Graphics Cards

[ATX350]

Hi all,

I am at a loss and have no idea where to turn from here. I need help

I am experiencing PC crashes in every game I attempt to play on this new rig. The frequency of crashes depends on the graphics settings but even on the lowest possible settings, I am still crashing within about 30 minutes.

Any ideas?

Parts list:

* Radeon 6900xt
* Asus ROG Strix B550-i
* AMD Ryzen 5900x
* Super Flower Leadex V Gold PRO 850W
* Crucial Ballistix 3600 memory
* Sabrent 1TB Rocket 4 PLUS NVMe
* Samsung Q7 32"

Describe your problem:

Every game hard crashes my PC. I lose power and have to power it back on. Lowering graphics to the lowest settings buys me time but ultimately I still crash.

This I've tried:

1) Updated all drivers/firmware.
Mobo - Version 2423
GPU Drivers - 22.1.1
Radeon software - 2022.0107

2) Replaced PCIe riser card - https://www.amazon.com/gp/product/B092JTTLYF/ref=ppx_yo_dt_b_asin_title_o01_s00?ie=UTF8&psc=1

3) Disabled enhanced sync (reference - https://community.amd.com/t5/graphics/6900xt-sapphire-toxic-driver-crashes-in-games/m-p/504586/highl... )

4) Configured AMD chill to fix FPS between 30 - 60 fps

5) Tried under volting

6) I have tried 5 different power supplies. Originally Corsair SF750 and ultimately landing on Super Flower Leadex V Gold PRO 850W

EVGA Supernova 850 GM, 80 Plus Gold
SilverStone Technology SX1000 Platinum, 80 Plus Platinum 1000W
Cooler Master V850 SFX Gold Full Modular, 850W, 80+ Gold
Corsair SF Series, SF750, 750 Watt
Super Flower LEADEX V Gold PRO 850W

7) I've tried running OCCT and the power test immediately crashes my PC. I get no errors, PC just crashes.

Using 2 separate PCIe cables 

9) tried 2-3 different GPU driver versions 

My issues seem very similar to those reported here. I have already RMA'd my first 6900xt and I have experienced identical issues on both. I am running on my 2nd 6900xt. 

https://community.amd.com/t5/graphics/i-m-close-to-rmaing-my-gpu-rx6900xt/m-p/505146#M82240

[ATX350]

Here is an hwinfo log from the last crash I experienced: https://onedrive.live.com/view.aspx?resid=50DCA0139F845428!21778&ithint=file%2cxlsx&authkey=!ACL-WWV...

[ATX350]

Update: Disabled PBO/CBS and enabled DOCP as recommended in this thread but am still crashing. 

https://www.reddit.com/r/ASUS/comments/kq9sld/b550f_gaming_bios_ver_1202_where_to_find_pbo_and/ 

[goodplay]

Near the end of your log, the card is cooking 107 edge/119 hotspot.

What is your case make/model, how many cooling fans, what cooling setup on cpu ?

What model ram sticks ?

Forget reddit, get real advice on bios setup (optimal non oc) from an overclockers  forum 5900 owners thread (overclockers uk or overclock.net are a couple to start). 

[ATX350]

I am using A Meshlicious ITX case with an EK closed loop water cooler setup

Case - SSUPD Meshlicious Mini-ITX
GPU Waterblock - EK-Quantum Vector RX 6800/6900 - Copper
Pump - EK-Quantum Kinetic TBE 120 DDC
NVMe - Sabrent Rocket 4 Plus 
Fans - 2x Noctua NF-A14
Radiator - EK 280 Coolstream 

As far as what RAM I'm using, it's the following. Note that I have reseated the RAM a few times along with replacing the original RAM with an identical model/make also. 

RAM - Crucial Ballistix 3600 MHz DDR4 DRAM Desktop Gaming Memory Kit 32GB (16GBx2) CL16 BL2K16G36C16U4B

I did notice last night that my RAM isn't actually on the motherboard's QVL. Should I replace the RAM? https://rog.asus.com/motherboards/rog-strix/rog-strix-b550-i-gaming-model/helpdesk_qvl_memory 

[Souran]

Try with driver 21.10.2

I got a 6900XT (XFX ZERO WB), 5950X, any driver later than 21.10.2 is totally unstable for me. Latest driver is a nightmare, hard crash every 20mins.

[FederaluL]

I would look for a way to add at least another 240 rad in that loop. they recommend at least 120mmfan/rad/100w draw. with the processor you've got you probably reach well over 400w draw under load (and very high temps). where did you plug the fans/pump on the mobo? eventually try setting up a curve for them to try get lower temps on your setup. GPU could crash because of that once it goes over 105degrees jct temp.

[Geforsikus_2021]

Try not to set xmp for RAM in the bios, but the usual one, or reduce the value not 3600 MHz by 3400, or by 3200. And run the osst test.

[Matt_AMD]

Also, have you tried without the riser cable? They can cause problems.

[ThreeDee]

Maybe try running everything in a mesh front ATX case with the GPU's stock heatsink/fan .. as well as a good air cooler (NH-D15?) on your CPU and see if crashes persist

[ATX350]

Okay so I am abandoning my SFF dreams and am moving to the new 011 Dynamic Evo with dual 360 radiators to solve the heat issues. I'll keep you all posted once the new rig is built. Thanks for your help. 

[Geforsikus_2021]

Considering what you have done with the PC, one thing remains, if you fumble in the PC theme, take out the brochure (magazine) from the motherboard, and pull out all the RAM bars except 1 (look at the book in which channel 1 RAM bar should remain.) start the PC, try tests, and so with each RAM bar 1 pc, I just didn't find where you wrote how much RAM you have in general. In general, test, if it does not help, remove the cooler from the processor and carefully remove the processor, and reinsert it, there were cases that the amd processor mount could not hold the processor in the socket and it got out a little and there was no contact. We write that ASUS has hardware on the b550 chipset that is connected both externally and internally (Bluetooth WiFi and USB ports), If there is this problem, then they say the computer is hanging tightly, and restart only with a button. Possibly an unsuccessful motherboard (defective)