Here is the bug report nvidia-bug-report.log.gz (123.8 KB)
I have been running into the same issue! My machine will boot and work with my nvidia card for ages, and then something will happen where it’s like the card is effectively gone. It first happened because my machine booted without the TV being plugged into the computer. It also has happened where a video game crashes and causes nvidia errors in dmesg (which sadly I haven’t recorded.) After the game crashes and I reboot the machine, I only see the intel card. I’ve tried reinstalling dkms and rebuilding initramfs (which installation of the driver should do by default.) Anyway I attached the bug report log.nvidia-bug-report.log.gz (65.1 KB)
(bizarrely I was able to fix my issue by moving the hdmi cable from the nvidia card to the built in hdmi, rebooting, switching back, and rebooting again. Pretty sure this is a hardware issue.)
There was no sign of any nvidia device in your logs, so this really looks like a hardware issue (or some setting in bios)
i have a similar problem. nvidia-smi us not working. i am working on a project which uses cuda cudnn and it says nvidia not detected. i have attached bug report file for refence. thank you for helping me outnvidia-bug-report.log (1.8 MB)
Hi generix, I just installed Nvidia rtx 3090 graphic card from Software and Updates utility. The driver is properly installed but the system does not detect it. Can you offer any help?
I am getting a error while installing the driver from the Ubuntu’s Software & Updates application.
I have did a dual boot for Ubuntu with windows 10.
I have similar problem on my server. 2080 working, but 3090 not. I will be happy if you provide any ideas about it.
nvidia-bug-report.log.gz (485.7 KB)
I can only see two 2080 ti on the pcie bus, both working.
In case there should be also a 3090, please check power connectors, reseat the card, check if it works in another system.
I have the same problem on my ubuntu 20. My GPU is 3070ti laptop, Here is my resulting file
nvidia-bug-report.log.gz (132.2 KB)
Same problem here on the following system:
- Thinkpad P14s Gen2 with T500
(note: GPU can’t be disabled or explicitly switched in Bios) - (K)Ubuntu 22.04.1 with kernel Linux 5.15.0-58-generic
The Nvidia was working at some point, but has now vanished (unclear what triggered that). Clear sign that it is still powered all the time is the fan/battery drain.
Debugging info from what I have seen in this thread:
$ nvidia-smi
NVIDIA-SMI has failed because it couldn’t communicate with the NVIDIA driver. […]
$ lspci | egrep -i “(vga|nvidia|3d)”
00:02.0 VGA compatible controller: Intel Corporation TigerLake-LP GT2 [Iris Xe Graphics] (rev 01)
$ dkms status
bbswitch/0.8, 5.15.0-58-generic, x86_64: installed
nvidia/525.60.13, 5.15.0-58-generic, x86_64: installed
$ grep 10de /lib/udev/rules.d/*
/lib/udev/rules.d/71-nvidia.rules:SUBSYSTEM==“pci”, ATTRS{vendor}==“0x10de”, DRIVERS==“nvidia”, TAG+=“seat”, TAG+=“master-of-seat”
I would like to attach the log output of “nvidia-bug-report.sh”, but this just hangs in the “Processing upload” stage, so I could send it via e-mail or attach it later if it works then.
As far as I understand, no definite solution to this has so far been provided in this thread, so any input is appreciated. Thanks.
Hi, my computer’s GPU is NVIDIA GeForce RTX 2070 super, in powershell input “nividia-smi”, output nvidia driver version GPU memory, process etc. But when I’m using the ubuntu 20.04 LTS finding GPU not display anything. can’t detecing GPU. how can i reslove this please help! thx
I’m having the same issue after updating to 5.4.0-164-generic
here is the log file.
nvidia-bug-report.log (195.2 KB)
Hi, @generix .
Previously, one Nvidia 3090 card was intalled on my Ubuntu 22.04 and it worked well.
After I installed a second 3090 card (no NV link installed), I just got an error when excuting command nvidia-smi:
$ nvidia-smi
NVIDIA-SMI has failed because it couldn't communicate with the NVIDIA driver. Make sure that the latest NVIDIA driver is installed and running.
Trying to reinstall nvidia driver several times, the output of command nvidia-smi were still the ‘failed one’ showed above.
The command lspci | egrep -i "vga|display|3d" output nothing, so it seems that the two 3090 cards are not identified.
Here is bug log:
nvidia-bug-report.log.gz (107.3 KB) .
Please help. Thanks.
There are no nvidia gpus detected as you already know. Since the kernel doesn’t see them, they’re not detected at the bios level. Please check power connectors, try reseating them in the pcie slot, check if “Above 4G decoding/RBAR” is enabled in bios.
Adding some connectors or wires works.
Thank you!
Hello everyone,
Same here,
nvidia-bug-report.log.gz (265.1 KB)
I’m hitting a massive roadblock trying to get the dedicated graphics card working on a legacy laptop I recently acquired, and I’m hoping some of the hardware/driver experts here might have run into this specific error behavior before.
The Setup:
Laptop: Lenovo Yoga 500-14ISK
CPU: Intel Core i5-6200U
RAM: 8GB
Integrated Graphics: Intel HD Graphics 520
Dedicated GPU: NVIDIA GeForce 920M
The Problem: The 920M simply refuses to initialize. The driver briefly worked a single time during a series of manual PCI device driver updates on Windows, but after a single reboot, it dropped again. It consistently fails to start across multiple clean operating systems.
What I’ve Tried & Technical Specifics:
BIOS: Secure Boot is completely disabled to rule out driver signature blocking.
Windows 10: Clean install (activated via embedded license). The card throws errors in Device Manager (Error 43 / Error 14 layout behavior).
Linux (Ubuntu 22.04 & 26.04): On Ubuntu 22.04, I downgraded from the stock Kernel 6.x down to Kernel 5.15 to try matching the hardware era better. I’ve tested both the NVIDIA 470 and 390 proprietary drivers.
The Logs (dmesg): In Linux, I consistently hit severe initialization loops right at boot: NVRM: GPU 0000:01:00.0: RmInitAdapter failed! (0x22:0x56:667) NVRM: rm_init_adapter failed, device minor number 0
Root Cause Clue: The logs indicate failures to copy the VBIOS to system memory, meaning the driver crashes hard right as it tries to negotiate power states via ACPI or handle early PCIe bus handshakes. I’ve tried forcing modprobe options like NVreg_DynamicPowerManagement=0x00 and NVreg_EnableMSI=0, but the adapter fails before it even reads them.
Because Lenovo has marked this model as End of Service, official recovery media isn’t an option to check the factory state.
Has anyone successfully dealt with this specific 0x22:0x56:667 initialization failure on legacy mobile GPUs? Is this a known ACPI/BIOS memory mapping conflict with newer modern OS kernels, or is there a specific legacy driver/kernel/grub parameter combination required to coax this firmware to initialize correctly?
I included an nvidia-bug-report.log.gz in this link:
Any insights or technical pointers would be massively appreciated. Thanks!