Aurora R8, bricked

---

@DromendeDaris wrote:

Aurora R8 ... with the latest i9 and the RTX2080TI.

1. I've already faced a lot of issues with the horrible Killer network card, but that's another topic.

2. I had the exact same "space invader" screen full of artefacts.  I'm part of the failure rate of these cards, it seems.

3. And it looks like this means my RTX2080TI is dead.

4. I went for the Windows restore point option. This actually worked, the restore point was only half a day old, so I hadn't lost much. I was back to productivity for a few hours,

5. and then faced the exact same GPU crash. 

6. To be more specific: the Alienware logo briefly shows, then the screen goes all black. Auto repair does not launch, no matter how many times I reboot. This machine is now completely bricked. 

 

---

1. Common (for about 10 years now). Glad to hear you found the "work around".

2. I read it was bad video-ram on some. Shame on Nvidia for not catching that or alerting OEMs.

3. I would definitely remove it (and just use the on-board video for now).

4. Almost never works.

5. Please remove it ... now. If willing, try to verify rest of the machine hardware is still working 100%.

6. If BIOS is POSTing, it is not actually "bricked".

If POSTing, press F12, and run ePSA Diagnostics.

If that passes, you can completely restore back to Dell Factory, or just clean-install Windows.

https://www.dell.com/community/Alienware-Desktops/Aurora-R7-M-2-NVMe-bootable-options/m-p/6073081/highlight/true#M3401

Thanks for responding.

1. I didn't find a work-around. I assume you mean the driver-only option? Doesn't work. I've been going back and forth with Killer support for weeks, and we exhausted all options now. So now I have a crazy expensive desktop with failed networking. No help from Dell at all, all they keep doing is asking me to completely reinstall Windows and run diagnostics, none of which solve anything.

2. Yes, read the same thing, and also that they have a new batch based on other chips. Surely I hope a replacement would be of the newer batch. I bought this machine for the next 5-6 years. I need reliability. I've already wasted dozens of hours on this machine and its many problems. It seems every problem with it is my problem, not that of the manufacturer.

5. It's getting late here and tomorrow I have a busy day at work, but may try this tomorrow night. I prefer to first know from Dell what their plan is. New machine? New GPU? Onsite install or self install? Refund? When? I need to know because it takes me a week from clean Windows to my desired software stack (I use it intensely for programming, photography, gaming, design work) so I need to know first whether to spend that time on this machine or not.

1. Correct. On my Aurora-R6, it has the on-board Killer Networks e2400 Gigabit Wired Ethernet. I only use the simple-driver supplied by Microsoft Windows Update. It works fine (long online gaming sessions, VPN access, etc.). The WiFi is Intel, but I don't use it.

2. Unknown. I do know if your video-card is bad, it can drag-down the whole machine.

3. Once the hardware is 100% functional, you do a clean-install. Normally, you only install enough software to stress-test it.

https://www.dell.com/community/Alienware-General-Read-Only/fixed/td-p/5627124

For now, I would suggest do you real work on old-machine or different machine.

Update: I finally got in touch with somebody from Dell.

They refuse to believe me, in that this is a GPU malfunction that requires a replacement. Despite this being a well known problem. They have no "approval" to replace it, nor does the situation qualify for replacement. They only focus on getting the machine to boot again and running diagnostics tools, whilst completely ignoring the underlying hardware problem. None of the diagnostics tools ran so far mention any GPU failure. Unsurprising, as this failure happens randomly during run-time. 

What they will do instead: they will pick it up tomorrow, and then send it to Poland, for advanced diagnostics. I will lose the machine for a whopping 3 weeks. 

Surely I expect them to get the machine in a bootable state again. The big question is if their diagnostics is sane enough to conclude that this GPU is broken. If not, I'd get back the exact same machine and it's just a matter of time before the GPU crashes again. So far, as said, none of the diagnostics tools detect this problem, so I have a bad feeling about it.

This expensive, high-end machine was supposed to be a joy. Instead it's broken, a nightmare, a liability. I cannot even trust to put anything on it as the entire system crashes randomly, and it will always be my problem, not theirs.

It's been a hard lesson finding out what Dell support really means when you need them, but I now know enough. I will never buy Dell again and recommend anybody in my circles to stay away from it.

Some general hardware troubleshooting thoughts...

* Have you powered off, disconnected the power cable and monitor, opened the case, reseated the RTX 2080 Ti in the motherboard slot, then retested?

* While the case is open, I would also check all cabling, ram, etc. to make sure all is firmly connected

* Have you tested all of the RTX 2080 Ti video out ports (DP1.4/HDMI2.0b/USB Type-C) on the monitor? Without knowing the specific monitor model, we cannot know what testing you might be able to do

Thanks for the support. All I can do now is wait what I'll get back in a few weeks from now. If its the exact same PC with nothing replaced, surely it will inevitably crash again in a matter of weeks. And then we will repeat the cycle of shipping this box across Europe for no reason at all. Not in my interest, not in Dell's interest, but we can keep doing this for 2 years or so. I may then extend warranty to keep it going. 

Bitter? Yes very much.

---

@DromendeDaris wrote:

 

The big question is if their diagnostics is sane enough to conclude that this GPU is broken. If not, I'd get back the exact same machine and it's just a matter of time before the GPU crashes again. So far, as said, none of the diagnostics tools detect this problem, so I have a bad feeling about it.

 

---

Correct.

Remember that this machine additionally has an on-board Intel graphics GPU (that can not be removed) as well as the dedicated Nvidia card.

In ePSA (outside of Windows) ... I'm not sure which one gets used/tested.

And in general, ePSA can only test hardware to about 90% . Some flakey hardware only fails inside Windows.

Even on system with a "plainer" Intel CPU (no on-board GPU) ... I've seen them pass ePSA, and even idle Windows desktop. The dedicated GPU card only fails under actual "3D graphics" use where all the GPU-cores engage and the card is actually being used at > 80% utilization.

Your first post said you have had the machine for 6-weeks, so all my answers have been with this in mind. I've been working with Dell and Alienware for many years ... so I had a pretty good idea of the path you were leading down.

Good luck with your repair.

Chris,

The GPU is dead. There's nothing to test. This is a common issue of the founders edition of the RTX2080TI, yet nobody at Dell seems willing to acknowledge it. It's a broken GPU that needs to be replaced, it's as simple as that. No amount of troubleshooting is needed because the trouble is already found, a faulty component. 

Furthermore, as said, it's going to be shipped to Poland soon. It's out of my hands now.

Still appreciate that you responded. 

Hi @DromendeDaris,

Are you getting something in the monitor? No boot and no video are different things. 

If the problem is the discrete video card, then you should get video by connecting the monitor to the on-board card (display-port). Sometimes, the port will be covered with a plastic protector that can be removed to access it. Please try it and let me know if you get an image.

@Alienware-Eimy Thanks for responding.

The machine is already disconnected and prepared for shipping, so I'm not going to troubleshoot anything at this point.

Even if doing that hypothetically, I fail to see the point. I believe you that it may be possible to get this machine working again with onboard video, but what does that bring me? I bought it with the RTX2080TI, which is broken and needs to be replaced. The problem is not getting this machine to boot, I'm quite sure that can be done, the problem is a broken GPU! Dell hasn't committed yet to replacing it, so now I have to wait for weeks what the team in Poland decides. If all they do is the diagnostics scan, surely the problem will not be detected, and I get back the exact same machine. Which then after a few weeks will crash again, as the GPU is broken.

It's this problem, to be clear:

https://www.techspot.com/news/77445-nvidia-addresses-failing-geforce-rtx-2080-ti-cards.html

Surely you must agree that Dell can't place back this broken component into my machine? I expect this machine to last for years, not crash the entire setup and taking along Windows with it every few weeks. It will be an expensive ordeal for both me and Dell as we'll keep shipping broken products back and forth. 

So if you want to help me, tell the people in Poland to replace the GPU with the new batch that doesn't have the problem. The alternative will be even more expensive. Nvidia shipped you a broken product, and you shipped it to me. Surely not intentional, but that's how it is. Return it to Nvidia, get a good copy, and ship it back to me. Solved. Painful? Yes. But a whole lot less painful that endlessly "not repairing" something that is just plain broken.

Hi Everyone,

Also following this thread with interest.

My Aurora R8 arrived a week ago and I am experiencing the same issue. 9900k and 2080ti machine.

Eventually, the issue leads to the machine not being able to boot. I am able to get the machine running by reinstalling Windows and using a USB recovery key, but consistently experiencing artifacting.

Have reached out to Alienware Tech Support and will see what happens.

@Code_Guru  Sorry to hear the same thing happened to you, and thank you for chipping in and sharing your experience. It looks like the exact same issue. Same hardware, same problem.

I too had multiple artefact crashes. A crash may take down all of Windows and sometimes it does not and you can recover. Then it's a matter of time before it crashes fatally and you have to start all over again. 

It seems all help offered so far here as well as by Dell officials is to recover the machine back to a bootable state as if that solves the problem. What are we expected to do exactly? Reinstall Windows every 2 days?

Discussing the real problem, a broken GPU, seems to be avoided. That, above all, is my main frustration. This problem should be addressed between Dell and Nividia, not between Dell and its customers.

Anyway, I hope you get better support than I do, because for me it certainly could not get any worse. I posted a week ago that my machine would be shipped to Poland for diagnostics. Guess what? It's still in my home.

After a call with Dell where they mentioned to pick it up the next day, I offered to put it in the original box, which I still have. No, they would bring their own box. Fine. Then they called that they had to ship an empty box from Germany to the Netherlands. It took close to a week for this box to arrive. It's just a cardboard box with foam panels. Today this box arrived, and tomorrow they will finally pick it up. So another week lost for no reason at all.

So on top of dozens of lost hours in productivity for troubleshooting this machine, I also have to negotiate with my employer to stay home for 2 days to take care of receiving and sending a freaking cardbord box. Which takes a week to ship over 200km. And now I'm facing another 3 weeks without this machine where I have to work around this lost productivity with incapable hardware, whilst not even sure if the problem will actually be addressed or they just ship it back with Windows working, yet the faulty GPU still not replaced.

It feels like a customer-last experience, its humiliating and infuriating. As said, I hope you get better support, as nobody deserves this treatment after such a great expense. Hope you keep us updated on what happens next.

@DromendeDaris - Very sorry to hear about your experience. My own experience with premium support today was much more positive. I have on-site service in my Alienware warranty in Ireland, so I will keep you (and the thread) updated on the outcome. 

Yes, I would be in agreement with your diagnosis. Gamers Nexus, who take their testing pretty seriously, did a number of videos on the 2080ti showing that the operating system, driver, temperatures, clock speeds, etc. have little impact on the artifacting. The primary issue seems to be on the manufacturing or quality control side before leaving the Nvidia (or partner board) factory.

https://www.youtube.com/watch?v=JIRfPlC15uc

I was running fine for the first five days before it started happening. I tried to recreate various test environments over the weekend to see if I can identify the root cause, but no real conclusions for now.

Sometimes, the machine will crash almost immediately on a relatively low load task (e.g. having a few browser tabs open), other times it might run for a few minutes, or even a few hours on higher loads.

Some crashes can be solved with a simple reboot, whereas other crashes seem to corrupt the entire operating system, requiring a System Restore, or even a full clean install via USB.

For example, Metro Exodus crashed immediately yesterday, whereas I was able to play about two hours of Battlefield V without any issues. I launched The Division 2 this afternoon and it crashed immediately, but then I relaunched it on the hard drive instead of the SSD and I was able to play at high settings without issues.

**bleep** May Cry 5 got about 20minutes before crashing (on the SSD) and requiring a full reinstall. 3D Mark TimeSpy Benchmark crashed within about 3minutes, whereas the Unigine Heaven Benchmark runs fine.

I might do a bit more analysis to see if this is an SSD vs. HDD issue, as I do seem to be able to get a bit more functionality on the HDD, but it's clearly not working regardless.

Happy to give Alienware a chance to replace the card, as I can see how it would have passed an initial burn-in or installation test. From reading other user experiences with the 2080ti, many other cards seem to degrade after a few days or few weeks. It's not as simple as 'Dead on Arrival' so hard to pick up when shipping a new machine.

---

@Code_Guru wrote:

The primary issue seems to be on the manufacturing or quality control side before leaving the Nvidia (or partner board) factory.

Thanks for the YT video.

"Manufacturing level issue" is my guess as well. I'll even go so-far as to suggest it's the lead-free solder (in general) or failure of proper BGA soldering with it.

I don't see how it could be a "chip silicon" problem if most cards work properly. However, only Nvidia knows the build/revision of each chip.

@Code_Guru  Thanks for the very useful information. Good for you that you seem to get some serious support. I suppose I am born in the wrong country to receive the same. 

That video is very useful, and it is "comforting" to hear that this is not some imagined problem that can be solved with a driver, that instead it is a true hardware defect. The load you put on it, the firmware, even the operating system...none of these factors matter. 

Like you, I agree that the video ended with a wrong conclusion that this could be detected via a simple "turn-on" test at Nvidia's manufacturing. That's not how the problem manifests at all.

For me, it happened after about 6 weeks, I had zero GPU issues before. The card even announced its death via max fan bursts that produce a deafening noise, randomly, in Windows. The day after was the first artefact crash. After a successful Windows recovery via a restore point, it then crashed again after about 4 hours. Since then, I could not boot at all anymore. I'm sure with the onboard video option it could be brought back to a working state, but that obviously does not solve my problem.  

I can only concur with the behaviour you're seeing. I had artefacts crashes both in Window under low load and in a game under heavy load (BF1). Furthermore, I had soft crashed where I could reboot, and hard crashed taking down all of Windows. 

This behaviour is exactly why I'm so concerned with the next step. It's going to be diagnosed in Poland. I worry they get the system bootable again, run some diagnostics, conclude it's fine, and send it back. As said earlier, Dell support in the Netherlands said I did not qualify for replacement, so I consider this scenario of sending me back this mess to be possible, if not likely.

With such clear evidence that this is a bricked card, it's maddening that Dell/Alienware goes into hiding and in no way claims responsibility. Due to the way in which the problem manifests itself, we can establish that it was hard if not impossible to prevent. It takes days or weeks to surface, and manifests itself randomly and inconsistently.

Fine. But now that we know this, how about somebody from Dell/Alienware reading this steps up and takes care of the replacement? It's going to have to be replaced anyway, so there's no "win" in not granting a replacement. It just means I will have to keep sending it back, as this card will keep crashing like every known case of this issue. 

I'm going to write to that guy of the video, as maybe my story is of some use to him.

Thanks once again for sharing your misery, I'm genuinely happy for you that you do seem to get serious support.