Drivers & Software

[Nakedjustice]

Summary of issue:
Hello everyone! Just upgraded my PC from 1080 + 3600x to 6950Xt + 5800x. Now I'm encountering a problem, when I play Fortnite (or Ready or Not or Grounded) with DX12 turned on 22.11.2 driver, my PC spontaniously turns off and reboots. No error messages, nothing at all. In Event Viewer I only found BugCheck 1001 error.

What I have tried:

Tried undervolting - didn't help (I did it for temp reasons on 22.5.2 driver)
Tried tweaking PBO - didn't help
Cruve optimizer - Any magnitude changes lead to reboot cycle even on the desktop. Tried negatives from -2 to -20. Ryzen Master setted it up to -28 on first 2 cores and -30 on each else.

As of now downgraded to 22.5.2 things seem stable unless I play games with DX12 api. If I do switch to DX12 - same sudden blackscreen shutdowns.
I have turned off MPO in 22.5.2, not sure if it did help anything.
After turning MPO off I tried to get back to 22.11.2 but got same random shut downs and my second screen started to flicker and show artefacts.
I saw plenty of stuff fixed in newer driver versions so I wanted to ask what can I try here.

Setup OS - WIN10 22H2, Build 19045.2364
MOBO - MSI MAG B550 TOMAHAWK MAX WIFI
GPU - MSI RX6950XT GAMING X TRIO 16GB
CPU - AMD Ryzen 7 5800X
CPU cooler - be quiet! Pure Rock 2 but I want to switch to Deepcool 280EX because of the average temp of my chassis.
RAM - DDR4 16GB (2*8GB) Kingston FURY KF432C16RBK2/16
PSU - DEEPCOOL PQ1000M ATX 1000W Gold
Monitor - MSI Optix G241, 1920x1080, 144hz, through HDMI
Monitor 2 - HP 22w, 1920x1080, 60hz through HDMI to DP cable.

[syraphx]

I'm in pretty much the same boat, but I'm replacing my PSU today for a Corsair RM1000x (single rail)...

5800x
deepcool castle 360EX
AMD 6950xt (bought from AMD direct)
Tomahawk B450 (non max) with latest Bios so it's compatible with the 5800x
G.Skill Ripjaw DDR4 3000 (don't have exact model)
2x SSD
Windows 10 home freshly installed (updated to latest non Win11 updates)
I'm using latest driver from Adrenaline
I can stress test with Furmark and prime95 (both at the same time) for over 30mins without crash, even on my EVGA Supernova G3 750 (multi rail) PSU.

As soon as I start Cyberpunk or Grounded, looks like everything shutdown, but RGB and fan keep going. Screen goes black, mouse, keyboard and headset lose power.

So... waiting to grab my Corsair RM1000x and test, but my friend gave me a 1000W Rosewill capstone 1000w (low tier quality PSU) and it's crashing as well, same symptoms.

If Corsair fail as well, I'm thinking of returning the card and grabbing an RTX, but I'm not even sure AMD direct will offer refund... since the box is open...

[Nakedjustice · Accepted solution]

Hey brother, I got my problems fixed. Wasn't CPU for me but GPU. The problem was in PSU, and 2 lines wasnt enough to power up the gpu at maximum utilization, so I reconnected it with 3 separate cables. Basically, 3 outputs from PSU to 3 inputs in GPU, and everything works like charm. Previously, it was connected via 1 separate line and 1 chain 6 it was like 1 + 2. Now it is 1+1+1.

Second screen flickering and artifacts happened due to cable malfunction (hdmi to dp wasn't good enough, I suppose), so I just bought DP to DP cable for one monitor and used HDMI to HDMI to second. Dunno if anything of this will help you, but I still decided to tell ya my variant of fix for this.

[Maddion]

I got the same GPU and after the newest driver update i got blackscreens all the time when its on load.

tried everything from undervolting to complete reinstall Windows and ddu stuff but nothing helps. 

I have 3 Monitors with different Hz (144hz,280hz,60hz) but doesnt matter if i plug in only one of those i got the same black screen and flickering.

PSU: 1000W Bequiet Platinum 80+ 3 Single 8 pins for GPU.. so PSU is not the problem.

i just sent my GPU to MSI. i hope they fix or replace the card asap. 

[eebiii]

Please dont take this the wrong way, I just need to know. Why would you think that 2 - 8 pin cables would work on a 3 - 8-pin card?