The system is experiencing frequent crashes due to nvlddmkm.sys issues.
The system is experiencing frequent crashes due to nvlddmkm.sys issues.
I understand you're facing the same issue and are unsure about the next steps. Let's work through it together.
I'm experiencing the same challenges and believe some details could assist you. I created an account to track this. My issue began when the DP port on my monitor malfunctioned, causing intermittent connectivity issues that reduced bandwidth slightly. This happened after previous problems with unstable cables and mixed refresh/resolution multi-monitor configurations. The more components I added, the more sensitive the driver became. Recently, I switched to an Intel-based system with an integrated GPU handling three monitors, while the 2080 manages the main display with the problematic DP port. By rearranging these elements, I noticed the DP signal occasionally fails, likely affecting crash triggers in my games. I also encountered this with HDMI ports, though not specifically DP-related. It's hard to determine if Gsync matters or if the 10-bit setting plays a role, or if it simply lowers the bandwidth needed. Still, this might help narrow things down. As always, try simplifying variables until the problem stops, record changes, and eliminate other factors. If needed, I'm open to hearing about similar cases. P.S.: A new GPU setup means rechecking connections, which has boosted my confidence. If anyone spots a connection or bandwidth issue, I'd love to hear about it!
Experiencing the same problem on a 4090 FE with a 1000W PSU (z690 board). This appears only during RT workloads. Initially assumed it was related to power consumption, but testing at full GPU load (around 700-750W total system draw) showed no issues. However, with RT tasks—such as Witcher 3, Control, Cyberpunk 2077, or Forza Horizon 5—the GPU usage drops to 30-50% (about 400-500W), triggering the nvlddmkmsys error. The system crashes regardless of the game. Native DX12 runs for TimeSpy (Normal/Extreme) also fail under heavy artifacting, similar to overclocking beyond stable limits. Some suggest a PCI-E bus fault. With five NVME drives on an ASUS add-in card, I’m wondering if that’s the cause. Reducing GPU speed by 300-500MHz and setting power limits to 75% resolves most issues, though occasional crashes remain. If underclocking by 400-500MHz and limiting power to 75%, TimeSpy Extreme runs smoothly without artifacts. For those facing this, could you share your motherboard chipset/model? I’ve tried several fixes without success: disabling XMP, lowering resolution, swapping PSU, altering cable connections, modifying PCI-E settings in BIOS, and adjusting bus speed. I have a Corsair HX1000i PSU and plan to use their native 12+4pin 12VHPWR cables tomorrow to test. Switching to a Z790 motherboard might help too. System specs: 12900kf, 32GB DDR4, 6000MHz, 4090 FE Z690-E, Gaming WiFi, NVME Gen4 (4TB), Gen3 (1TB).
Created an account to share my experience with similar problems. I'm using a Zotac 4090 Trinity OC. I first encountered this issue about two weeks ago after switching off HDR while playing Tarkov for a few hours. I've tried almost everything—DDU, Vbios, Diff Ram, Mobo, Psu, driver updates, and different BIOS versions. It's outside my return window, but I'm likely to need an RMA. Have you come across a fix? Thanks!
I encountered a problem with my previous response not appearing. I attempted to replace the Nvidia Squid power adapter with a Corsair 12+4 pin 12V cable, but the issue persisted. This suggests the power adapter isn't the cause. It seems the problem might be related to a z690 board or a faulty card. Many users are reporting similar failures, which is concerning given the high failure rate. I'm curious if anyone with a z790 board experiencing this issue has encountered it, especially since real-world testing isn't being conducted. I have a z790 board ordered and will consider returning it if the problem continues.
also made an account to hop in and say i am also running into the same issues in event viewer. What happens to me is that both my monitors will lose conenctiion to the gpu but the power will stay on. but if i dont turn my pc off quick enough when i go to reboot the mobo status light for gpu issues will light up (and pc will be unusable for like 30 minutes and it has to stay off or it wont fix ) If i do turn it off like quickly itll let me boot again. I had this with a 3070 ti and all the parts below and even before all the "new" parts it was happening with the 3070ti in an old rig, upgraded cpu mobo ram psu and it was still occuring, i rma'd the card right before chrismas and it happened again even witth the new one, i traveled to family till now and browsed forums all holiday season and found that maybe xmp on was the issue. turned xmp off yesterday and it seemed fine, but today when i reconnected the sata drive, games would crash again. today after it happened with the sata drive connceted, but not yesterday i thought that it could be power related. I am not sure if it could be a power related issue but the crashes dont fully turn off my pc which leads me to think otherwise.\so i unplugged my sata drive, but after the crash i was actually able to see event viewer errors, and they had the same info as OP. i found this thread because of the event viewer line, and saw that someone said they can always force the crash using 3dmark, i ran timespy and it crashed my comp. i waited to turn off pc so it could potentially log as much as possible and gpu mobo light is on and hasnt fixed in like 45 minutes. hoping theres any advice, im at such a loss for this issue edit: added picture of the most common crash, and a bsod i recieved after the new card it would either do this, or just lose connection to both monitors and the pc would just stay on unless i held down the power button ( changed everything except gpu in october rma'd gpu in december) new z690 aorus elite ax new i7-12700kf new corsair h100i elite 3 corsair af 120 fans new corsair 32gb 5600 ddr5 new evga supernova p6 750watt psu evga ftw3 oc 3070 ti (rma'd already) 2 tb m2 ssd 2b sata ssd not plugged in 2 tb sata hd not plugged in