing seems to replicate itself onto any new Windows install USB's I make with the media creation tool, and instantly takes over any blank USB's I plug in, making autorun.inf's in hidden folders with invalid characters in their path names (like colons especially "D:\System Volume Information\:\autorun.inf:"), as I found out with some of the file monitoring NirSoft utilities.
It's a freaking monster. I think I might try just getting a friend to make a (hopefully) 100% clean windows USB for me and drop it off at my house, that's literally the only thing left to try, because I've exhausted every other option I can think of to wipe my drives and do a clean reinstall without the malicious files setting themselves up from the very beginning. And yeah I've probably done 6-8 reinstalls in the past couple weeks, across my Gram and 2 different Thinkpads I had sitting around/in storage.. nothing's worked so far. Always end up with WMIprvSE.exe running along with sketchy looking rundll32.exe processes and COM Host Provider services to go with them.


But to get back to the Gram specifically, I've been trying to nail down where this could've started, and I think it has something to do with all the kernel debugging settings in the bios. It looks like "Legacy UART0" is set to handle any kernel debugging output, although consent for debugging looks to be Disabled so maybe another setting or combination of settings in the other menus, like serialio.
I did notice crucially that in the windows setup logs it has some stuff right at the beginning where it's loading the bluetooth drivers (bth0.inf or something) that look inconsistent with the rest of the driver setup, and if that has anything to do with the "child item" (bth0) being enabled along with Connectivity (CNVi), and that it says that being enabled makes UART0 *unable* to be *disabled* - well maybe UART0 is being forced on despite what it says under the actual kernel debugging settings, and it happily broadcasts our kernel's activities the moment you boot up, (oh and there is a Network Kernel Debugging driver sitting there in Device Manager.... so displaying kernel debug info over any bluetooth/wifi/wired networks) at which point attackers sniffing for the right ports can get the info they need to worm their way into Windows and make you their obedient little WMI client machine.

It took me a lot of yelling and lack of sleep to put all this together, I know it sounds probably downright freaking insane, but I took dump logs, screenshots, file copies, the whole 9 yards to prove it, if anyone's interested.

Whoever it is that currently owns my entire Gram via this WMI attack has done a great job of making me hilariously paranoid lately, lol, but I need some input from everyone else here; check your Group Policy editor, sort the administrative templates by Managed or Unmanaged, are almost all of them showing up as managed? If so you might have the same problem as me and not even know it.

Either way, I'd really like to get some real clarification on what all these bios settings do exactly, what they SHOULD be set to by default to avoid some kind of catastrophic and easily-preventable security hole that might've resulted from LG shipping a broken bios setup, and why all of them were left like that, and the bios so thoroughly unlocked in the first place. (and I'm not complaining about it being unlocked, btw... I didn't even know manufacturers had access to this many options and I'm loving the idea of tweaking all of them myself to potentially let my computer run better than they might've been able to do themselves. BUT - I'd really at least like to have the assurance that running the laptop with everything default isn't horribly compromised and leaving me open to needless personal info leaks right out of the gates)

 

Anyone using this laptop in hot environments? I'm in 38C 80% humidity most days this time of year and finding the CPU gets hot, the area above the F8-F10 keys is too hot to touch really. Can this cause damage long term do you guys think? Maybe I should take the back off and give it a clean I suppose, could be some dust adding to the issue. If I run Speccy it says the CPUs are about 65-70C.

 

Shouldn't hurt anything. CPU can run at 90 with no issues, it's within spec.

 

That's a lot... I can maybe comment on a couple things

Optane
We don't have optane enabled memory/storage on this unit, it's useless here, and would not help it achieve anything further anyway. The default microsoft AHCI driver is plenty capable, and in my case Intel's driver was causing performance issues. As soon as I uninstalled that driver my problems disappeared.

Malware
If you haven't found any specific malware, how do you know you are/were infected with malware? System32 file modified dates being different is not generally an issue unless you've found evidence of files that are known to only legitimately have a specific date. You aren't going to discover anything useful opening binaries in notepad I don't think. There are a lot of processes that run under svchost and utilize rundll32, one on my machine happens to be nvidia stuff. Have you used Process Explorer to try to get better info about your process trees to see what is utilizing rundll32 to load a library? Download it here: https://docs.microsoft.com/en-us/sysinternals/downloads/process-explorer and use the 64-bit executable.

Also

This is typical as well, you don't want your user to have this type of access generally, can you specifically tell me which things you are talking about? Usually windows will throw a User Access Control dialog at you so you can confirm that you want to do something with an escalated privilege. Think of it as an analogue to sudo on a unix-like OS.


Configs
Just picking an example. Just because a Thunderbolt settings are set to disabled for instance doesn't mean that is a wrong configuration setting. I'm using an eGPU over TB3 and it works fantastically at whatever defaults LG used. I don't think there's a more strenuous TB peripheral than that.

It seems like you're digging into a lot of things and not quite understanding exactly what they mean and inferring or perhaps being told erroneously that they should be configured a certain way but that isn't universally true. Something being disabled isn't specifically a bad thing depending on what it is, and context is very important there.

 

OK, thanks for feedback - appreciated.

The outdoor temperatures have cooled off to a more reasonable 34C these days and the LG17 is running much cooler and just fine.

On another note I use Win10 Home mostly that the LG came with and even today the update says ver 2004 is not yet ready for my device and to wait.

 

Hi all. I've had a Gram 15 (15Z980) for a few months now and love it! The amount they've packed in to such a lightweight laptop is really impressive. Excellent screen too.

Anyway have finally gotten around to buying an eGPU for it (a Gigabyte Aorus Box GTX1070), and am having real problems with the Thunderbolt 3 aspect. The eGPU doesn't work when I plug it in, I just get a "limited functionality" message. I can't find any reference to a Thunderbolt controller in device manager even with the eGPU plugged in. I thought I'd solved it when I went into the secret BIOS and found Thunderbolt disabled, however after enabling it it's still not working. I've downloaded the LG Control Centre and it doesn't show any drivers relating to Thunderbolt. It's like this laptop doesn't have Thunderbolt, but I know it should do! Were there any 980 model Grams without Thunderbolt? Have tried the eGPU with a friend's HP Elitebook and it worked immediately...

Btw have been reading this thread for a while, great to see some really useful information being shared by the community! Have found little to no information on the Gram elsewhere on the web. Keep it up

 

Most of the Gram 15 did not have Thunderbolt except of the full loaded version, i7, 16g and 2x512 SSD. Some of the lesser versions were without Thunderbolt. I had to go for the full version to get Thunderbolt. I use Akitio Node Pro with an RTX2080 and works perfectly, some issues in only the beginning with the NVIDIA drivers. I checked at the BIOS, I have the 15Z980-R.AAS9U1

 

Which exact model Gram 15 is it? There are sub-models of the 15Z980.

I would look something like this in AIDA64



Also should just be listed as Thunderbolt controller under System in Device Manager

 

Thanks very much for the info guys, in all my research I had never seen it mentioned that only the i7 model had TB3. Mine's the i5 8250U version (256GB SSD/8GB RAM) and doesn't have any reference to Thunderbolt controller in Device Manager or AIDA64, so safe to conclude it doesn't have TB3 after all.

Bit of a shame as that was one of the main selling points originally. Ah well.

 

That's a bummer. Maybe try to sell it and grab one with TB3?

 

I have the LG 17 from 2019, the one sold by Costco. Here's an unusual request. Currently there's enough storage (512GB), but I can see down the road (2 years) when I'd like to populate the second SSD slot.

Many of you have describe problems filling in the second slot and some have pointed to a bug in the BIOS. Here's the request: What SSD do I look for in 2022 or so for this second slot? Do I try to match the existing and original one (and thus go searching for an antique)? Also, how to I update the BIOS to this rig, since LG's update software seems rather anemic?

Thanks in advance.

mjclifford

 

I've run two Toshiba NVME 512GB (KXG50ZNV512G) drives since I got it last year no problems.

 

Oddly enough try searching for new drivers on the Firmware object in device manager. Worked for me, someone else earlier in the thread discovered that new firmware can be found using this. I was surprised actually, I'm so used to finding new firmware on the manufacturers site.



I wouldn't worry too much about incompatibility, but since the 2nd slot is only a PCIe 2x slot, you can get a somewhat cheaper drive, like the WD SN550 and save a little cash. If it doesn't work return it for a different one is all I could suggest.

 

I bought the LG17 2019 in May last year and at same time bought a 1TB SSD NVMe PCIe 3x4 M.2 second drive that worked without any issues at all. I read the system SSD that the LG ships with is a slower than the extra drive I got, but I wasn't that bothered and cannot see much difference for my needs. I am just happy to have the extra drive space.

 

I have an i5-8265U model (imported from Korea), but do have TB3 and use it daily with a Razer Core, so it is not limited to i7. Is there even a Gram17 with 8250?

 

EDIT: I think I'm wrong, see the following posts.

Not only that, TB3 controller usually doesn't appear if you aren't actually plugging a TB3 device into it. Sometimes, this even fools the TB3 drivers on my computers (2 laptops so far and 1 desktop).

It's, IMO, a horrible design flaw, but Intel is still the only source for TB3 chips, so we all have to live with it.

 

I just pulled my eGPU and reboot just to make sure and I still see the TB3 controller in AIDA64 under:
Devices > Windows Devices > System devices > Thunderbolt(TM) Controller - 15D9

Edit: I am assuming the driver has to be installed for this to show up...

 

A somewhat similar problem I'm facing is that the mini-sd card reader most of the time just doesn't show up in my device manager. It occasionally does work, but most of the time, nothing can be found and if I insert a mini-sd card, nothing happens. Has anyone else experienced that? Other than that the machine performs awesome and the latest ThrottleStop update has fixed my wake-from-sleep bluescreens.

 

I'm looking at my TB3 controller right now in Device Manager, though there have been times where it didn't show up in the past.

I never followed up in this thread, but someone else couldn't find it, either.
https://forums.lenovo.com/t5/ThinkP...device-manager-for-X1-Carbon-6th/td-p/4314736
As it turns out, it may have been a Thinkpad UEFI setting.

 

I caved and got this bad boy from costco... Loving it so far. I set dpi to 100% and It's surprisingly usable. I'm a sucker for maximizing space, we'll see if my eyes complain in a few days.
I got into the hidden bios and thought I enabled the options to allow for undervolting but can't seem to see the changes take effect when I run CB15 and observe numbers of TS.

Any tips on this front?

edit: nm I restored dpi to 120%. Opening coding is overwhelming on my eyes at that dpi.
edit2: I applied the undervolts in the bios and they seem to hold. Wish I could get them through TS so I could experiment more easily.

 

LOL.. My eyes are nowhere near good enough to for that, i think I'm using 125% in Linux and 150% in Win 10.

I don't think there is a way to undervolt without a 3rd party app like Throttlestop. It's easy enough to set it up to run at startup. I haven't bothered trying to undervolt in Linux, I would imagine there is a tool for it out there.

 

Costco will have this unit at $1250 (down from $1499) 8/05 so hoping I can drop by to readjust the price with their customer service.

 

Right you are, just went looking around and found this coupon that will be valid on 08/05

 

Nice Deal!
Long time lurker here and finally decided to register.
I found the following which is Gram 17 like but only better and not that much heavier.
It's the LG Ultra PC 17 ... Take a look and let's chat it up

https://www.lg.com/us/laptops/lg-17u70n-r.aas8u1-ultra-slim-laptop

And yes its U.S.A. available...

https://www.amazon.com/gp/product/B...&pf_rd_p=edaba0ee-c2fe-4124-9f5d-b31d6b1bfbee

Specs...
LG Ultra PC High Performance Laptop - 17" IPS WQXGA (2560 x 1600) Display and Intel 10th Generation Intel Core i7-10510U CPU, NVIDIA GTX1650 GDDR5 4GB, 16GB DDR4 2666 MHz RAM - 512GB NVMe SSD, Dual Cooling

 

Was considering this bad boy but I already have a powerful 17"... that said if this was my only machine then definitely. Only worry is I hope they slapped more cooling for the cpu (14nm and GPU)

 

Linus Tech Tips actually reviewed it.
Looks like dual cooling



I can't seem to paste a pic here so see 8:55 minute mark

 

The aspect that kills this particular one for me is the lack of Thunderbolt 3, a 1650 isn't nearly enough GPU for 1600p.

 

It has it - Search the web page carefully toward the bottom.
Spec Tab is a little incomplete (doesn't show TB3 for LG Gram 17 Costco either).

 

Where exactly? If you are referring to this, it's about the Gram 17 not the Ultra PC 17

 

My bad. You are absolutely right. It is regarding the Gram and not the Ultra.
I got blinded in all the excitement. Oh well

 

Forgot to mention one thing regarding cooling...
Did you notice the the vented bottom on the Linus Tech Tips video?

Has anyone tried running bottomless to see how temps are affected?

If it's a decent enough drop, the dremel may be soon to follow, but it's not something I'd like to do unless it's confirmed to actually improve things.

 

New LG gram 17 owner here, loving the lightness and the screen... however, this morning the CPU temp was up to 94 degrees C (coretemp monitoring).
I was charging via USB-C at the time, but that shouldn't matter I would think.

Are these normal temps or should I be worried that I have a "bad" CPU? Except for the temps it is working fine... hope you fellow owners can shed some light...
spec is i7 1065G7, 16Gb, 512Gb SSD...

 

I was able to Undervolt with Intel Extreme Tuning Utility with setting as follow...

-.110 (that's negative offset) for CPU Core Voltage
-.110 (that's negative offset) for Cache Voltage
13 Watts Turbo Boost Power Max reduced from original 14.5 Watts
3 Second Turbo Boost Power Time Window

Stress Test ran for 10 Hours on AC Power with Temps Averaging 85-87C.
Not the greatest, but better than pegged at 90C+ the whole time.
CPU Max Core Freq also averaged better staying at or above 3Ghz.

 

Is this just idling? Check task manager and see if there was something consuming a lot of CPU. If not, that's a pretty poor TIM application (heat sink paste..). It's not terribly difficult to take it apart and re-paste, but still. It could understand if it was under heavy load, it WILL sit at 90+ under heavy load that's just the way it is. Let us know under what conditions this is happening.

 

Newer Intel WiFi drivers are causing me disconnection problems. I did a lot of pocking around and nothing worked. I did changes to my WiFi router, used different settings on the laptop and nothing worked. Finally my solution was to install the OEM LG driver which is very stable and never experienced a single problem. You might ask, so why bother updating drivers, well I've been always able to update with original Intel drivers to all my machines and this is the first time I have this issue. Mine WiFi adapter is the Intel AC 9560 160 and the OEM driver is from 01/2019, the newer Bluetooth and the video drivers from Intel work without issues.
Any thoughts?

 

I am using the latest drivers without any issues, might be that I have a different router though. Make sure the firmware on your router is up to date as well (sounds like you probably did that).

 

Thanks for your help hfm. No it wasn't idling, but not exactly 100% cpu usage either. I was trying out a Geforce-Now streaming game, but maybe there was some windows indexing going on in the background. Temperatures seem very unstable, maybe that's normal for this type of cpu or because of bad TIM.
I am used to a 6700HQ which only reaches 90+ when under load after like a minute or two. This cpu just shoots up. Not sure if I should open the laptop and try to apply paste or to request a replacement. Or would a replacement do the exact same thing?

 

The hard truth is that this unit does not have an extremely capable heat management solution. The 2020 model looks to be a little better than the 2019 model, but it's still subpar IMO. Under a decent amount of load the heat management solution is going to get saturated rather quickly. If the TIM application is subpar, it can be slightly improved using a better quality product, but there's only so much you can do. Another aspect to this is that they use the same exact heatsink/pipe assembly in the entire Gram Line: 14,15 and 17. I understand why they do this. Same components just in increasingly smaller form factor, it can save them a significant amount of engineering dollars and manufacturing cost just designing one assembly to cool a 15W CPU in increasingly smaller units. It even looks like nearly the same main and port layout boards in all units. There seems to be a little difference between the 14/15 and 17 for the smaller board for ports on the right side of the unit, but the MAIN board looks identical between all three, just less empty space n there as the notebooks get smaller since the boards for the ports are separated and just connected by cables they can use them in all the units.

Here's shots of the 2020 line...
Gram 14


Gram 15


Gram 17


Obviously there's room to improve the heat management in the 17, but I get why they would use the same parts across the line since it's essentially the same 15W CPU.


Back to the heatsink and managing heat in general outside of software based solutions like undervolting.
If you look in the history of the thread you'll see where people have created their own assemblies with dual fans or larger heat sinks attached to the heat pipe, which requires some modification of the chassis to allow exhaust from the 2nd fan.

I myself have tried grizzly kryonaut, results were marginally better but there was still a little core temp disparity there where two cores on the same side of the die were always hotter. With the 2019 model's 3 screw mounting it's tough to get a good even contact point. I most recently installed coollaboratory liquid metalpad and that created more even core temps. But, under heavier loads all of the paste performed almost the same due to the heat sink and heat pipe assembly just getting saturated to where it just couldn't remove heat any faster no matter how good the thermal conductivity between the die and pipe was. From pictures it looks like the 2020 model uses a new 4 screw method to mount it, which is way better as it's more even pressure, and the heat sink part of the assembly was made a little larger and is angled slightly to try to provide more surface area. Not sure exactly how much better it is, but just looking at it I would imagine it has the capability to remove slightly more heat.

One think that can help tremendously is using thermal pads to bridge the heat pipe to the chassis so that the magnesium alloy of the bottom chassis can sink away some of the heat, but then you run into a comfort problem if you are just using it on your lap where the bottom of the unit will get hot.

I'm hoping they yet again improve the heat sink assembly for the Gram 17 (and I guess the entire line in general) in 2021. I'm not holding out much hope though. If they continue to use the same exact heatsink/heatpipe/fan assembly across the entire line they will never put something larger in the Gram 17 that can utilize the empty space.

 

Thanks for the quick reply. I actually tried two different routers and yes I keep the firmware up to date. I might try again with the newest Intel drivers since I stop using them a few months ago and I will see if there is any improvement. The OEM driver is also an Intel of course so I don't really know what so special about that driver and the newest, but if it works for you it could be the "air" in my parts of woods .