General Discussions

[jakefrances]

Hi,

I've been having issues since building my computer about 6 months ago. There was a period of a few months in which everything appeared to be working well, but outside of that narrow window, my system has been frequently crashing while playing or recording games. Sometimes, it does not crash for a few hours. Sometimes, it crashes upon loading the game. It crashes much more often while recording, but it also crashes often while only playing a game. Once, a crash corrupted my Windows OS and I had to reset my computer. That was a few weeks ago. Since then, things have gotten worse.

I have my BIOS, chipset and GPU drivers up to date. I've tried reinstalling Windows about 15 times. I've no idea what to do. How do I know which part of my computer is broken? How do I go about determining that? I can't just return every part as defective under warranty. I suspect my graphics card is broken, but how can I be sure? What is the first step when confronted with what appears to be a broken system?

My Hardware:
Drive: Seagate FireCuda 530 2TB Solid State Drive - M.2 PCIe Gen4 ×4 NVMe 1.4
CPU: AMD RYZEN 9 7950X Processor
RAM: G.Skill Trident Z5 NEO RGB Series (AMD Expo) 64GB (2 x 32GB)
Mobo: ASUS ROG Strix X670E-E Gaming WiFi 6E Socket AM5
GPU1: ASUS Radeon RX 7900 XTX 24GB GDDR6 (Ref. Model)
GPU2: ASRock Challenger Arc A380 6GB GDDR6
PSU: Corsair RM1000x Fully Modular ATX Power Supply - 80 PLUS Gold

My OS:
Edition: Windows 11 Home
Version: 22H2
Installed on: ‎7/‎20/‎2023
OS build: 22621.1992

If you need additional information from me, please let me know. Any help is appreciated.

Thanks.

[ThreeDee]

There were some issues with reference model vapor chambers (coolers) on the 7900xt(x)'s ... how are your GPU temps?

There were some issues with ASUS motherboards frying CPU's .. how are your CPU temps?

EXPO enabled in BIOS? Do you crash with EXPO disabled? Have you tried running with just one stick of RAM and see if it crashes?

Describe your 'crashes' .. do you get a BSOD? .. does the PC just shut down? .. do you just get kicked back to your desktop and whatever game/app just closes on its own? Does everything lock up?

Have you tried running with just one GPU? .. maybe your RM1000x isn't up to the task of running those particular GPU's together .. or maybe there is a driver conflict of sorts running those 2 GPU's together

Have you tried running with PBO/CPB disabled? .. or C-STATES disabled...?

Do you have Memory Context Restore disabled? IF you do, you need to make sure you have Power Down Mode disabled too

[jakefrances]

I don't use the Intel GPU for anything other than encoding, if I am recording. I have also disabled it to no avail. The Intel GPU doesn't get used at all, as far as I can tell, unless I explicitly tell Windows or a program (like OBS) to use it.

I've tried both AMD EXPO enabled and disabled to no avail. I have not noticed unusual temps on the CPU or GPU1. I have hit 110 junction temp when gaming in rage mode, but I don't think I've seen 110 with default settings. I have seen 100, though, in some games. Other than that, there is a high power draw of around 100 W while idle from GPU1, which is unusual but not unheard of.

The crash it typically a freeze, which leads to a full black screen, which then sometimes leads to a restart or a failed restart, or just sits on the black screen until I physically restart it. When I am recording, sometimes the freeze is followed by a full green screen, which then sometimes leads to a restart or failed restart.

I will run the other tests and report back.

Thank you.

[ThreeDee]

For testing purposes .. I'd remove one GPU or the other and see what happens .. or rather .. what doesn't happen

[jakefrances]

I'm okay with running more tests, but would it be appropriate to just send my GPU1 to be repaired under warranty? Or do I need to include the tests I've done? I'm kind of sick and tired of dealing with this. I was hoping there was a solution besides more arbitrary tests.

[ThreeDee]

testing stuff is for  your benefit .. what if it's running dual GPU's that is your issue and you get your brand new RMA'd card back and you still have the same issue(s) .. 

maybe it is a faulty GPU .. maybe it's your PSU .. maybe it's your motherboard's PCIe x16 slot ..? maybe it's some other software causing issues that you have installed .. hard to say without isolating parts as best as you can and running .... tests

[jakefrances]

I removed the graphics card drivers and then removed the AMD GPU. I then cleared BIOS and installed the latest Intel Arc Control. I then played Elden Ring on maximum graphics with maximum ray tracing at 4k while recording AV1 at 1080p and experienced no crashing. I am still running the game in the background. The frame rate is horrible, obviously, but I have not experienced any green or black screen of death.

At this point, what more tests need to be done? I believe this means that the AMD card must be broken.

What would be the next step? Or do you seriously think I should do further testing? I know that it comes with a 3 year warranty from ASUS. I want to either try to get my money back or a replacement.

If you think more tests really are needed, let me know. But I've been struggling with this for awhile and am becoming impatient.

[BigAl01]

I think you have one of the AMD 7900XTX video cards with the vapor chamber problem (lack of sufficient fluid so it gets hot spots); this was referenced by @ThreeDee in the beginning of this discussion topic.  As soon as you mentioned a temperature of 100 C, I knew that was your problem.  I recommend you run on the Intel Arc video card for now and RMA the 7900XTX.

[jakefrances]

could be vapor chamber or not. Regardless, the Intel Arc works fine and the AMD 7900 XTX is crashing like it's a 10 year old card, or something. I'm going to remove all the driver again, remove the Intel and try to play on the AMD card exclusively, for one final test. If there's still constant crashing, I don't think any further tests need to be done.

[BigAl01]

The thing to watch is the temperatures of the GPU.  If it's getting above 70 C then you should be concerned.  If it's beyond 80 C then it's time to figure out the problem.  Assuming you have good case ventilation (please confirm), then all arrows are pointing to the GPU having a cooling issue.  Don't run it over 100 C because you are looking an imminent failure. 

[jakefrances]

they removed junction temp in the latest adrenalin driver update. Not only do games and OBS work on the Intel card but the software encoders are magically working with the amd card removed. AMD claims 110 junction temp is normal, but regardless I know that a high-end card is not supposed to constantly crash when playing a normal game at default settings. That's not suppose to be happening. Why check the temps at this point? It's obviously either physically damaged or it's a lemon.

[jakefrances]

Also my case has 15 external fans. Airflow is not an issue. This is not my fault.

[BigAl01]

Then RMA the video card.  Also do some research on driver conflicts between the Intel Arc drivers and the AMD Adrenalin drivers.  It appears you have been experiencing some conflicts that stop when one of the video cards is removed.  Also, I would never consider running  a CPU or a GPU at 110 C and expect it to last more than a few months.

[jakefrances]

The AMD GPU has always had issues. This is the nail in the coffin for me. Junction temp is not the same as GPU temp. The GPU temp never gets anywhere near 110, but I get what you're saying. Regardless, the stupid card cannot play a game, which means it does not do what it advertised. Hopefully I can get my money back instead of a replacement, so I can get a better card.

[jakefrances]

Idk why I never thought to just take each card out and use each one individually. That should have been the first thing I did. I still kind of want an AMD card, though. Just sucks they rushed this crap out or whatever happened. It could have been damaged during shipping. Who knows.

[jakefrances]

Thanks, guys. Nice meeting everyone. Depending, if I get my money back, I may get a 4090, or a non ref 7900 xtx. I still like AMD. These things happen. There have been problems with NVIDIA cards, too. Hopefully, the process of making this right isn't too torturous. I'll let you guys know what happens. I'll report back in a week or two.

Thanks, again.
Jake

[jakefrances]

I don't think it's even covered by ASUS warranty.

This seems so strange, though. My A380 handles gaming better than the 7900 XTX, which is false advertising, isn't it?

Why does it matter if it is covered by a warranty or not? I didn't receive what I paid for. This is either a scam or a defective item.

[jakefrances · Accepted solution]

I spoke with a friend and he assured me that the card should be covered by ASUS warranty even though it is an AMD ref.

So, all is good. Going to ship the card out today or tomorrow. Thanks, gang for basically allowing me to collect my thoughts here. Thread can be closed.