Graphics

[Stringy_Lanky]

Hi all, please could I get some advice on this issue that has resisted my best attempts to resolve?
Built this new rig in December 2020 and borrowed a Sapphire RX 5700XT Pulse BE from a friend until the RX 6800 XTs were in stock. Computer ran fine with no issues.
Installed the new RX 6800 XT and returned the RX 5700 XT to my friend. Big mistake that was…

Issues
Computer crashes to black screen with no keyboard & mouse input, no sound output. The LEDs and fans continue to function. Motherboard debug light indicates a VGA problem, so it is the graphics card. Must be rebooted with a hard reset. Crashes do not seem to correlate with high intensity usage, as it crashes occasionally in 10-60 minute intervals when I am doing light web browsing. GPU Temperatures hovering around 41oC, while browsing, fans kick in above 50 oC.

EDIT 1: Seems highly likely to crash if I pause Netflix for more than a minute or two and the computer enters an idle state.

Steps taken

1. Unmounted and remounted graphics card.
2. Used two separate PCI cables instead of a single pigtailed cable (Duh!).
3. Uninstalled and reinstalled GPU driver.
4. Uninstalled and reinstalled GPU driver with no internet access to prevent Windows downloading a placeholder driver.
5. Installed DDU, cleaned all display drivers and AMD chipset. Reinstalled drivers and updated BIOS.
6. Read this thread about multiple screens causing issues, so unplugged my second monitor. (https://community.amd.com/t5/graphics/rx-6800-xt-black-screens-when-second-monitor-is-connected/m-p/...). Achieved greater stability and thought the issue was resolved.
7. Installed Unigen Heaven Benchmark, Unigen Superposition Benchmark, and Cinebench R23. Each ran High or ultra-settings without issue and ran Heaven and Cinebench together without an issue.
8. Had a day or two of gaming without issue, playing mostly Total War Warhammer 2, Arma 3 and Wargame Red Dragon.
9. Issue reoccurred after watching a few episodes on Netflix and then the day after while light web browsing.
10. Read this thread that suggested reducing the voltage to the GPU. Reduced voltage by 10%. (https://community.amd.com/t5/graphics/asus-tuf-6800-non-xt-complete-computer-freeze/m-p/439723/highl...)
11. Issue reoccured. Tried disabled Zero RPM fan tuning.
    
12. Issue reoccured. Highly likely to crash if watching Netflix and the video is paused, but has no issues watching multiple episodes conscecutively.
13. Due to intermittent nature of the issue, will report back later to see if resolved. Any suggestions welcome!

System

Hardware Specs

• Processor: AMD Ryzen 7 5800X 8-Core Processor 3.80 GHz
• GPU: Radeon RX 6800 XT Gaming X Trio 16GB
• Motherboard: MSI MPG B550 Gaming Edge Wi-Fi Motherboard
• RAM: G.Skill Trident Z Neo RGB 16GB (2x8GB), PC4-28800 (3600MHz) DDR4, 16-19-19-39, 1.35v, Dual Channel Kit, XMP 2.0
• CPU Cooler: be quiet! Dark Rock 4 CPU Cooler
• Case: be quiet! Pure Base 500DX RGB Case Black
• Storage:
  Samsung SSD 840 EVO 250GB (OS),
  Samsung 970 EVO NVMe SSD 1TB (High resource intensive games),
  WDC HDD WD20EZRX-00D8PB0 2TB  (Low resource intensive games & other media).
• PSU: CoolerMaster V850 Modular PSU Gold
• Monitors:
  HP e243i 24-inch,
  BenQ XL240T (unplugged in step 6).

Software Specs

• OS: Windows 10 Home, version 20H2
• GPU Driver: AMD Adrenaline (27.20.14501.33001)

[fenix144]

I seem to have the exact same issue as you. Same model card too. 

[dkydev]

Same issue, same GPU. Been debugging for nearly a week since building this PC. Will be watching this thread for a solution (hopefully).

Similarly never experiencing crashing while under load. Only crashing while idle or light workload. Temporary solution is to play games 24/7. Fingers crossed for a software/driver fix.

Specs:

• Processor: AMD Ryzen 9 5900X
• GPU: MSI Radeon RX 6800 XT Gaming X Trio 16G
• Motherboard: ASUS ROG Strix X570-F Gaming Mainboard
• RAM: Corsair Vengeance LPX Schwarz 32GB Kit (2x16GB) DDR4-3600 CL18 DIMM
• CPU Cooler: Corsair Hydro Series H100i RGB Platinum
• Case: Corsair 5000D Airflow Midi Tower ATX
• Storage: Corsair Force Series MP600 SSD 1TB M.2 2280 PCIe 4.0 x4 (removed heatsync to use heatsync on motherboard - will be testing new drive asap)
• PSU: Corsair RM750x v2 - 750 Watt

[mackbolan777]

Try using the high  performance power plan and disable sleep/hibernation modes. Why not migrate the OS onto the NVME 1TB? Because you'd need to reinstall the games if the 840 took a dive anyway. There's a newer driver out as well, which did you try it yet? Last thing I might suspect is a defective PSU. I use the Corsair RMx series now and the RM's in prior builds. The Rmx's have a 10 year warranty now. I would say the CPU is ok because this all worked with the 5700. It could boil down to a defective card as MSI has been having a tough time with the Trio and the Merc.

[Stringy_Lanky]

Update:

Step:

14. Installed Memtest86 to check RAM and found one stick was producing 10k errors in <2mins. Suspect RAM fault could be playing a role and it was only noticeable now that RAM is theoretically the bottleneck in system performance. Removed the Error causing stick to operate the machine only on one and issue reoccurred. 

15. Made warranty claims with the retailers for GPU and RAM. Will report back to see if warranty claim is confirmed as a fault and accepted.

if not then I’ll investigate PSU and power/sleep settings thanks for the tips!

[mackbolan777]

I'd say you got it with the RAM alone. If you swapped slots/sticks and still got errors, it's most likely the board. But I would start with the RAM first.

[dkydev]

Probably right about the board.

Here is a very long thread on the MSI community forums regarding the 6800XT Gaming X Trio: https://forum-en.msi.com/index.php?threads/rx6800xt-gaming-x-trio-black-screen-random-signal-loss.35... 

In the meantime it appears running a single display with HDMI is more stable than DP, at least with the latest BIOS (020.001.000.049.000000).

[mypg0306]

do you enable DOCP/XMP at ram rated speed?

[Stringy_Lanky]

Yep I had XMP enabled automatically and it detected and applied the correct speed. At one point I disabled xmp by resetting to default BIOS settings. Had no effect on the issue. Thanks for the suggestion!

[Stringy_Lanky]

Hopefully the MoBo is still good. Thanks for the link to the thread. Clearly a common problem without a clear solution yet. Looking forward to whatever MSI does to solve it soon!