• Welcome to TechPowerUp Forums, Guest! Please check out our forum guidelines for info related to our community.
  • The forums have been upgraded with support for dark mode. By default it will follow the setting on your system/browser. You may override it by scrolling to the end of the page and clicking the gears icon.

RX 7900XTX game crashes suddenly

Status
Not open for further replies.

ColonelKurtz

New Member
Joined
Sep 4, 2024
Messages
11 (0.04/day)
Location
Turkey
Hello. I upgraded from NVIDIA 3060ti to RX 7900XTX 2 weeks ago. When I first bought and I installed clean with DDU removed NVIDIA driver and there was no problem 1st day. I played Counter Strike 2 for almost 4 hours and nothing happened, but when I entered the game again the next day, the games started to crash. And there is no error code (But I catch some check photos please) The game just crashes and I go back to the desktop. I don't know what the problem is. All my drivers are up to date.
Since then, crashes continue to occur. I'm also getting crashes in other games. -> Red Dead Redemption 2 - Mafia Definitive Edition - Squad - Squad 44 (I test these so far)
Using Adrenaline default mode.

I tried:
DDU (Previous driver and latest one still crashes)
MemoryTest (0 failure)
XMP Off - On (still crashing)
Power cables plugged correctly I am sure.


System Specs:
GPU: Sapphire Pulse AMD Radeon RX 7900XTX
CPU: AMD Ryzen 5 5600X
PSU: MSI MPG A1000G 80 Gold Plus 1000W
CASE: MSI MPG GUNGNIR 300R
LIQUID COOLING: MSI MAG CORELIQUID M360
MB: MSI B550 Gaming Gen 3
RAM: G-Skill RipJaws 16X16 (32 GB) 3600MHz
HDD 1: Western Digital Black 1TB
HDD 2: Western Digital Blue 2TB
Sata SSD: Kingston 500GB
Sata SSD 2: Kingston 240GB
SSD NVMe M2: Kingston 1TB


GPU Tempeture: 75c while playing game max setting. CPU Tempeture: 60

Some problems & Case & Adrenaline Default - Check photos please.
 

Attachments

  • Screenshot_6.png
    Screenshot_6.png
    439.4 KB · Views: 363
  • Screenshot_1.png
    Screenshot_1.png
    288.2 KB · Views: 482
  • Screenshot_5.png
    Screenshot_5.png
    26.3 KB · Views: 430
  • IMG_7188.JPEG
    IMG_7188.JPEG
    1.5 MB · Views: 448
My first thought would be power supply, are all PCI connectors getting their own cable run?
 
My first thought would be power supply, are all PCI connectors getting their own cable run?
Exactly I am sure I plugged it corectly like this.
 

Attachments

  • Screenshot_8.png
    Screenshot_8.png
    148.4 KB · Views: 217
My first thoughts would also be PSU or memory. Test out memory first, with memtest86 boot USB stick to rule that out.

I just recently dealt with black screen crashes with GPU fans to full on my XFX 7900 XTX, it ended up being my 1200 watt corsair HXi (2019 version). It passed PSU tester test and other video cards - rx 6700xt & rtx 2070 super - ran fine, but with the 7900 xtx it would crash occasionally. Took me a year to finally properly diagnose it after it would crash almost every time.
 
My first thoughts would also be PSU or memory. Test out memory first, with memtest86 boot USB stick to rule that out.

I just recently dealt with black screen crashes with GPU fans to full on my XFX 7900 XTX, it ended up being my 1200 watt corsair HXi (2019 version). It passed PSU tester test and other video cards - rx 6700xt & rtx 2070 super - ran fine, but with the 7900 xtx it would crash occasionally. Took me a year to finally properly diagnose it after it would crash almost every time.
I test with memtest86 already. But how said 1000W is enough and I just bought 1 week ago.
 

Attachments

  • Screenshot_9.png
    Screenshot_9.png
    3.8 MB · Views: 171
  • Screenshot_10.png
    Screenshot_10.png
    4.3 MB · Views: 155
  • Screenshot_11.png
    Screenshot_11.png
    3 MB · Views: 159
Last edited:
I was dealing with a game support team (specificly Forza Motorsport 8...and separately, a little bit of this forum) before deciding a RMA/repair is needed for my PowerColor Hellhound 7900XTX. It survived every stress tests and benchmarks I threw at it, but staying in a heavy menu screen long enough could cause either graphical glitches (Forza Motorsport 8) or crash (F1 Manager 2023). I was hoping that it was a Forza game engine issue instead of a hardware issue. Yes, I know I was high on copium during the whole fuss. For a very long time the Forzas are the only heavy game I play. I touched F1 Manager the first time after I changed GPU and, boom, I knew there really was something wrong in the GPU.

Here's what the support team had thrown at me other than those already mentioned, in no particular order. All of them didn't help my case, but it might help yours. Good luck.

- No overclocking (XMP, curve optimizer, whatever)
- Underclock the GPU
- Clean boot (on Microsoft terms, disable all non-Microsoft startup applications and services)
- A new administrator account
- Make sure GPU slot and contact is clean
- Disable Smart Access Memory on BIOS
- Reinstall the affected game on different drives
- Leave only single drive and single monitor connected
- A fresh OS install
 
You've got some weird cable issues in your photo.

These appear to be different types of cables connecting to the GPU. Two that have shrink wrap, and one with just loose cables.

1725421217161.png


Then there's this stray that I don't know what it's doing.

1725421244397.png


You said you changed power supplies - please say you're not mixing cables from both supplies. If so, get rid of any old cables, and only use the cables included with the new MSI PSU.
 
Reduce PSU load :
-Use only one storage
-Try one memory stick
-Turn off any RGB fans
-Change speed of PCIE16 lane
After that reinstall fresh OS and the driver
 
Do you have the latest chipset drivers installed?
Try playing games with a 60 fps cap in lowest detail (for minimum GPU load) to see if the issue persists. Did you play the same games before with the 3060Ti without a problem?
You should also try your old card with the new power supply to rule out a faulty PSU.
 
You've got some weird cable issues in your photo.

These appear to be different types of cables connecting to the GPU. Two that have shrink wrap, and one with just loose cables.

View attachment 361949

Then there's this stray that I don't know what it's doing.

View attachment 361950

You said you changed power supplies - please say you're not mixing cables from both supplies. If so, get rid of any old cables, and only use the cables included with the new MSI PSU.
1725443170942.png
 
My first thoughts would also be PSU or memory. Test out memory first, with memtest86 boot USB stick to rule that out.

I just recently dealt with black screen crashes with GPU fans to full on my XFX 7900 XTX, it ended up being my 1200 watt corsair HXi (2019 version). It passed PSU tester test and other video cards - rx 6700xt & rtx 2070 super - ran fine, but with the 7900 xtx it would crash occasionally. Took me a year to finally properly diagnose it after it would crash almost every time.
how the heck is a 1000w psu crashing? was it defective or overloaded?
 
I test with memtest86 already. But how said 1000W is enough and I just bought 1 week ago.
Download HWinfo64 and see what the voltages are while you are gamming by setting it in log mode. If you see the 12V rails drop below 11.50V, it is more than likely your supply.
 
Hello. I upgraded from NVIDIA 3060ti to RX 7900XTX 2 weeks ago. When I first bought and I installed clean with DDU removed NVIDIA driver and there was no problem 1st day. I played Counter Strike 2 for almost 4 hours and nothing happened, but when I entered the game again the next day, the games started to crash. And there is no error code (But I catch some check photos please) The game just crashes and I go back to the desktop. I don't know what the problem is. All my drivers are up to date.
Since then, crashes continue to occur. I'm also getting crashes in other games. -> Red Dead Redemption 2 - Mafia Definitive Edition - Squad - Squad 44 (I test these so far)
Using Adrenaline default mode.

I tried:
DDU (Previous driver and latest one still crashes)
MemoryTest (0 failure)
XMP Off - On (still crashing)
Power cables plugged correctly I am sure.


System Specs:
GPU: Sapphire Pulse AMD Radeon RX 7900XTX
CPU: AMD Ryzen 5 5600X
PSU: MSI MPG A1000G 80 Gold Plus 1000W
CASE: MSI MPG GUNGNIR 300R
LIQUID COOLING: MSI MAG CORELIQUID M360
MB: MSI B550 Gaming Gen 3
RAM: G-Skill RipJaws 16X16 (32 GB) 3600MHz
HDD 1: Western Digital Black 1TB
HDD 2: Western Digital Blue 2TB
Sata SSD: Kingston 500GB
Sata SSD 2: Kingston 240GB
SSD NVMe M2: Kingston 1TB


GPU Tempeture: 75c while playing game max setting. CPU Tempeture: 60

Some problems & Case & Adrenaline Default - Check photos please.
Any manual PBO settings and/or CurveOptimizer negative steps for the 5600X?
 
It looks like you may have plugged an EPS cable in place of the PCIe one. You should've used these two (both connectors from the 16-pin cable and a single one from the 8-pin):

1725446319929.png
1725446339044.png


Alternatively, you could try two of those, but the first option is preferred:

1725446339044.png
1725446339044.png
 
Last edited:
As usual people on this forum have this weird obsession with blaming everything on PSUs.

You cannot mix and match EPS and PCIe, they have differently shaped connectors.

PSU issues do not crash games back to desktop, they crash the entire system = PC shuts down, OP is getting driver errors clearly indicating something to do with the GPU.

I'd return/RMA the GPU.
 
Last edited:
make sure the 7900xtx is level to the Motherboard…

just remove all nvidia drivers software and any overclocking Utilities, (afterburner, etc) that you are using with the nvidia card.
(IMO, on AMD cards use the adrenaline software to overclock AMD cards. )


then download 24.5.1 driver if playing “Black Myth :Wukong”, (IMO wait until 24.9.1, for the fix for Black Myth :Wukong) …. (install adrenaline software too)
profit!.

overclocking the 7900xtx too much (and undervolting) can cause weird problem like those.
 
Small world

I would start by rewiring that card, run 3 power cables to your power supply instead of daisy chaining
 
From personal experience I got game crashes to desktop and GPU driver resets when I got curve optimizer too far from what the CPU can tolerate.

For the record I dont see any differences between the 3x8pin connectors powering the card that some indicate... Just that the far right is not shown whole.
Its not clear but I think all 3 are 6+2pins (I think I see the division in the plastic housing and also the cables to the 6 + 2 pins are different angled

Untitled_175.png

Also GPU/VRAM OC or GPU UV can cause this too...

Small world

I would start by rewiring that card, run 3 power cables to your power supply instead of daisy chaining
I think the only daisy chaining going on is the 6+2. I see 3 different cables there

Here

Untitled_176.png
 
Yeah but always run 3 cables on these high power GPUs, This is why 12VHPWR was needed
Thats what Im saying and shown in last pic...
3 cables
 
I think there is a problem with AGESA ComboAm4v2PI 1.2.0.Ca update (if you did a bios update) and/or latest windows 11 updates , got some crashes too and ive changed 2 PSUs to test + underclocking everything (helped a bit but still happened). What im trying right now and seems to help is running the game in compatibility mode: win8, no crashes so far but i feel that i have a bit less performance.
 
I think there is a problem with AGESA ComboAm4v2PI 1.2.0.Ca update (if you did a bios update) and/or latest windows 11 updates , got some crashes too and ive changed 2 PSUs to test + underclocking everything (helped a bit but still happened). What im trying right now and seems to help is running the game in compatibility mode: win8, no crashes so far but i feel that i have a bit less performance.
Win version?

Meaning 23H2 or anything else?

Did you get the KB5041587
 
Daisy chaining is simply not an issue with single rail power supplies.
 
Status
Not open for further replies.
Back
Top