17 R4, graphic crashing

Hello,

I have the same problem since 1 week.

I have all drivers up to date and reinstall Windows but this problem is still there.

I can work and look at videos but when I play games, my computer freeze.

I launched the SupportAssist to scan the hardware and it freeze too when checking the graphic card (GeForce 1080). So it should be the problem. I think it's due to a recent update (less than 2 weeks) but i can't find witch one.

I tryed previous drivers to my graphic card but the problem is still there.

Does anyone knows what the problem is?

P.S: Sorry for the faults, I'm French and not used to write in English.

General Information

After few days of active research, i have found this solution witch works for me:

https://answers.microsoft.com/en-us/windows/forum/all/video-dxgkrnl-fatal-error/463585b5-3fe3-4044-82bb-35d946af8367

The default windows driver (date 07/12/2019) resolved my problem.

I hope it can help someone else.

Hi Stuart_N

Did you manage to solve this issue?

I have the exact same issue on the exact same computer (Alienware 17 R4, GPU is GeForce 1080 8.00GB, CPU Intel Core i7-7820HK, 16GB RAM) since few days, after a BIOS update to 1.9.0 (not sure of what was my BIOS version before the update).

As soon as the load is increasing, with CPU temperatures about 85°C and GPU Temperature about 70°C, the same crash as you describe happens : freezes, and buggy lines / squares blinking randomly, random colored bars at random patches on the screen...

The problem happens exactly the same way and same conditions on Windows and on Ubuntu.

With Furmark + a CPU Stress test, less than a minute is needed to get the crash.

This cannot be an excessive temperature issue (excessive temperature shutdown would have been about 100°C, but here the crash happens below 90°C), but it seems to be linked with some kind of instability when getting warm.

By experience, this generally is caused by insufficient core voltage, but adding 100mV in the BIOS for turbo boost makes no effect (CPU ID HWMonitor tells no voltage difference, and crash still happens)

We need the help of DELL on this subject.

Hi all,

Apologies for not replying with any updates regarding this issue, for people searching for similar issues.

After contacting Dell support, I was advised to send my system to the repair depot.

After doing so, they advised me the issue was motherboard related. A certain section of the motherboard had short circuit and they were able to replace the affected section without replacing the whole board.

However, after receiving my system back, I was having issues with some software, stating "No Board Found" or something similar (can't remember exactly). So I sent it back again (covered under repair warranty) and they are currently in the process of replacing the whole motherboard (at least that's what I gather from their communications).

Since I am not familiar with computer repairs myself (I only know the very basics), I would advise anyone else having an issue similar or exactly the same as this to set up a service request and have the Dell technicians take a look at the system.

The time frame was very reasonable (3 weeks from sending my system to receiving it back) and the cost was also reasonable (around the £200 margin).

Sorry I cannot provide a fix, my computer knowledge only covers basics, but this is what I had to do to resolve my particular issue.

Hi, no problem and thank you very much for your fast answer!

As I'm well trained about laptop repairs and electronics, the encountered issue with my laptop (same as yours), which is instability when getting warm (just after a BIOS update in my case) is most likely to be something like an incorrect hardware voltage (a voltage which is too low causes more troubles when getting hot, while running perfectly fine when cold).

BGA soldering issues/cracks would be causing troubles even when cold, so this can be ruled out.

Short circuit could be ruled out in our case, as there is no over power consumption, and no abnormal hot points with thermal camera. And every part of the computer is working fine (apart from this crash when loaded and hot). The computer is successfully passing all tests from BIOS.

BIOS voltage adjustments may be really tight and not working very well on all laptops considering variability of CPU and GPU manufacturing for example : DELL optimizations about those voltages in new BIOS versions may have been pushed a little bit too far.

In any case in my case, the issue has been triggered by BIOS update. This is why I'm would be interested in having DELL looking at this issue, potentially with a BIOS update (even with a beta) before risking to pay a new motherboard in case the issue may be solved by a different method.

Thank again for you answer, do not hesitate to keep us informed

Looking forward for Dell feedbacks on this topic!

I successfully solved the problem. I'll need more time to ensure that crash aren't coming back in a few days, so that I can give you a stronger confirmation.

It was related to a bad clock setting about the Intel GPU, which was fixed to "1100 MHz" while the default value was supposed to be... "default". For Intel HD 630, "default" seems to mean Dynamic, 350 MHz to 1100 MHz depending on usage and... temperatures, of course.

So we put it back to default, restarted the computer, and we are now unable to crash the computer. Which is good news!

I used Intel(R) Extreme Tuning Utility to do so. It's persistent between boots, even if you boot another OS.

The fact that the crash were graphics glitches on the screen was a pretty obvious indicator of an Intel GPU misconfiguration rather than a CPU thing (as screens are connected to the Intel GPU). It seems that getting the GPU forced to 1100 MHz wasn't a good idea, probably because this GPU is not working fine at 1100MHz when getting hot (or it would probably need an higher voltage to keep running fine when getting hot).

Dell, if confirmed in the following days, could it be possible to release a new BIOS fixing this issue for everybody?

Hi

Just wondering what the outcome was in this thread. Understand from several forums that we're not the only ones having problems similar to these. Anything from Dell?? Other than swapping motherboard (which obviously doesn't work).

Would it be possible to get a more detailed run-through of how you used the "Intel(R) Extreme Tuning Utility"? And more importantly, has the problem returned??

Hi, sorry for the delay (holidays etc)

Some new information : the problem recently came back and quickly became stronger : after few times the computer was sometimes unable to start completely or at all, depending on we weren't sure exactly what.

The Dell auto test then signaled a memory problem, so we cleaned the memory modules contacts and put them back in place, and the problem disappeared again... !

Is this related to the initial issue? It's not sure but not impossible : Just before Intel Extreme Tuning Utility was used, we disassembled the computer to check again everything on the main board (without finding anything). So in case of weak contacts on memory modules, the fact of having it disassembled could have moved contacts a little bit... but it's strange, as memory modules were already removed / placed back before with no effect (first time we did a thermal paste replacement with no change).

So it looks like this problem can also be caused by weak contacts on RAM modules. Cleaning them seemed to solve the issue for now... if something new happen, we'll keep you informed!

Great job!

Same machine, same problem. The GPU around 70 degree froze and shut the system off.

i didn't fix the problem but what i did let me played  my games.

what i did is retored the system BIOS to the factory one as described: https://www.dell.com/support/article/en-ca/sln114853/alienware-system-overheats-or-shuts-down?lang=en

and hadn't face the problem yet. clearly there was someting in the bios update tha messed things up.

Hope it helps out someone.

Ayy hold up, you saying 70c?

My 17 R4 with a GTX 1080 OC edition wont reach 70c because I replaced the thermal paste so I would advise you do that soon because a GPUs maximum temp is 70c.