*** The Official 2019 MSI GS65 Stealth with RTX GPUs Owners and Discussions Lounge ***

The worst that could happen would be getting the laptop stuck in a boot loop if the ram is really poorly configured imo but that can easily be fixed by resetting the bios (holding power button for a few minutes)

Removing the power limit could also result in thermal throttling if your laptop has poor cooling. A simple repaste should overcome this most of the time.

 

It's not realy bios editing, just bios configuring. Most other brand's don't show these settings by default so to enable them they have to modify and reflash the bios. That's where most of the risk lies. There is still however risk involved.

That being said, this bios settings will only help if you are being power limited before being thermal limited, this is almost never the case with non ramped up fans and stock thermal paste.

Also the quoted settings will give about a (55 * 100/50) + 31.999 ~= 142W power limit. I kept IMON slope at default and set offset to -10000 ~= 65W total and that was enough, with my undervolt, to never power limit throttle while maintaining maximum 6 core turbo clock.

 

I just finished installing a second ssd and applying liquid metal paste.

After application while pulling the liquid metal nozzle away a saw to my horror a tiny drop fall from the nozzle and land presentably somewhere on the board. I could not find the drop so ended up spending two hours washing the whole side of the board down with 2 large bottles of 91% isopropyl alcohol and 3 cans of compressed air. The flipped board design actually helped here as the motherboard was already completely removed at this point.

While it ended up booting up with no issues there will always be this lingering thought of a tiny drop of liquid metal still somewhere in the board just waiting to fry something at the worst possible time.

 

If you know what you're changing and read up on what they are and the safe ranges then there really isn't much of any "risk". Now if you take out some limiters and overvolt or have prolonged long temps that are not within safe ranges then yes you can potentially damage or degrade your hardware. With mild overclocking and tweaking the most you'll see is hardware lock ups, probably corrupted OS drives (constant crashing) and if it doesnt reboot then restarting BIOS etc...

 

So BB is having a bit of a sale again 1399 for the 1660ti/144hz/16gb/512 model or 1699 for the 2060/240hz/32gb/512 model i can get that last one new on ebay for 1600 it comes out to a final difference of 218.00 between ebay and BB due to taxes.....I know imma need to repaste these and ive done so with my gs60 and GS63VR already so its not new to me.....BB adavnatge is the physical store being there in case i need it...always much simpler than just shipping stuff out..id love to go with HIDEVOLution but pricing is a bit much for me to have it modded by them and all also i couldnt find the 9th gen models on their website.......anyhows...any suggestions? is 200 bucks worth the piece of mind of having an actual retail store to complain to? Your thoughts?? I have someone coming over from the US on the 25th so i need to make a decision well before then so i can make sure they have it..BB id buy today or tomorrowso someone can pick it up.

 

anyone?

 

I think HIDEvolution's prices are too high. I also don't trust Best Buy. 240Hz screen with a power gimped 2060 is ridiculous as well. If these are the choices, I'd get the 1660ti 16GB and buy a 32Gb kit of micron rev.E or Samsung b-die. From pcpartpicker.

Sent from my SM-N960U using Tapatalk

 

Getting it at a retail store is always best when it comes to return/swaps. Having your buyer test it out for possible return is a lot of peace of mind. I'd also go with the 2060 over the 1660ti. 99% of games, sure the 1660ti is within spitting distance of the 2060, but it's nice to have the option of ray tracing and dlss to play with. More ram is nice too, while not terribly difficult it's nice to not have to open up the GS65 for upgrades whenever possible. Also, the 240hz panel is a newer screen. You won't be hitting 240fps but I have a feeling the 240hz screen is slightly brighter (based off reviews i've seen) than the 144hz screen which is borderline too dim for my tastes. I'm looking for a 240hz panel to upgrade mine with to verify that though.

 

Ill be taking it aprt anyways as im going to install 2x1tb nvmes on it...so i will be repasting with kryonaut and new fujipoly pads...ive heard these are pretty mch a must on these to get adequate cooling..am i wrong??? I mena one teardown should be all i veern ened to do if i get it all done in one go?

 

Make sure you get 0.5mm and 1mm thermal pads. The GS65 uses both. I used 1mm Thermal Grizzly pads and 0.5mm fujipoly pads on mine. Also, and this is just me, I got better performance using Phobya NanoGrease Extreme over Kryonaut. I read somewhere Phobya is good for laptops where they don't always have optimal pressure from the heatsink to the cpu/gpu. I wasn't happy with Kryonaut's performance so I redid it with Phobya and had a noticeable improvement. Granted maybe I applied it better or better distributed pressure on the heatsink, I don't know really but it was an improvement. I had a large temp difference between the cores prior and now they're a bit more close to each other. Kryonaut is good stuff, I use it on everything else but for this application I just wanted to give you some personal experience on my end.

Also, I used a copper heatsink on my nvme drive to help shield the keyboard from these hot m2 drives. I keep reading about dying LED's/damaged keyboards so for a little piece of mind I tried this. Don't know if it really helps any, but I feel it'll prevent a hotspot by distributing the heat.

https://imgur.com/MTOhj5o

 

Hey everyone,
It's been 8 months since the early 2019 GS65 launched and I decided to make one last review on it.

https://www.theeverydayenthusiast.com/home/long-term-review-early-2019-msi-gs65

Inside are all the various tweaks and optimizations I've made these past 8 months along with updated benchmarks for most new titles (also in 2560 x 1080 Ultrawide). This laptop has held up really well and I expect it to last me until the next generation of RTX GPUs.

Hope this helps!

-JRey

 

Can u ell me where u got the heatsink for the nvme?? Also i have IC diamond,AS5,Kryonaut here with me can i use any of these or do u truly suggest going with phobya or gelid?

 

https://www.amazon.com/gp/product/B07QNRP9HK and I just used some Kapton tape to keep it together. It comes with some type of thick rubber bands, but I didn't like them. Felt loose, made me think it would come loose and the last thing you need is a piece of copper come free in your laptop and shorting everything out.

Eh, I really don't feel comfortable recommending one over another. Like I said, it was just in my personal experience phobya worked better. Maybe your heatsink sits better than mine and it's not an issue? Maybe I applied the paste better with the phobya, I have no idea. I just remember having an 8-10C difference in core temp between the cores but after switching to Phobya the cores were with 4-5C of each other and over a couple of degrees lower. Anything will be better than the stock paste, so probably just stick with what you have but just keep it in mind. I can never leave well enough alone and just keep tinkering.

 

I just ordered some GC gelid extreme and CM maker gel nano.....Imma try the Cm maker gel and see how that does, read it should work well .if not its good to have options...i might just put a line of ICD and be done with it ...i have artic clean remover so i can clean up stuff pretty well...gelid seems like a safe bet also as its been used alot but so has ICD so wish me luch also what method of application did u use...spread or linein the middle or pea? IM going to look into those nvme heatsink lik eu im a bit scared of having those inside the laptop but may work with that tape mod u did ill order them and see how securely i can get them on

 

I use the spread method, just because I feel laptop heatsinks don't apply sufficient pressure.

Just take your time, I believe I took mine apart half a dozen times before one night I got a little impatient and reckless, damaged the antenna connection for the wifi chip. Replacing the antenna was not fun, let me tell you.....

Which leads me to suggest..... upgrade the wifi chip while you're at it because.... why not?

https://www.amazon.com/gp/product/B07Q4YYQC2

 

Not bad for stock ram lol

Still debating on jumping to 32GB though

 

Ah damn just made me spend more money!!!! Just ordered it...does it make a differece truly?

 

I prefer Intel over Killer but the AC 9560 that came with the laptop isnt bad at all.

You're pretty much going:
WiFi 5 (802.11ac) -> WiFi 6 (802.11ax)
1.73Gbps -> 2.4Gbps in bandwidth
https://ark.intel.com/content/www/us/en/ark/compare.html?productIds=99446,189347 More Read.


So to actually make full use of it you need an ax router.. or if you just want future support I guess.

https://www.amazon.com/WiFi-AX200-PCs-802-11AX-Bluetooth-NGW/dp/B07TLBNSZQ/ here is the link if you're interested as well.

 

I haven't tested my self but user reported even if non wifi 6 router, signal was improved with newer wifi card.

Sent from my BLA-L09 using Tapatalk

 

Actually I take it back. The new Killer AX1650 uses AX200 (Intel) hardware. On top of that is all the "Killer" tweaks and optimizations.

So at this point its preference.

 

Hey guys,

First post here ... I have a MSI GS65 Stealth-666 (i9-9980H / 240 Hz display) that I’ve been using for about 40 days. Since the day I got it, I’ve frequently tried to use it with an eGPU: a Sonnet Breakaway Box 650 and an RTX 2080 Ti.

When connected to the eGPU, I got fairly frequent whole system hard-locks. The mouse cursor and keyboard would completely freeze up, the screen would stop refreshing, and sound playback would permanently stutter on whatever was in the sound buffer. This would happen between once and four or five times per day.

I started trying to troubleshoot this freeze by updating drivers (motherboard, Nvidia, Thunderbolt 3), by changing the physical Thunderbolt 3 cable connecting the laptop to the eGPU, updating the BIOS and EC... to no avail.

My next troubleshooting steps were to use Right Ctrl + Right Shift + Left Alt + F2 in the BIOS to look at the Thunderbolt controller settings under the hood. I tried experimenting with a few option settings, and then when I rebooted, the Thunderbolt controller stopped responding.

At this point, my Thunderbolt/USB-C port is completely dead in all possible ways. USB devices do not recognize. Thunderbolt 3 devices do not recognize, whether cold or hot-plugged. In the BIOS, it says the Thunderbolt security status is “Unknow status” (typo is intentional), and it doesn’t change when I set the security between “None” and “User authorization”.

I’ve tried all kinds of things to restore Thunderbolt to working again:

• Reset the BIOS by holding in the power button for 60 seconds with AC power disconnected.
• “Load Optimal Defaults” in the BIOS.
• Re-flashed the latest BIOS update from the MSI website.
• Re-flashed the latest EC update from the MSI website.
• Tried 3 different versions of Thunderbolt 3 drivers in Windows 10 v1903.
• Booted into an Ubuntu 19.04 Live USB environment and checked if Thunderbolt devices (or USB devices plugged in via the USB-C port) were recognized there. They weren’t, despite verifying that the right drivers and kernel modules were loaded. I was unable to “force on” the Thunderbolt controller using WMI, either.
• Looked for a Thunderbolt firmware update on the MSI site, but they don’t seem to have one for my laptop at all (not even the “original”). I’ve never tried flashing the Thunderbolt NVM (firmware) before, by the way.

I’ve ruled out software by trying two different operating systems. I’ve ruled out the system BIOS. I’ve ruled out hardware by trying different USB-C / Thunderbolt 3 devices with different cables, including an external SSD and two different eGPUs with two different eGPU enclosures. I’m 100% sure that the problem is the Thunderbolt 3 controller not recognizing.

The Thunderbolt settings in the BIOS are set to defaults now because I reset the BIOS, so it’s not due to my customizations (unless an earlier customization somehow broke the Thunderbolt controller or its firmware).

When I try to change the “Windows 10 Thunderbolt” setting in the BIOS, it gives me a prompt that suggests the BIOS needs to communicate with the Thunderbolt controller to store a setting, but it immediately fails and says host could not be reached — like the Thunderbolt controller is not responding to the host system.

Is there any way to force it to reset back to defaults, or can anyone provide me a way to forcibly flash the latest Thunderbolt firmware? If not, is my only option to RMA the unit and wait weeks for it to ship back and forth from City of Industry (on the exact opposite side of the country from where I live)?

The rest of the system functions fine, but I’m not going to continue to use a $4899 laptop (I paid for several expensive RAM and SSD upgrades from a system integrator) with a broken port when I have a top-of-the-line eGPU that I want to use for higher performance.

Also, anyone else having issues with Thunderbolt 3 eGPU stability, or is that just me? Could be the TB3 controller was bad on arrival...

 

It's hard to tell what was changed in the Thunderbolt options under the BIOS and many times one of these options can upset the controllers.
I would first try resetting the settings back to defaults if you can find out what the default options were.
Also, have you checked with MSI Tech Support to possibly obtain the TB firmware and try updating it?

 

I remember what the defaults were, so I set them all back to defaults. But I actually did an experiment to determine if the options were “host-side” (stored in the UEFI BIOS settings) or “NVM-side” (stored in the Thunderbolt controller) — I changed the settings to mostly their opposite values, wrote them down, then flashed the latest system BIOS, which wiped out all of my BIOS settings... When I did this, all of the Thunderbolt settings reverted to their out-of-the-box defaults. That means all the settings available in the BIOS are host-side, with the exception of the “Windows 10 Thunderbolt Support” option that is either Enabled, Enabled + RTD3, or Disabled... this one seems to try to program the Thunderbolt NVM (and fails). It defaults to “Enabled” out of the box no matter what it was set to before the BIOS flash, but I assume that’s just the host side default if it can’t communicate with the TB3 chip.

I have a web ticket open with MSI to get the original firmware, but I’m not hopeful of being able to actually flash the firmware, because the method of forcibly applying power to the TB3 controller under Linux, that seems to work for everyone else to make the Linux Thunderbolt drivers recognize the controller, didn’t work for me.

For further debugging: I know it’s not the Thunderbolt 3 *port* in the machine itself that’s damaged, because when I plug in the TB3 cable from my eGPU enclosure, the eGPU’s power supply’s relay clicks on, and the GPU’s fans start to spin. So the eGPU is aware that it’s connected to a valid port, electrically, but the controller it’s talking to is “braindead”.

BTW, the system's full part number is GS65 Stealth 9SG-666US-BB9988H32GXXDX10PH.

 

Does MSI just submit form replies to everything? It's like they didn't even read my request. Wow.

Thanks, MSI, for confirming that I should've bought a Razer or stuck with Apple.

 

Try calling them directly.

 

So i got my GS65 today..its the BB 240hz/2060/32gb/512gb one..do the tweaks in the 8750h guide apply to the 9750h?

 

I posted a short optimisation guide here recently. Try going back a few pages.

 

Is it the one on this page?

http://forum.notebookreview.com/thr...owners-and-discussions-lounge.827160/page-155

 

Do u guys limit the turbo multiplier in TS?

 

Limiting clock speed is for the weak

 

I do for normal office use. I think mine is 32x or 36x keeps its at 35°C idle and about 65°C max.

It's fast enough for web surf, office use.

I hotkey to my "performance" profile on TS when gaming.

Sent from my BLA-L09 using Tapatalk

 

Hello all,

I've noticed recently that the Killer network software (specifically process kfeco10x64.sys) is causing intermittent high CPU usage/fans spin up to high in bursts--this issue only occurs when I am connected over ethernet cable--do I need this software? If I just uninstall the whole Killer control center will this cause any issues?

Thanks,
N123

 

We dont all have LM on em....hehehhee

Im going to repaste and repad im going to use Cm maker gel nano do i use the line method on cpu and x on gpu or do i spread? I have fujipoly pads on stby also to switch them out

 

Are you repasted?

 

Kryonaut on CPU and CM MasterGel Maker Nano on GPU/VRMs.

Swapped all pads with Artic.

Sent from my BLA-L09 using Tapatalk

 

I personally use the line method for long rectangular dies and the dot method for square dies.

 

does someone know why my cpu stays hot after repaste with liquid metal? (9 gen)

 

why different pastes? I have both should I do the same?

 

Oh my kryonaut took 1 month to ship and I only bought 1g. So didn't want to spend all of it to do both.

I have 4g of the other and easy to find so opted to use that for vrm and GPU.




Sent from my BLA-L09 using Tapatalk

 

Hey if you get Optimus working it will lower scores by a pinch.
Reinstalled OS 1809
installed iGPU and chipset
Installed stripped Nvidia Driver (left optimus and driver supports)
Silent Option

its back, Optimus is working but I have no access to the control panel and Geforce experience didn't install correctly so I uninstalled it.

Gaming was not impacted with Optimus that I could recognize However benchs are now 6200 for timespy and 15200 for firestrike after the changes on the same settings for MSI Afterburner and Throttlestop
I did flash the 2060 90w bios to test after confirming scores, but given I used Gelid Extreme and not LM I did not like how loud the laptop was getting.

9750h/2060/32GB/512GB Stealth 478 9xx
Currently running 1605 mhz @ 725mV
Wow Classic @ 10 with fps locked at 120 and CPU set to step from 35-42 on all cores
64C. Runs under 60% both fans. When Idle, I get to see the little white light by the power button


I cannot get any active OC in Afterburner without Optimus turning off.

Fixed loadline to pull correct voltage for CPU in bios and can maintain a -65 on Core i7 using throttlestop.

Hopefully this helps..

 

OH I forgot, GET OPTIMUS 8+ hours on battery I hit 3.75 playing wow

 

You repasted ?

 

Yes some VRMs have paste, others pads

Sent from my BLA-L09 using Tapatalk

 

I didnt know havet taken mine apart yet...can i replace paste with pads?

 

"but given I used Gelid Extreme and not LM"

Yeah but I may repaste LM

 

those who put a new ssd in their machines di u ue thermal pads on them?

 

Decided to take new 3.9Ghz throttlestop and pair it with the 90w 2060 vbios:
CPU hit 80C briefly
GPU stopped at 74C





Just finished OC Scanner, running another bench with updated clocks

 

You mind telling me what thickness of the pads u use i have 0.5mm and 1.0mm available

 

Also my cpu wont go over 40-42w on aida stress test