Question GPU nightmare - help?

May 9, 2024
10
0
10
Hello, Tom's Hardware forums, is my Asus Strix 5700 XT GPU dying?


Issues:
Drivers crash randomly
some games stutter and freeze periodically.
-Rare bluescreen ( happened a few times)


OS:
Windows 11 (terrible I know)


What I've tried so far:

2x DDU- uninstalled and reinstalled drivers twice
Updating Drivers- Updated multiple times
Reverted Drivers- Nothing happened
Checked windows integrity

minidump file:
https://mega.nz/file/AHUkAA5J#Te4nf7NfnrtUeXB27EsTje5YdsxlzwVW74t27VPutV8


Someone else indicated that the minidump file indicated "dxgmms2.sys"



https://gpuz.techpowerup.com/24/05/09/978.png


Don't mind the current out of date drivers as updating them and reverting them did nothing previously.



Currently my bios is out of date if that is of any revelance, not sure if that would affect drivers anyway, this gpu has seen some heavy use, I have mined for a little while (3 months maybe tops) other than that it's been mainly used for gaming. I've repasted the GPU twice since I got it.

any potential feedback is welcomed, as I got only 1 reply on reddit.
 

Eximo

Titan
Ambassador
Hardware can fail at any time. Motherboard BIOS, doesn't hurt to update.

Have you looked at updating the BIOS on the GPU? Those can get corrupt every once in a while.

 
  • Like
Reactions: EntityS
May 9, 2024
10
0
10
Hardware can fail at any time. Motherboard BIOS, doesn't hurt to update.

Have you looked at updating the BIOS on the GPU? Those can get corrupt every once in a while.

Thanks for the input, ill try that, didn't know that gpu bios was even a thing lol
 
May 9, 2024
10
0
10
You need to give full system specs, including make/model of the power supply.
Full specs:

MOBO:MSI B450 GAMING PRO CARBON MAX WIFI
CPU: Ryzen 5 3600x
RAM: CORSAIR Vengeance LPX 16GB (3600mhz)
GPU: ASUS ROG STRIX AMD Radeon RX 5700 XT
SSD: WD Blue 3D NAND 1TB - WDS100T2B0B
PSU: EVGA 600 BQ series
 
May 9, 2024
10
0
10
Hardware can fail at any time. Motherboard BIOS, doesn't hurt to update.

Have you looked at updating the BIOS on the GPU? Those can get corrupt every once in a while.

Update:

After updating vbios, some issues got better ish, others not so much.

Sometimes, now it stutters and freezes periodically in some games instead of crashing, and it's 50/50 whether it unfreezes or just gives me a driver crash popup.

Additionally, I had a game freeze, and then my PC crashed (no blue screen) it just shut off and restarted.
 
Apr 19, 2024
67
13
35
Well, is there a way to take the card and use it in another system for testing pusposes?
I mean in order to narrow the problems, we suggest to put the card in another working system.
After some tests we can see what's next.
 
  • Like
Reactions: EntityS
May 9, 2024
10
0
10
Well, is there a way to take the card and use it in another system for testing pusposes?
I mean in order to narrow the problems, we suggest to put the card in another working system.
After some tests we can see what's next.
The problem is that I don't have another system. I'll do a clean install of OS and see if it helps fix it, but either way, I am looking to upgrade in a few months (GPU, PSU, and most likely the storage), so I might just bite the bullet if it doesn't work and wait until then.
 
Apr 19, 2024
67
13
35
The easiest way to eliminate a hardware part at a time is to check it in another system.
Yeah if you can would be nice to make an upgrade.
But until then we don't throw any hardware part unless we tested it and the result indicate it as a faulty part.
 
Hello, Tom's Hardware forums, is my Asus Strix 5700 XT GPU dying?


Issues:
Drivers crash randomly
some games stutter and freeze periodically.
-Rare bluescreen ( happened a few times)


OS:
Windows 11 (terrible I know)


What I've tried so far:

2x DDU- uninstalled and reinstalled drivers twice
Updating Drivers- Updated multiple times
Reverted Drivers- Nothing happened
Checked windows integrity

minidump file:
https://mega.nz/file/AHUkAA5J#Te4nf7NfnrtUeXB27EsTje5YdsxlzwVW74t27VPutV8


Someone else indicated that the minidump file indicated "dxgmms2.sys"



https://gpuz.techpowerup.com/24/05/09/978.png


Don't mind the current out of date drivers as updating them and reverting them did nothing previously.



Currently my bios is out of date if that is of any revelance, not sure if that would affect drivers anyway, this gpu has seen some heavy use, I have mined for a little while (3 months maybe tops) other than that it's been mainly used for gaming. I've repasted the GPU twice since I got it.

any potential feedback is welcomed, as I got only 1 reply on reddit.

Other than the PSU, which should be tested, specially if it have a few years working ...

... Have you used some monitoring software (like hwinfo64 portable - "sensors only" option) to check the working temps under load of both CPU and GPU, and if posible VRAM temps too?
 
May 9, 2024
10
0
10
here are 2 test you could made use occt to stress test psu and then gpu-z on rendering option for gpu check temps voltages and fans speed values with hwinfo that you could made a log off it .
I tried to run OCCT a bunch of times, and all tests resulted in the program freezing/PC freezing after 17ish seconds in. However, the PC never turned off by itself or throttled down the CPU fan. I am not sure how to fix that

gpu z render:

View: https://imgur.com/BwJsfVJ
 
Do you ever have problems/issues while not loading the GPU with games or the benchmark you just tested?

Im asking cause indeed the PSU is not a very good model if I remeber correctly the 5700 XT needs 2 x 8 pin PCIe power cables, right ?

How old is the PSU, I mean how many years of use does it have ?
 
May 9, 2024
10
0
10
Do you ever have problems/issues while not loading the GPU with games or the benchmark you just tested?

Im asking cause indeed the PSU is not a very good model if I remeber correctly the 5700 XT needs 2 x 8 pin PCIe power cables, right ?

How old is the PSU, I mean how many years of use does it have ?
The benchmark didn't have any issues with loading GPU or any it made aware to me.

Yeah the 5700xt uses 2 8pins

I've had this power supply since I built this PC, so since 11/26/2020, or about four years. I also bought a braided power cable extension set for this system.
 
Wait, you just wrote:

"I tried to run OCCT a bunch of times, and all tests resulted in the program freezing/PC freezing after 17ish seconds in. However, the PC never turned off by itself or throttled down the CPU fan. I am not sure how to fix that"

So What component was the benchmark loading when the PC according to what you wrote "freezing/PC freezing after 17ish seconds in"?
 
May 9, 2024
10
0
10
Wait, you just wrote:

"I tried to run OCCT a bunch of times, and all tests resulted in the program freezing/PC freezing after 17ish seconds in. However, the PC never turned off by itself or throttled down the CPU fan. I am not sure how to fix that"

So What component was the benchmark loading when the PC according to what you wrote "freezing/PC freezing after 17ish seconds in"?
I was testing power.

I just tried running the power test again, and it didn't freeze or stutter, 1hr tested on full GPU and CPU until
it gave me a 1 WHEA error.




OCCT:
00:00:00 - Info - Test schedule started at 2024-05-16 18:27:26
00:00:00 - Info - Power - Initializing (Duration : 01:00:00)
00:00:00 - Info - Power - Started (Duration : 01:00:00)
00:21:32 - Warning - 1 WHEA error(s) found
01:00:00 - Info - Power - Test stopped


View: https://imgur.com/sMclFXU
 
May not be the GPU at all if programs are freezing/crashing. GPU drivers are notoriously extremely sensitive to any system memory instability, and the maximum rated speed of memory for your Zen 2's memory controller is 3200MHz with two sticks.

Most will overclock to at least 3800, but 3600 is indeed overclocking, and it is up to you to make sure voltage and timings are such that the result is absolutely stable. The quickest way to rule this out is to try again with memory set at 3200 or lower.

Also, many former mining cards will eventually need the GPU's memory underclocked some to continue working, as that's the part that sees the most wear in mining.
 
  • Like
Reactions: EntityS
Try runing Cinebanch R20 or R 23 (you can get it from guru3d), run the multi-core test a few times and see if that also freeze.

If it does, try (as BFG-9000) wrote, to set the memory DOCP/XMP to 3200, and benchmark again.
 

scout_03

Titan
Ambassador
you could see in the occt test temp are very high do you have any debug led staying on when it boot will start to solve the temp issue then you told us you use extension cable there are from same maker of psu .