G752VS Crashing Issues (shuts off and restarts automatically)

[Hapablap]

Hey,

I've had my g752vs for a few months now and recently I've been running into a problem with it crashing while playing some games.
I uninstalled ROG Gaming Center and that seems to have reduced the frequency of the crashes but I'm still experiencing some.

Currently I'm playing DOOM and my laptop seems to crash randomly. The computer simply shuts off and restarts automatically. I've been unable to find any logs about the crash in my event viewer.
I've set up some logs for my GPU and CPU and temperatures at the time of crashing looks like this:

Seconds GPU HD0 HD1 Temp1 Core 0 Core 1 Core 2 Core 3
82356 77.0 36.0 32.0 68.0 67.0 69.0 61.0 61.0
82359 77.0 36.0 32.0 68.0 71.0 67.0 64.0 63.0
82362 76.0 36.0 32.0 68.0 66.0 68.0 64.0 61.0
82365 0.0 36.0 32.0 70.0 69.0 62.0 64.0 60.0

For the last two crashes the GPU temp hits 0 right as it crashed (perhaps indicating that this is the piece that is failing/shutting off to avoid overheating).
Has anyone been experiencing the same?

[Captgeech21]

Hey all,

I'd like to not only add my 2 cents and attempts/progress/lack there of as well as gain feedback. I have sent a PM to Bahz to ask what to do in my situation, but I would truly like to get input from anyone still on this thread.

I purchased my ROG G752VS-RB71 in February of 2017 from ExbaliburPC.com. They did a "free" ram upgrade to 32gb and I added a SSD on top of the standard HDD.

Everything seemed fine with this laptop for the first few months. After installing Elder Scrolls Online, I started to get the crashes we are all talking about. White screen, some audio, reboot and then a Kernal 41 stating critical power failure. As with many, my temperatures in Gaming Center for video card never went above 60-65 for CPU or GPU. As I have been building computers for quite some time, my immediate thoughts were heat, temps etc. causing restart, but as with many, there isn't much to support this. Also, around the time I started getting the ESO crashes, I had upgraded Nvidia drivers (very very close together).

Since then, it has started to occur at times in other games. The worst games are ESO and Destiny 2. I may play for 6 hours with no problems or get 5 crashes in a half hour. Like I said, there were no crashes the first couple months I had the laptop, and I played Elite Dangerous with no issues. After reading this thread, I tried Elite Dang. again and I crash every single time I enter a training immediately.

What I have tried thus far has been rolled back Nvidia drivers, used DDU in safe mode to remove Nvidia and G-experience and reinstalled Newest drivers.

I have reinstalled Windows, I have changed bios settings, I have updated bios to 310 from 302 (i believe cant remember).

After reading the thread, I just removed ASUS gaming center and used GPU z to monitor voltages/power settings etc and it continues to occur.

Some games seem worse than others, and like I said I can stress this thing out with no issues or even game hard for a while with no issues and then BAM.

I am guessing I have 12 months ASUS manufacture warranty, which would mean i'd need to RMA it soon. I am not sure what their policy is on adding ram, SSD etc and whether it's covered still or if i would need to remove them before sending in.

Something I am very interested in getting feedback on is this:

After trying GPU Z and seeing the stats, I see the ram is GDDR5 (Micron). Now on a MSI thread I found a while back, I found that there was Vbios update for MSI 1070 nvidia cards due to the micron ram having issues versus other cards that have Samsung memory modules. The issue in summation is that the voltage changes due to gaming and such wreck havoc on the Micron ram and cause the crashes. Apparently the vbios update for the MSI cards/laptops worked well to resolve this but i have been unable to find a safe or descript vbios to try.

Any feedback would be appreciated because i've been dealing with these crashes for about 6-7 months now and am running out of patience and time.

Thank you all and good luck!

[Captgeech21]

Okay,

So I wanted to make an update to my above post and figured a separate post would be easiest to differentiate.

After uninstalling ROG Gaming Center and using GPU-Z to monitor frequencies of the GPU and GPU Memory, Elite Dangerous crashed every time I went into a specific training mission.

It is crazy to see how much the frequency of the memory and even in some cases the GPU spikes in Gaming Center.

Then I re-read the Gaming Center deletion posts and realized someone mentioned adding After burner or even gaming center back in and setting a "standard" preset instead of the extreme or overclocked setting.

With "Standard" on Gaming Center, the crashes for that mission have stopped. Even played ED for a few hours with no issues.

While this is a great step in the right direction, it is frustrating to realize that my GPU can't really be put to it's full potential (not even accounting for overclocking) without a factor of extreme instability.

The more I have explored the frequency side of this equation, the more this is looking to be identical to the MSI GTX 1070 threads I have found discussing faulty Micron memory on the 1070 GTX cards.

I do not believe this is an issue as believed before by others and by ASUS to be a "mainboard" issue. I am guessing that due to the G752 having a dedicated, but soldered on GPU 1070GTX, that replacing the motherboard has also replaced the faulty Micron graphics card memory. And that is why people are seeing results when RMA'ing back to ASUS for a swap. I have also read only a fraction of responses that an RMA did not work, even with a PCB/mobo replacement. Maybe they are replacing with another faulty Micron memory board?

Now if PCB is not the motherboard, and the motherboard does not actually house the GPU (i.e they are able to install a new mainboard OR motherboard and for whatever reason are able to keep the previous 1070GTX module) and the issue is going away for people, there could be an alternative answer or even simultaneous issues that are causing this problem.

Sorry if that doesn't quite make sense. My understanding of dedicated cards on laptops for the 10 series is limited. I know that prior the "Dedicated" cards were still soldered on. However, with the 10 series, I know it's essential a desktop card (power wise) or close to it, so I am not sure if they can remove that card in some way while swapping out boards.

If indeed replacing the PCB is also in turn replacing the GPU, then I think the problem with the Micron memory is exactly what is causing all of these issues.

In the MSI threads, they stated MSI was swapping out the cards with non faulty micron or Samsung memory module cards. Or that you could flash the GPU bios (vbios) with an 8A version that supposedly suppressed the clock frequency spikes past a certain point. The replacement would definitely be the better option, as it is similar to not allowing Gaming Center to upclock the GPU or memory and hence, you will see slightly diminished performance.

Some other info.: To test the performance of my GPU quickly without having to run many a hour of bench marks, I mined some Ethereum 😃 With Gaming Center on Standard the GPU mines 25-26 Mh/s. With it on extreme, it mines 26-27. So not a huge difference, but definitely kind of a bummer.

If anyone has any thoughts on updating bios for GPU or if they have found a new bios to fix the issue versus just suppress it, please, please let me know!

Thank you all

[WinterElf]

Captgeech21 wrote:
Okay,

So I wanted to make an update to my above post and figured a separate post would be easiest to differentiate.

After uninstalling ROG Gaming Center and using GPU-Z to monitor frequencies of the GPU and GPU Memory, Elite Dangerous crashed every time I went into a specific training mission.

It is crazy to see how much the frequency of the memory and even in some cases the GPU spikes in Gaming Center.

Then I re-read the Gaming Center deletion posts and realized someone mentioned adding After burner or even gaming center back in and setting a "standard" preset instead of the extreme or overclocked setting.

...

Yeah I had the exact same thoughts and saw the same things about MSI forum and their custom BIOS (check my previous answer page 18).

Would be nice that Asus do actually something, even if it means a Bios that allow for less overclock, but at least a all the time stable system.

[danmaku]

I replaced my laptop. Played Arkham Knight again a bunch and it didn't shut down, or hasn't yet at least.

However, I now have the problem that some other laptops have where the audio port heats up for no reason

[Dreamonic]

Regarding recent posts in this thread, especially for the RB71 models that have the 6700HQ CPU specifically, the Extreme profile in the Gaming Center has no importance like it does with the 6820HK models (XB78K, XB72K, RB71K, Q72S). The Gaming Center is cancerous for overclocking, and if you feel the same way about XTU, you can then OC at the BIOS level if you have the 6820HK that supports it. That aside, if new installation of Windows and basic notebook drivers installed followed with latest BIOS is still giving you the same issues with system stability in games, you could also try flashing the GL502VS 1070 VBIOS over the G752VS 1070 VBIOS and test further. Or perhaps use MSI AB with the Curve Editor and find a feasible config. If it's happening while completely under stock/factory conditions still being related to the GPU Micron memory ICs, then a VBIOS update should suffice. Post a screenshot of your VBIOS version to see if there has been any newer ones out since. That or cross flash different vendor ROMs to see if any others yield improved stability over another. More on this later.

Current versions: 86.04.2A.00.0A (G752VS - original), 86.04.2A.00.3F (GL502VS - original), 86.04.42.00.07 (GL502VS - updated)

Now while thermals might be low and have been checked off as not being the problem (perhaps it's not GPU related), have you also checked your microcode version? As Skylake CPUs have had a HT bug that would cause a similar scenario most of you are reporting when specific conditions are met. Perhaps disable HT and repeat your tests again. IIRC, the version you want is BA should you use UBU to do it.

You also want to make sure you don't have SST enabled, which is HWP (Hardware P-States) in the BIOS under Advanced > CPU Configuration (once unlocked). This is a huge contributor to stability problems with the G752VS notebooks, as at the BIOS level it's enabled, meaning frequency/thermal/power package control is done internally. In the older BIOS versions, this was disabled by default. I have had people screenshot their HWiNFO "features" window with SST showing as enabled. Once I told them to disable this option there instability problems disappeared.

On that note, there is also something else most G752VS owners don't know, speaking more toward the 6820HK owners, and that is if you unlock additional BIOS options with AMIBCP v5.0.1, you can disable the ME State Control under the PCH-FW menu. This will allow full CPU VR and Power package control, exceeding 100W TDP @ 4.5-4.7Ghz without being limited so as long as you repaste with LM, as thermals are very hard to control beyond 1.35v if not using LM.

Also worth mentioning, I have tested multiple (10DE:1BE1) 1070 VBIOS ROMs out like the Clevo and MSI (using NVFlash v5.292.0). As mentioned above, also the GL502VS 1070 VBIOS which had no stability problems corrected the pink screen at startup for me a long while back. What's worth mentioning here is that if you decide to do the TDP mod with a HW programmer to the 1070 in your G752VS, expect power delivery problems once you exceed 150W for a short period of time (130% TDP from 115W). As I originally did the shunt mod on my G752VS 1070 VRM a long time ago and while I had impressive performance gains, due to the way the EC manages board/battery power, the system would get massive spikes (GPU usage dropping to 0%) like the AC adapter was being repeatably unplugged; battery in a constant charging state.

So in short, most of the ROG Notebooks I've had in my possession all required full disassembly, intake mods, repasting, adjusting heatsink assemblies, standoff mods, custom heat plates, max fan states, BIOS/VBIOS mods just to get them all at a comfortable performance level where thermals are effectively controlled. Such a shame these notebooks require so much work to make them decent enough for their cost. Tomayto, tomahto.


Oh yeah,

All models after the G750 series are of BGA package for both CPU and GPU. So if you send your notebook in for repair (RMA through ASUS), you will likely get a refurbished one in return. Meaning if you had a good quality CPU or GPU previously (could OC well with lower voltage), it is likely going to be a silicon lottery again when you get your new MB, as everything is replaced with the MB short of the NVMe SSDs. They do it like this for a quicker turn around time to you (the customer) while they send their serviceable parts to another department for repair. Likely the actual issue is not found or not escalated to a higher rep who then sends it to another facility. So faulty components are likely to keep circulating as "passed" refurbished notebooks through RMA since as I've seen in the past, their QC for stability tests is literally as simple as booting up and shutting down to be stamped as a "pass" of approval. No synthetic testing or specific load tests are done to determine if related to CPU or GPU components. That is another department (usually outsourced) if the refurbished board comes back again.

[ForgeOfGods-G75]

6 month after rma (replaced mb) confirm that crashing still persist, not so often now but still pc isnt stable sofar. Crash changed, now ita not crash but freeze, 2-3 times a week maybe, doesnt matter what i do, playing, wathing movie or sleeping, its randomely freeze up, none errors just a kernel power error after i reset it, thats annoying af , only one good part that its rare freeze.

[Captgeech21]

@Dreamonic,

These are some valid points and good tips/info.

The extent to which you explained your efforts on multiple models is exactly why I inquired to Bahz whether i'd get my SSD and ram upgrades transfered, back, or need to take them out. If it's going to continue as an issue for me, and require substantial tweaking of VBIOS, BIOS unlocking etc, i'd rather know I can send it in and get a fixed or new unit. It's sad that you are stating ASUS has such poor technical RMA procedures (At least on a general scale). After working on Geeksquad for years, we had quite expensive and exhaustive programs that any computer displaying any instability symptoms would undergo at a minimum. It doesn't take too long (a couple days with very minimal human interaction) and it's great to pinpoint down problems.

I would definitely have to get a confirmation from a higher tier rep on the phone that i'd either get back my laptop repaired or a new laptop, not a refurbished unit. Never liked that idea with my phones either. Can't tell you how many Palm PRE's or HTCs ive gotten refurbs back on with much lower quality specs or performance.

As another update. With standard on Gaming Center, I have not been able to cause a crash yet. I think I will go heavy on some ESO and Destiny for a while and see if I get anything. Im not willing to accept a limit through VBIOS on my very expensive graphics card as a solution, unless it's literally my only option for stability.

We shall see.

[Captgeech21]

Guess it's time to start looking at vbios flashing. Just had a crash while playing ED even with "standard" ROG. =(

[Captgeech21]

Btw, my Nvidia bios version is 86.4.2a.0.a
Going to see if i can find any gtx 1070 for the g752 series bios updates out there.


@Dreamonic

I hope my earlier responses don't come across as ignoring some of the other feedback and info you gave above. I am indeed trying to look into all avenues you explained to look into before I do an RMA. This whole situation is just very frustrating, and i've been dealing with it for around 6 months now.

Anyhow, I am looking into vbios swaps. As I said before, I have flashed mobo bios many a time, but never done vbios (never had to). Does swapping to another model laptop with a 1070 pose any risk? Mainly, can I pick a wrong model with a 1070? Or are all of the 1070 GPU going to be about the same? My understanding is that even the 1080 (at least desktop versions) are the same GPU with different firmware/bios to allow for more shader use or what not. So I am under the impression as long as its a 1070 mobile bios, it should be ok to use. I just don't want to hinder my card, or heaven forbid screw up the GPU.

Also, you mentioned unlocking advanced bios to look at some CPU config. Is this possible on the RB71 with the 6700HQ? I have found the G752VY but not the VS nor the RB71 specifically yet.

Any tips would be appreciated.

As per the issue with Gaming Center. The reason I was hopeful for standard settings over extreme is that even in a software suite as pitiful as Gaming Center, Extreme mode does show an increase in clock speed, memory speed and boost speed. It's not those set speed increases that are causing the crashes. But per the logs, every crash shows the GPU at 1847 for the 4-8 seconds before the lockup and restart. This was not happening when I set the Standard option, but for some reason, standard in Gaming Center isn't keeping the clock quite low enough and I have had a few crashes.

I am going to grab MSI afterburner and play with it to see if I can at least get a stable solution for now.

[nulmas]

@Captgeech21

I've been having the same problem, and I already RMA'd my laptop twice. I'm currently trying to get a replacement or my money back.

The laptop crashed with most games, but it was quite random. WWE2k18, though, turned out to be great to test if the issue had been resolved.

I did find something, though: if you use MSI Afterburner to underclock the GPU by quite a bit, the crashes stop. Of course, by doing that I wouldn't be getting what I payed for.

I also think the issue is with the Micron memory in the cards. It seems consistent with the issues I've seen mentioned.