*** MSI 16L13 (Eurocom Tornado F5)/EVOC 16L-G-1080 15.6" Owner's Lounge ***

I ran my GPU at 1.86GHz and 875mV for some time and wondered, why I got crashes in Warzone and Ghost Recon Breakpoint.
The crashes happened randomly, sometimes after 10 minutes of gaming, sometimes after more than 1 hour - and sometimes not at all.
Initially, I attributed them to buggy games, but when Cyberpunk 2077 came out, the crashes happened more frequently.
For reasons I can't fully grasp, it seems that inefficiently coded games tax the GPU in quite a reliable way to produce crashes.
Cyberpunk 2077 allowed to test in short intervals as it would reliably crash within 45 minutes of starting it.
After adding voltage to reach 900mV, all those games have become stable.
I.e. my undervolt was too aggressive but did not show up in most games or even the 3DMark Stress Test, Unigine Superposition, ...
Just as a FYI, in case you are experiencing games that sometimes crash at random.

 

Thanks for the info.

 

guys someone know if there is some display upgrade like 120hz 4k\2k for it? even not an aftermarket..some update for rtx 5000 ? someone tried it?

 

After doing some additional testing, it turns out that my statement was not precise enough.
I have set up 1.86GHz and 900mV in Afterburner, but for reasons unknown, the GPU will start into any kind of application with one frequency notch lower (I.e. setting 1.86GHz in Afterburner means a maximum of 1.848GHz in applications).
And funnily enough, it clocks one notch lower after about one minute, reaching 1.835GHz and staying there.
So, being the nerdy fiddler that I see myself as, I decided to try undervolting further only to find out that the GPU crashes within 15 minutes in Superposition, even when using 1.848GHz (resulting in 1.835GHz used) at 875mV and 881mV.
The next voltage notch is 893mV which I have not yet tested extensively, but I am not too optimistic.
Consequently, I have got the suspicion that my GPU is actually only stable at 1.835GHz @ 900mV in real world applications...

Long story short:
- set one frequency notch higher than actually planned (in Afterburner)
- monitor the frequency during load after some time to see which frequency the GPU stabilised at

 

And as another small tidbit I have just realized (the usual caveats and disclaimers apply, i.e. you do this at your own risk and although it works without problems on my device, there is a chance that it behaves differently on yours):
ASPM seems to be off for the PCI-E devices on this laptop by default.
I have previously struggled to cool down my PCH, where it would routinely reach uncomfortable temperatures.
As a remedy, I did add a copper heatsink (my pictures and description are in this thread) but the PCH would still heat up to ~85°C under taxing loads (e.g. gaming).
I would have loved to get its temperature down, but did not find any meaningful way - that is until I stumbled across ASPM.

My PCH is now registering a maximum of 68°C after stress testing (1x FireStrike Ultimate Stress Test, 1x FireStrike Extreme Stress Test, 4 hours of OCCT Memory Stress Test), 3 hours of Valheim and a few hours of idling.
All that has changed between the hot PCH and the cooler PCH was enabling ASPM in BIOS for 4 PCI-E devices:
- NVME SSD #1 (Sabrent Rocket 1TB) on Root Port #9
- NVME SSD #2 (Intel 660p 2TB) on Root Port #17
- Intel WiFi (Intel AX210) on Root Port #7
- Killer LAN on Root Port #6

I did use HWInfo to know which PCI-E Root Ports they are connected to and it looks like this:

I would assume that this is the same for each and every Tornado F5, but feel free to check on your end - it's a commonly available freeware after all

To enable ASPM in BIOS, you need to have access to the hidden settings.
There, go via Advanced -> PCH-IO Configuration -> PCI Express Configuration -> PCI Express Root Port #xy -> ASPM
And set ASPM to Auto.
Repeat for each of the 4 devices
As a visual guide, I have tried to summarize the steps in those pictures:

Afterwards, you can check with HWInfo again, as each device you have enabled ASPM for should show the status like this (i.e. something other than disabled):


For devices that don't have ASPM enabled (this example is my card reader), it looks like this:


Hopefully this is of some help to you as well

 

so we will not receive any gpu upgrade soon? what you think guys?

 

As a short update to this: my device is still not showing any adverse effects.
So, because of my curiosty, I decided to enable ASPM for the other Root Ports as well.
While this seems fine and stable as well, there is little additional temperature reduction (if at all) from doing that, which makes sense since those Ports do not contain any devices according to HWInfo.

Just as a warning though: I did previously also test the "DMI Link ASPM Control" which is one level above the Root Ports and this led to instability on my system during longer sessions (i.e. it boots and posts fine, but did BSOD on me in longer gaming sessions).
Consequently, I advise against changing the "DMI Link ASPM Control".