I didn’t manually set the timings in the bios. I’m not overlocking the RAM.
When I had one 4K screen, this issue occurred before.
An update. The RAM usage on my computer just exceeded 50%, and then the problem occurred again. I haven’t rebooted my computer.
Memory usage drops below 50%, but the issue exists. I guess this is not related to RAM.
I checked nvidia-smi, assuming there might be some leads.
I notice when pacman shows correctly, the VRAM usage is 28MB.
Tue Nov 7 08:18:01 2023
+---------------------------------------------------------------------------------------+
| NVIDIA-SMI 535.129.03 Driver Version: 535.129.03 CUDA Version: 12.2 |
|-----------------------------------------+----------------------+----------------------+
| GPU Name Persistence-M | Bus-Id Disp.A | Volatile Uncorr. ECC |
| Fan Temp Perf Pwr:Usage/Cap | Memory-Usage | GPU-Util Compute M. |
| | | MIG M. |
|=========================================+======================+======================|
| 0 NVIDIA GeForce GTX 1060 6GB Off | 00000000:01:00.0 On | N/A |
| 49% 49C P0 24W / 120W | 4478MiB / 6144MiB | 6% Default |
| | | N/A |
+-----------------------------------------+----------------------+----------------------+
+---------------------------------------------------------------------------------------+
| Processes: |
| GPU GI CI PID Type Process name GPU Memory |
| ID ID Usage |
|=======================================================================================|
| 0 N/A N/A 799 G /usr/lib/Xorg 2103MiB |
| 0 N/A N/A 1835 G /usr/bin/kwalletd5 1MiB |
| 0 N/A N/A 2053 G /usr/bin/ksmserver 1MiB |
| 0 N/A N/A 2077 G /usr/bin/kded5 1MiB |
| 0 N/A N/A 2079 G /usr/bin/kwin_x11 611MiB |
| 0 N/A N/A 2222 G /usr/bin/plasmashell 84MiB |
| 0 N/A N/A 2380 G ...b/polkit-kde-authentication-agent-1 1MiB |
| 0 N/A N/A 2382 G /usr/lib/xdg-desktop-portal-kde 65MiB |
| 0 N/A N/A 2622 G /usr/bin/msm_kde_notifier 1MiB |
| 0 N/A N/A 2628 G /usr/lib/kdeconnectd 1MiB |
| 0 N/A N/A 2663 G /usr/bin/kaccess 1MiB |
| 0 N/A N/A 2674 G /usr/bin/pamac-tray-plasma 1MiB |
| 0 N/A N/A 2705 G dotnet 14MiB |
| 0 N/A N/A 4584 G ...nyin/files/bin/sogoupinyin-watchdog 1MiB |
| 0 N/A N/A 4635 G ...inyin/files/bin/sogoupinyin-service 1MiB |
| 0 N/A N/A 5102 G ...sion,SpareRendererForSitePerProcess 658MiB |
| 0 N/A N/A 5649 G /proc/self/exe 26MiB |
| 0 N/A N/A 7671 G /opt/google/chrome/chrome 1MiB |
| 0 N/A N/A 7723 G ...799934969,575060744193374022,262144 322MiB |
| 0 N/A N/A 8068 G ...bin/plasma-browser-integration-host 1MiB |
| 0 N/A N/A 139772 G /usr/bin/dolphin 1MiB |
| 0 N/A N/A 208523 G ...ures=SpareRendererForSitePerProcess 413MiB |
| 0 N/A N/A 230879 G /usr/bin/flameshot 1MiB |
| 0 N/A N/A 252955 G /usr/bin/rustdesk 9MiB |
| 0 N/A N/A 273134 G /usr/bin/konsole 1MiB |
| 0 N/A N/A 612359 G /usr/bin/rustdesk 19MiB |
| 0 N/A N/A 1046775 G /usr/bin/dolphin 1MiB |
| 0 N/A N/A 1234188 G /usr/bin/dolphin 1MiB |
| 0 N/A N/A 1286570 G /usr/bin/dolphin 1MiB |
| 0 N/A N/A 1486337 G /usr/bin/konsole 1MiB |
| 0 N/A N/A 1491134 G /usr/bin/pamac-manager 28MiB |
| 0 N/A N/A 3267370 G /usr/bin/dolphin 1MiB |
+---------------------------------------------------------------------------------------+
The usage drops to 8MB when the window is black.
Tue Nov 7 08:18:07 2023
+---------------------------------------------------------------------------------------+
| NVIDIA-SMI 535.129.03 Driver Version: 535.129.03 CUDA Version: 12.2 |
|-----------------------------------------+----------------------+----------------------+
| GPU Name Persistence-M | Bus-Id Disp.A | Volatile Uncorr. ECC |
| Fan Temp Perf Pwr:Usage/Cap | Memory-Usage | GPU-Util Compute M. |
| | | MIG M. |
|=========================================+======================+======================|
| 0 NVIDIA GeForce GTX 1060 6GB Off | 00000000:01:00.0 On | N/A |
| 49% 49C P0 24W / 120W | 4466MiB / 6144MiB | 13% Default |
| | | N/A |
+-----------------------------------------+----------------------+----------------------+
+---------------------------------------------------------------------------------------+
| Processes: |
| GPU GI CI PID Type Process name GPU Memory |
| ID ID Usage |
|=======================================================================================|
| 0 N/A N/A 799 G /usr/lib/Xorg 2103MiB |
| 0 N/A N/A 1835 G /usr/bin/kwalletd5 1MiB |
| 0 N/A N/A 2053 G /usr/bin/ksmserver 1MiB |
| 0 N/A N/A 2077 G /usr/bin/kded5 1MiB |
| 0 N/A N/A 2079 G /usr/bin/kwin_x11 611MiB |
| 0 N/A N/A 2222 G /usr/bin/plasmashell 84MiB |
| 0 N/A N/A 2380 G ...b/polkit-kde-authentication-agent-1 1MiB |
| 0 N/A N/A 2382 G /usr/lib/xdg-desktop-portal-kde 65MiB |
| 0 N/A N/A 2622 G /usr/bin/msm_kde_notifier 1MiB |
| 0 N/A N/A 2628 G /usr/lib/kdeconnectd 1MiB |
| 0 N/A N/A 2663 G /usr/bin/kaccess 1MiB |
| 0 N/A N/A 2674 G /usr/bin/pamac-tray-plasma 1MiB |
| 0 N/A N/A 2705 G dotnet 14MiB |
| 0 N/A N/A 4584 G ...nyin/files/bin/sogoupinyin-watchdog 1MiB |
| 0 N/A N/A 4635 G ...inyin/files/bin/sogoupinyin-service 1MiB |
| 0 N/A N/A 5102 G ...sion,SpareRendererForSitePerProcess 658MiB |
| 0 N/A N/A 5649 G /proc/self/exe 26MiB |
| 0 N/A N/A 7671 G /opt/google/chrome/chrome 1MiB |
| 0 N/A N/A 7723 G ...799934969,575060744193374022,262144 330MiB |
| 0 N/A N/A 8068 G ...bin/plasma-browser-integration-host 1MiB |
| 0 N/A N/A 139772 G /usr/bin/dolphin 1MiB |
| 0 N/A N/A 208523 G ...ures=SpareRendererForSitePerProcess 413MiB |
| 0 N/A N/A 230879 G /usr/bin/flameshot 1MiB |
| 0 N/A N/A 252955 G /usr/bin/rustdesk 9MiB |
| 0 N/A N/A 273134 G /usr/bin/konsole 1MiB |
| 0 N/A N/A 612359 G /usr/bin/rustdesk 19MiB |
| 0 N/A N/A 1046775 G /usr/bin/dolphin 1MiB |
| 0 N/A N/A 1234188 G /usr/bin/dolphin 1MiB |
| 0 N/A N/A 1286570 G /usr/bin/dolphin 1MiB |
| 0 N/A N/A 1486337 G /usr/bin/konsole 1MiB |
| 0 N/A N/A 1491134 G /usr/bin/pamac-manager 8MiB |
| 0 N/A N/A 3267370 G /usr/bin/dolphin 1MiB |
+---------------------------------------------------------------------------------------+
I don’t know if this info helps. I tried to lower my VRAM usage to under 50%, but the issue still exists.
BTW, minimizing and then restoring the window will sometimes make the window return to normal.
Since the new stable update has been released. I’m going to update the system and do a restart. Perhaps do the memtest as well. I will keep this post updated to avoid creating a new topic.
Um, no, not with commodity hardware. ECC registered memory does exist, but it’s very expensive and commonly only reserved for heavy-duty workstations and servers. As the matter of fact, Linus Torvalds has still recently held a plea for putting ECC memory in regular consumer-grade hardware as well.
However, given what you report, it would appear to be a problem with either the memory controller, the GPU or the power supply, and if your machine is that old, this is quite possible. But then again the symptoms would be more severe — a steady buildup of the CPU load followed by a full-on freeze, for instance. So I doubt whether it would be the power supply or the GPU.
I think @bedna is probably closest to the mark, regarding the memory timings. 
I recon there’s something fishy with one of the two modules. Either
or there’s something off
in the different slots.
Before trying anything else I’d update the bios, your v. 1.0.12 from 2018 is 4 years out of date, latest is 1.1.26 from October '22.
3 Likes
You are misinformed.
Memory timings has nothing to do with overclocking (they can be USED to overclock to get most out of overclocking). Every motherboard has been tested with a bunch of memory (by law might I add), you should ALWAYS read this list and get the memory that has been tested on your specific motherboard.
Other memories can be used, but:
- They are NEVER guaranteed to work well (stuff like you are experiencing might happen f.ex)
- The timings STILL has to be set to fit your motherboard.
In most cases, if you get supported memory modules, the timings will be set automatically, but if not supported, you have to make sure they fit each other.
Check your motherboard manual (or website) and look for “supported memory list”,
Edit
A quick search for your mobo gave me these memories.
But first update your mb firmware like 6x12 suggested.
I checked the website you provided, and I am currently not using the memory listed there. However, this problem also occurred when I was using the memory that came with my Dell, so can we infer from this that the memory being used is not key to the problem?
If it is indeed caused by ECC, then I think the issue would be almost impossible to reproduce consistently and essentially unsolvable. But given the current situation, I suspect that it’s more likely there’s a problem with the way the system operates the memory. Of course, this would require more evidence.
1 Like
Yes, I’m going to update the BIOS version. The newest version I can find is 1.1.30, which was updated on July 13, 2023.
2 Likes
Even better, I only did a quick search.
1 Like
I’m afraid you misread. ECC exists so as to prevent data corruption in memory — or in transit from RAM to the drive controller and vice versa — due to electrostatic interference. So I think you probably meant the interference, rather than ECC. 
Yes, you are right. I misread ECC to single-event error (SEE ). 
1 Like
There is no way of telling that. All I can say is if you do not use supported memory, things, usually very strange things can occur.
Are you sure the old memory was not faulty?
Things like these are nearly impossible to analyze, since the testing grounds are flawed (your memory is probably not correctly configured, we are talking about fractions of a second here).
All I’m saying is that this is VERY hard for anybody to analyze, especially since you don’t use supported memory.
And you probably can’t post on dells forum, they will most likely just respond with “you are using unsupported memory”, but you could try.
If we were to entertain that suspicion, then the probability of cosmic rays causing this problem would also exist, and can’t be completely ruled out, right? In that case, the issue would be nearly unsolvable. I think we should start by verifying the most likely hypothesis. The current situation is that I’ve experienced this problem with both sets of memory. I changed one memory stick but didn’t change the system. If the problem persists even after updating to the latest system version, then I think it’s less likely that the memory is at fault. I can do a memtest on the current memory, but I don’t think it’s very useful to continue following this lead.
1 Like
They have always existed, since the dawn of the universe, Has nothing to do with faulty memory or not. That is what ECC can correct, normal memory never does. And cosmic rays are not a myth, they DO impact all electronics on earth, it’s just a matter of when things get faster and faster, every single little error can impact more than in the past.
It seems like you do not believe in cosmic radiation, OR think your memory gets impacted by cosmic rays (where did that even come from) hundreds of times a day, that is not the case. You might get impacted and not notice it, but if you on top of that has incorrect timings, well, chaos can occur.
I agree, and by this point, I would argue it’s no longer a hypothesis but rather a theory that the unsupported memory you use is involved one way or another.
So you not only got unsupported memory, you use DIFFERENT memory modules… Each memory has different timings they want.
It’s becoming more and more clear this is most likely unsolvable.
To be honest, I don’t think it’s any use trying to figure this out until you have hardware that is working well together. Sorry.
But a memtest might tell you things are not as they should.
Kinda the absolute first thing you should have done.
Edit
For clarification:
“Back in the day” these things were not as important, most memory just worked but you lost a tiny fraction of the speed (in most cases).
But today, when things are 10 000x faster, every little thing matters WAY more and it is VERY important to get correct hardware.
Software was also more accepting to errors than it is today. Security must be maintained within a system, and accepting changed memory f.ex is or SHOULD be a big nono on any os.
1 Like
I’m not saying I don’t believe it; I’m just saying that even if cosmic rays are causing this problem, I can’t solve it. I am simply looking for viable solutions, not to discuss whether I believe in cosmic rays or not. You seem to have completely strayed from the topic.
I think if it really is a memory issue, then indeed there’s no solution. However, given that others have already reported the same issue, I believe it’s highly unlikely to be a problem with the memory. Of course, you could suspect that the other person’s memory is also faulty, but that would seem like quite a coincidence. Since the cost of replacing memory is much higher than the cost of restarting, I would choose to restart. I’m just looking for possibilities beyond replacing the memory, so please don’t get hung up on whether or not there is a problem with the memory; it’s clearly a waste of time.
1 Like
I am already running a memory test. If the results show that the memory is fine, and since there are no other directions to explore at the moment, I guess I will have to give up looking for a solution for the time being.
YOU brought up cosmic radiation, not me.
I agree, the solution is most likely you getting supported hardware.
Yes there is, GETTING SUPPORTED MEMORY. But it’s not a guarantee it will fix it. What I keep repeating is that it is nearly impossible to analyze a system that is not “stable” because of incompatible hardware.
And there is probably none.
If you feel this is a waste of time, I step back and wish you good luck!