F5F Stay Refreshed Hardware Desktop GPU malfunction without total system failure or black screen

GPU malfunction without total system failure or black screen

GPU malfunction without total system failure or black screen

C
ClareCloud
Member
59
10-15-2020, 04:53 PM
#1
VRAM-heavy applications and software may sometimes stop unexpectedly after a few minutes of use, leading to errors like nvlddmkm.sys ID 153, 13, and 0. These crashes tend to happen when multiple demanding 3D tools are active together (for example, using Blender render alongside Substance Painter). Despite these GPU issues, the system remains stable otherwise—no visual glitches appear before or after a crash, and gameplay continues afterward. Certain titles such as Planet Coaster 2 can run smoothly indefinitely if started at boot, though they may occasionally stutter or experience brief GPU dips. Repeated restarts usually worsen the problem. Restarting the computer often helps, but shutting it down completely and rebooting tends to resolve it faster. The recurring issue consistently points to insufficient VRAM availability. Games display errors while Blender shows memory warnings, and ComfyUI logs CUDA access violations. FLUX-fp8 in ComfyUI can run for over 12 hours before problems arise. Rendering tasks can continue indefinitely without other programs, but they’ll stop once another heavy program launches. Blender stays open and renders normally, consuming about 5GB of VRAM versus 8GB reported by Blender itself. I’ve tried reinstalling drivers, ensuring cables are secure, checking storage health, and verifying RAM status. My PC specs include a Ryzen 9 3900x, RTX 3070, and various SSDs. This seems to be a fairly typical concern, but since many similar errors also cause blue or black screens, I wanted to seek advice. Could this relate more to hardware condition or driver problems? Thank you for your time and support!
C
ClareCloud
10-15-2020, 04:53 PM #1

VRAM-heavy applications and software may sometimes stop unexpectedly after a few minutes of use, leading to errors like nvlddmkm.sys ID 153, 13, and 0. These crashes tend to happen when multiple demanding 3D tools are active together (for example, using Blender render alongside Substance Painter). Despite these GPU issues, the system remains stable otherwise—no visual glitches appear before or after a crash, and gameplay continues afterward. Certain titles such as Planet Coaster 2 can run smoothly indefinitely if started at boot, though they may occasionally stutter or experience brief GPU dips. Repeated restarts usually worsen the problem. Restarting the computer often helps, but shutting it down completely and rebooting tends to resolve it faster. The recurring issue consistently points to insufficient VRAM availability. Games display errors while Blender shows memory warnings, and ComfyUI logs CUDA access violations. FLUX-fp8 in ComfyUI can run for over 12 hours before problems arise. Rendering tasks can continue indefinitely without other programs, but they’ll stop once another heavy program launches. Blender stays open and renders normally, consuming about 5GB of VRAM versus 8GB reported by Blender itself. I’ve tried reinstalling drivers, ensuring cables are secure, checking storage health, and verifying RAM status. My PC specs include a Ryzen 9 3900x, RTX 3070, and various SSDs. This seems to be a fairly typical concern, but since many similar errors also cause blue or black screens, I wanted to seek advice. Could this relate more to hardware condition or driver problems? Thank you for your time and support!

W
Waverabbit
Senior Member
643
10-17-2020, 05:53 PM
#2
The issue started when my 2080 began crashing due to insufficient VRAM. Upgrading the card resolved the problem for me.
W
Waverabbit
10-17-2020, 05:53 PM #2

The issue started when my 2080 began crashing due to insufficient VRAM. Upgrading the card resolved the problem for me.

G
gabbylife
Member
228
10-20-2020, 11:38 PM
#3
Thank you for your response. I believe you're correct. After reinstalling Windows, the problem persists. It's unusual since the card works properly unless it's consuming all available VRAM.
G
gabbylife
10-20-2020, 11:38 PM #3

Thank you for your response. I believe you're correct. After reinstalling Windows, the problem persists. It's unusual since the card works properly unless it's consuming all available VRAM.

D
derk4321
Senior Member
482
11-11-2020, 02:25 AM
#4
Yeah, I'm in the same situation. I switched with my wife's card and the problem transferred to her machine. I had to return hers, which means I've been struggling with mine too. It works okay for 2D and everyday tasks, but not for gaming. As soon as it uses more VRAM, the screen loses signal, though the computer keeps running quietly in the background. Of course, this happened right when new cards were released, so I've been holding off on replacing it. Unfortunately, all my gaming is still on the Switch for now.
D
derk4321
11-11-2020, 02:25 AM #4

Yeah, I'm in the same situation. I switched with my wife's card and the problem transferred to her machine. I had to return hers, which means I've been struggling with mine too. It works okay for 2D and everyday tasks, but not for gaming. As soon as it uses more VRAM, the screen loses signal, though the computer keeps running quietly in the background. Of course, this happened right when new cards were released, so I've been holding off on replacing it. Unfortunately, all my gaming is still on the Switch for now.