What leads to this kind of artifacting?
What leads to this kind of artifacting?
Hey, I've noticed this kind of issue mostly in Electron apps and with the Brave browser. It seems like it might be related to hardware or software problems. I also experience screen freezes in certain sections of apps like Spotify, which unfreeze when switching tabs. This problem appeared to resolve itself after a few frames while using alt+tab.
Additional details:
PC specifications
GPU: 4070 VENTUS 3X OC 12GB
CPU: i5-12600KF
MOBO: B760M GAMING X AX DDR4
RAM: Kingston FURY Beast DDR4 3200 MHz 16GB 2x8GB CL16
SSD: WD Black SN850X
Power supply: ATX 750W THERMALTAKE SMART BM2
HDD: Seagate barracuda st1000dm010 1tb 3.5" sata3
Mouse: Corsair Harpoon
Keyboard: Tactical Ozone Mini.
I have thoroughly evaluated each part using rigorous tests and various benchmarks (OCCT, Furmark, memtest), with no errors found—especially during OCCT and Furmark tests, where no artifacts or issues were observed.
Several potential causes were considered; here are the main possibilities:
1. Malware – likely the first consideration.
Run a scan using AV tools, such as the free version of MalwareBytes.
Link: https://www.malwarebytes.com/
2. RAM problem – might stem from file corruption causing the image to split.
For RAM, try downloading and executing memtest86:
Link: https://www.memtest86.com/
Instructions for installation and usage are available here:
https://www.memtest86.com/tech_creating-window.html
A single full pass (15 tests) is the minimum; 2 passes are ideal, while 4 passes are acceptable.
It may take some time, so it’s best to run it overnight.
- 1x 8GB takes about 1 hour per full pass → ~4 hours for 4 passes.
- 2x 8GB takes ~2.5 hours per pass → ~10 hours for 4 passes.
- 2x 16GB takes ~5 hours per pass → ~20 hours for 4 passes.
If no errors appear, RAM appears to be healthy.
If errors occur, replace the RAM.
For thorough RAM testing, 32 full passes are recommended (as per test #7).
Additional guidance:
https://www.memtest86.com/tech_individua...descr.html
However, most people won’t attempt 32 passes due to the time required, especially with ample RAM. Tech support generally agrees that 4 passes suffice.
3. GPU/monitor issue – likely because the screen doesn’t refresh automatically; hovering the mouse over it causes the app below to disappear.
Try booting in Safe Mode with Networking to see if the problem persists. If it does, the issue is probably related to third-party applications launched at startup. If the problem disappears in Safe Mode, the fault may lie with the Windows kernel or hardware.
Could the power supply unit be of poor quality and unsuitable?
🤔
This suggests a low-quality PSU that should be avoided.
For an opinion on your PSU, consult the official tier list:
https://docs.google.com/spreadsheet...JW...1973454078
You’ll see it’s Tier C.
For a new PSU, consider models from Tier A such as Seasonic Focus, Vertex, PRIME, Corsair RMx/RMi/HXi/AXi, or Super Flower Leadex Gold/Platinum/Titanium. A 750W unit works well.
(My three PCs are also powered by Seasonic. I own PRIME 650 Titanium (Tier A+), PRIME Ultra 650 Titanium (Tier A+), and Focus PX-550 (Tier A). Detailed specs with images are in my profile.)
Hey, first off, thanks for the reply. I'll give some additional information as it might help. When running PC Help Hub's file repair script, I get 3 errors, one of which never ends up actually getting fixed. I tested my GPU with OCCT, and every test came back just fine, both VRAM and the 3D tests came back just fine.
I wanna believe my GPU isn't the problem here, as if it was, I assume I'd also see this happening in games, but it doesn't happen in any type of games, at all, just on these Electron / Chromium based apps as I said. My first thought is also system corruption, probably coming from the RAM. I would check if the error is replicable through safe mode, but it is extremely random and happens whenever it feels like happening.
When I built this PC I wasn't as knowledgeable as I am today and I had a limited budget so I reduced the expenses through my PSU, which is a really dumb move, but I do plan on replacing it. I'll run the memtests today, and get back to you. As for scanning with malwarebytes, I only got warnings for process hacker and Monero wallet, which are fake positives, so I'm not sure how that helps my case.
Why not rely on the built-in commands provided by the operating system for fixing kernel issues?
🤔
1. Launch Start.
2. Look for Command Prompt (or cmd.exe), right-click the correct option and choose Run as administrator.
3. Enter the following command to initiate a system repair and press Enter:
SFC /scannow
4. If the scan detected and repaired corrupt files, repeat the same command just in case:
(if no errors were identified, proceed to step #5)
SFC /scannow
5. After the scan shows no issues, run this command for a quick verification and press Enter:
DISM /Online /Cleanup-Image /CheckHealth
6. When corruption is confirmed, enter another command for a deeper analysis and press Enter:
DISM /Online /Cleanup-Image /ScanHealth
7. If problems are found with the system image, type a different command to address them and press Enter:
DISM /Online /Cleanup-Image /RestoreHealth
8. After completing the repair process, enter another command and press Enter:
SFC /scannow
9. Finally, after all steps finish, simply press Enter to close the terminal window.
Note: Commands like SFC /scannow, DISM /ScanHealth and DISM /RestoreHealth might require a significant amount of time.
Do not close the terminal window or stop the system while these scans are running, even if they seem to stall.
👍
Solving PC problems often involves narrowing down possibilities. This means testing one component at a time until the problem is identified.
Indications you're facing suggest multiple potential causes. Testing in Safe Mode can provide valuable insights. Even if it doesn't show up in Safe Mode, the issue may lie with third-party software or drivers that run during normal startup. If this is the case, there are methods—though time-consuming—to trace the source. First, perform a memtest and attempt to reproduce the error in Safe Mode. If unsuccessful, I can assist further.
I have also used Windows manual commands, but haven't encountered any problems. I just returned home and the memtest was quite quick; it completed four passes in roughly four hours. I'm not sure if that's a positive sign, but at least I didn't get any errors.
I plan to recheck the Windows disk checking commands, then try to trigger the issue more often. After that, I'll switch to safe mode to see if I can reproduce it. My GPU tests with OCCT and Furmark also showed no errors. I've run HDSentinel on my disks as well and everything is fine. I'm still not completely sure what's happening, though.
DISM & SFC showed no issues at all. Everything looks good with the OS files. The OCCT GPU scans also came back normal, which makes me think the problem might not be with my GPU.
Tested my GPU using 3DMark's stress test, port royale, and time spy. Results were all within acceptable limits, less than 1% deviation, which seems typical. Ran DDU recently and no issues have appeared today.
You sure that you selected all 15 tests before starting the scan? Since by default, not all tests are selected when you initially launch memtest86 program.
It is possible that GPU drivers had some issue and using DDU + driver reinstall fixed it.
Just keep an eye out when the issue comes back.
If it does come back, try to reliably reproduce it within normal boot. Once you have that set, try to reproduce the same steps in Safe Mode and look if it happens there as well or not.
Also, do look into new PSU.
Yes, I chose all of them, completely confident. I'll try it on the PSU. If the problem recurs, I'll revisit the discussion. The only thing I'm noticing right now is that my GPU seems to be lagging in benchmarks, so... that's it.