PC Graphics

[drgouh]

Hello there,

I know this is a common subject, I've already read plenty of threads about it, but none of them really worked for me and I was hoping that some good soul on here could tell me what exactly is the issue.

While playing my screen goes black (or sometimes the GPU is lost then recovered after a few seconds but rarely) at random times. I'm 99% sure that this is a software related issue.

I've already tried a clean installation of all my drivers (deleted everything with DDU, Control ¨Panel, C:/AMD) then updating first through Windows Update and then thanks to AMD driver (both using and not using auto detect). I've tried both Stable and Beta drivers. My BIOS is up to date, my drivers are, I did all Windows updates too. I don't know what to do anymore but it keeps happening and it's so random, I'm not even playing GPU-heavy games.

Oh I also tried to tweak my GPU settings with Radeon as explained in this thread.  It worked for a while but not for long unfortunately... 

This thread also mentions a VBIOS issue, but I'm a bit scared about playing with that, I'm ready to do it if that's the issue though.

If I can provide anything to help you analyze the crash, please tell me.

Here is my System:

Motherboard: MSI B450 Tomahawk Max

PSU: Corsair RM650x V2 80PLUS Gold  

Processor: AMD Ryzen 5 3600

Graphics card: Sapphire Pulse Radeon RX 5600 XT 6G (299-5E411-002SA) (bought in April 2020) more here

RAM: Crucial Ballistix 16Go (2*8Gb) DDR4-3600

Storage: CT500MX500SSD1 (SSD) Crucial

I've tried to gather some info about the crash, here and here.

I'm also contacting Sapphire about the issue, hopefully they'll help me too.

[Matrixx]

Dear, it seems like driver problems for your card. I had similar problems and I advise you that the first thing you do is download the MemTest64 that runs under Windows Desktop, that is, with Windows started, leave it testing for at least 20 minutes and tell me. a greeting P.S. The program is free and portable.

[Coffeecat]

https://community.amd.com/t5/graphics/black-screen-crash-subject/m-p/441763#M71655

[drgouh · Accepted solution]

Thanks everyone for your answers, I just received an answer from Sapphire, haven't had any issue since I did what they told me. So I'm sharing their solution and I'm marking the thread as solved until I run into more issues.

Thanks again for your support.

Dear Customer : Please follow below steps.

1. Within MB BIOS setting page , please set PCIE speed to Gen3 instead of AUTO if your MB does not support Gen4.

2. Install Catalyst V21.1.1 Optional driver here - https://www.amd.com/en/support/kb/release-notes/rn-rad-win-21-1-1 , then reboot the system.

3. Go to CCC and make sure that disable Metrics Overlay in Performance tab.

4. Adjust Power Limit to +20% in CCC 5. Check if your GPU is overheating that more than 100 degree C when in load , if yes , please add one more case fan to cool down the ambient temperature. 6. Use only one signal monitor to isolate the problem first.

[efenstor]

Just my two cents for those who suffer from this or similar problems, especially on Linux and with the ASUS motherboard (at least this helped me after three years of suffering):

1. Make sure that all of your DIMM memory modules have identical timings (e.g. all four are 15-17-17-35). If not, that is the problem number one which should be fixed. It is the best if all of your memory are the same model and by the same manufacturer. After that you can safely use D.O.C.P. overclocking.
2. In BIOS: Just as drgouh mentioned above, find the "Force PCIe gen speed" setting and set it to "Gen3" instead of "Auto".
3. In BIOS: Find the "CPU PCIE ASPM Mode Control" setting and set it to "Disabled".
4. In the kernel command line (GRUB_CMDLINE_LINUX_DEFAULT line in "/etc/default/grub") add amdgpu.aspm=0 and do "update-grub".
5. (helpful for overall stability) In BIOS: Set "Global C-state Control" to "Disabled".
6. (may be also helpful) In BIOS: Unless you have a very old PCIe card set "PCIEX16_2 4X-2X Switch" to "4X".