Nvidia-smi "No devices were found" on Zephyrus GU605 (2025) with Fedora 42 KDE

My work got me a new Zephyrus GU605CW with RTX 5080. I’ve used exclusively Ubuntu for decades, but the first Ubuntu install in this new laptop was struggling with the new hardware. So I thought this would be the opportunity of learning the Fedora landscape, which, if I understood correctly, are more wired towards the latest technology.

It hasn’t been as smooth as I’d hoped, this is now the third time I reinstall Fedora 42 KDE. If your solution is trying a different version, dropping KDE, or get older cuda versions. That’s good too, I just need these drivers to develop on tensorflow/pytorch containers. So here are the steps I followed this last time.

1. Install Fedora 42 KDE from live USB erasing the whole harddrive with default partitions. I selected encrypted drive (needed for work). When it finished I rebooted, but it wouldn’t shut down due to the rtsx_pci bus error, I waited 10min then I shut down by holding the power button. The solution to this is adding pci=noaer to the grub, but on this pass I didn’t bother to stick to solving the nvidia driver issue.

2. When starting for the first time, it showed the welcome dialog and I was able to select third party repositories.

3. First update

$sudo dnf update -y

4. Step number 4 from this guide How to Set Nvidia as Primary GPU on Optimus-based Laptops :: Fedora Docs but using xorg-x11-drv-nvidia-cuda instead of xorg-x11-drv-nvidia

$ sudo dnf install gcc kernel-headers kernel-devel akmod-nvidia xorg-x11-drv-nvidia-cuda xorg-x11-drv-nvidia-libs xorg-x11-drv-nvidia-libs.i686

I waited the 10min, even thoughps aux | grep kmod didn’t show any processes.

5. the guide says this step is to force the configuration to be read.

$ sudo akmods --force

Checking kmods exist for 6.14.0-63.fc42.x86_64 [  OK  ]
Files needed for building modules against kernel
7.14.0-63.fc42.x86_64 could not be found as the following
directories are missing:
/usr/src/kernels/6.14.0-63.fc42.x86_64/
/lib/modules/6.14.0-63.fc42.x86_64/build/ Is the correct kernel-devel package installed? [FAILED]

5.1. As the log asked, I checked anf ound the kernel-devel was not matching the version akmods was using: 6.14.0-63.fc42.x86_64 vs 6.14.11-300.fc42.x86_64.

5.2. Reboot to try to start with the latest kernel installed. I pressed Shift while booting and grub indeed showed both versions.

$ sudo dnf install kernel-devel
Updating and loading repositories:
Repositories loaded.
Package "kernel-devel-6.14.11-300.fc42.x86_64" is already installed.

Nothing to do.

5.3. We try again akmods, this time it works, and Dracut doesn’t fail at least.

$ sudo akmods --force
Checking kmods exist for 6.14.11-300.fc42.x86_64 [  OK  ]
$ sudo dracut --force
(no output, it just waited a bit before showing the prompt again)
$ modinfo -F version nvidia
575.57.08

6. Because of the kernel mismatch I thought to rebuild just in case

$ sudo akmods --kernels $(uname -r) --rebuild
Checking kmods exist for 6.14.11-300.fc42.x86_64 [  OK  ]
Building and installing nvidia-kmod [  OK  ]

7. Reboot

$ nvidia-smi
No devices were found

I have tried every guide and I cannot get the drivers to load properly. Hopefully someone more experienced than me can help me diagnose the issue and get this installation going.

Thank you for reading this far. In case it helps, on my first install when I tried different things I posted in the asus forum. I’ll keep both sides updated if anyone replies

nvidia-smi "No devices were found" on Zephyrus GU6... - Republic of Gamers Forum - 1103351

So, I decided to try again with the open drivers:

$ sudo sh -c 'echo "%_with_kmod_nvidia_open 1" > /etc/rpm/macros.nvidia-kmod'
sudo akmods --kernels $(uname -r) --rebuild
Checking kmods exist for 6.14.11-300.fc42.x86_64 [  OK  ]
Building and installing nvidia-kmod [  OK  ]

And reboot, aaaand nvidia-smi still gets no devices.

Here is a list of guides I have used:

One thing that is starting to bug me is that none of them tests or uses nvidia-smi… Is this not a thing in Fedora? After seeing how all those solutions work for everyone else except me, I must be missing something. If someone is ready to help me I promise to be responsive with any traces you request.

I also looked at sudo dmesg before and after nvidia-smi call, so here is the trace after the call happens. I honestly don’t know how to interpret it:

[  888.730968] nvidia 0000:01:00.0: Enabling HDA controller
[  888.812860] NVRM: testIfDsmSubFunctionEnabled: GPS ACPI DSM called before _acpiDsmSupportedFuncCacheInit subfunction = 11
.
[  890.567787] NVRM: kgspHealthCheck_TU102: ****************************** GSP-CrashCat Report *****************************
**
[  890.567802] NVRM: GPU at PCI:0000:01:00: GPU-add48344-5a39-2942-8c87-5e9e01f28f0b
[  890.567805] NVRM: Xid (PCI:0000:01:00): 120, GSP task exception: load access page fault (cause:0xd) @ pc:0x140c7f6, parti
tion:4#0, task:3
[  890.567812] NVRM:     Reported by libos partition:4#5 kernel v3.1 [0] @ ts:2
[  890.567814] NVRM:     RISC-V CSR State:
[  890.567815] NVRM:         sstatus:0x0000000200000020  sscratch:0xffffffffa30144d0     sie:0x0000000000000220  sip:0x00000
00000000000
[  890.567817] NVRM:         sepc:0x000000000140c7f6     stval:0x0000000000000000  scause:0x000000000000000d
[  890.567818] NVRM:     RISC-V GPR State:
[  890.567818] NVRM:         ra:0x000000000140cea2   sp:0x00000047a9a0f5b0   gp:0x0000000000000000   tp:0x0000000000000000
[  890.567820] NVRM:         a0:0x0000000000000000   a1:0x00000047a2820530   a2:0x0000000000000004   a3:0x00000047aa041000
[  890.567821] NVRM:         a4:0x0000000000000000   a5:0x0000000000000000   a6:0x0000000000001010   a7:0x0000000000000004
[  890.567822] NVRM:         s0:0x00000047a9a0f740   s1:0x00000047a2a44cf0   s2:0x0000000000000002   s3:0x00000000017d4e9a
[  890.567823] NVRM:         s4:0x000000000411bb70   s5:0x00000000001a8000   s6:0x00000047a29805b0   s7:0x0000000000001500
[  890.567824] NVRM:         s8:0x000000000411c088   s9:0x0000000000000000  s10:0x0000000000000000  s11:0x00000047a297e5b0
[  890.567825] NVRM:         t0:0x0000000000000020   t1:0x0000000000000001   t2:0x0000000000000000   t3:0x0000000000000020
[  890.567826] NVRM:         t4:0x0000000000000000   t5:0x00000047a9a0f3c1   t6:0x0000000000000020
[  890.567827] NVRM:     Stack Trace:
[  890.567828] NVRM:         0x000000000140c7f6
[  890.567829] NVRM:         0x00000000017d4e9a
[  890.567829] NVRM:         0x00000000017dd84e
[  890.567830] NVRM:         0x00000000017e001c
[  890.567831] NVRM:         0x00000000017d411a
[  890.567831] NVRM:         0x000000000141629e
[  890.567832] NVRM:         0x0000000001a285da
[  890.567832] NVRM:         0x0000000001a4b698
[  890.567833] NVRM:         0x0000000001b887c0
[  890.567833] NVRM:         0x0000000001a2d332
[  890.567834] NVRM:     Local I/O Register State:
[  890.567835] NVRM:         0x01450800:0x00000000   0x01450900:0xbadf202b   0x01450a00:0x00000000   0x01450c00:0x00000000
[  890.567837] NVRM:         0x01454a00:0x810400d0   0x01454b00:0x010800d0   0x01454c00:0x00080000   0x01400200:0x00000040
[  890.567839] NVRM:     ------------[ end crash report ]------------
[  890.567862] NVRM: GPU0 GSP RPC buffer contains function 4128 (GSP_POST_NOCAT_RECORD) and data 0x0000000000000005 0x000000
00017d4e9a.
[  890.567864] NVRM: GPU0 RPC history (CPU -> GSP):
[  890.567865] NVRM:     entry function                   data0              data1              ts_start           ts_end   
          duration actively_polling
[  890.567866] NVRM:      0    73   SET_REGISTRY          0x0000000000000000 0x0000000000000000 0x0006380235eb6b27 0x0000000
000000000          y
[  890.567870] NVRM:     -1    72   GSP_SET_SYSTEM_INFO   0x0000000000000000 0x0000000000000000 0x0006380235eb6b24 0x0000000
000000000            
[  890.567871] NVRM: GPU0 RPC event history (CPU <- GSP):
[  890.567872] NVRM:     entry function                   data0              data1              ts_start           ts_end   
          duration during_incomplete_rpc
[  890.567873] NVRM:      0    4128 GSP_POST_NOCAT_RECORD 0x0000000000000005 0x00000000017d4e9a 0x0006380235f33a81 0x0006380
235f33a82      1us y
[  890.567879] NVRM: kgspRcAndNotifyAllChannels_IMPL: RC all user channels for critical error 120.
[  890.567891] NVRM: kgspHealthCheck_TU102: ********************************************************************************
**
[  890.567894] NVRM: nvCheckOkFailedNoLog: Check failed: Reset required [NV_ERR_RESET_REQUIRED] (0x00000062) returned from r
pcRecvPoll(pGpu, pRpc, NV_VGPU_MSG_EVENT_GSP_INIT_DONE) @ kernel_gsp.c:4877
[  890.567897] NVRM: nvAssertOkFailedNoLog: Assertion failed: Reset required [NV_ERR_RESET_REQUIRED] (0x00000062) returned f
rom kgspWaitForRmInitDone(pGpu, pKernelGsp) @ kernel_gsp_gh100.c:952
[  890.567908] NVRM: kgspInitRm_IMPL: Max GSP-RM boot attempts exceeded: 4/4
[  890.567975] NVRM: RmInitAdapter: Cannot initialize GSP firmware RM
[  890.570064] NVRM: iovaspaceDestruct_IMPL: 1 left-over mappings in IOVAS 0x100
[  890.570081] NVRM: GPU 0000:01:00.0: RmInitAdapter failed! (0x62:0x62:1941)
[  890.573061] NVRM: GPU 0000:01:00.0: rm_init_adapter failed, device minor number 0

This issue was my savior: RTX 5070 TI Mobile not being detected on latest open driver · Issue #876 · NVIDIA/open-gpu-kernel-modules · GitHub

It seems that 575.57 is the problem, versions 570 and 575.64 were recommended so I followed those steps:

$ sudo dnf config-manager addrepo --from-repofile=https://negativo17.org/repos/fedora-nvidia.repo
sudo dnf remove \*nvidia\* --exclude nvidia-gpu-firmware

Reboot.

Check the versions available sudo dnf list --available --showduplicates akmod-nvidia nvidia-driver

$ sudo dnf install akmod-nvidia-3:575.64-1.fc42 nvidia-driver-3:575.64-1.fc42

Just in case we build for the nth time.

sudo akmods --rebuild

Reboot.

Not yet out of the woods:

$ nvidia-smi
bash: nvidia-smi: command not found...
Packages providing this file are:
'xorg-x11-drv-nvidia-cuda'
'nvidia-driver-cuda'

So we find if those packages exist with this newest and greatest version of nvidia drivers

$ sudo dnf list --available --showduplicates xorg-x11-drv-nvidia-cuda nvidia-driver-cuda
Updating and loading repositories:
Repositories loaded.
Available packages
nvidia-driver-cuda.x86_64       3:575.57.08-1.fc42 fedora-nvidia
nvidia-driver-cuda.x86_64       3:575.64-1.fc42    fedora-nvidia
xorg-x11-drv-nvidia-cuda.x86_64 3:575.57.08-2.fc42 rpmfusion-nonfree-nvidia-driver
$ sudo dnf install nvidia-driver-cuda-3:575.64-1.fc42

Success!

$ nvidia-smi
Fri Jun 20 11:54:03 2025        
+-----------------------------------------------------------------------------------------+
| NVIDIA-SMI 575.64                 Driver Version: 575.64         CUDA Version: 12.9     |
|-----------------------------------------+------------------------+----------------------+
| GPU  Name                 Persistence-M | Bus-Id          Disp.A | Volatile Uncorr. ECC |
| Fan  Temp   Perf          Pwr:Usage/Cap |           Memory-Usage | GPU-Util  Compute M. |
|                                         |                        |               MIG M. |
|=========================================+========================+======================|
|   0  NVIDIA GeForce RTX 5080 ...    Off |   00000000:01:00.0 Off |                  N/A |
| N/A   45C    P8              7W /   80W |       2MiB /  16303MiB |      0%      Default |
|                                         |                        |                  N/A |
+-----------------------------------------+------------------------+----------------------+
                                                                                          
+-----------------------------------------------------------------------------------------+
| Processes:                                                                              |
|  GPU   GI   CI              PID   Type   Process name                        GPU Memory |
|        ID   ID                                                               Usage      |
|=========================================================================================|
|  No running processes found                                                             |
+-----------------------------------------------------------------------------------------+

I hope my pain helps someone

2 Likes

I have yaoshi16ultra-5070ti_laptop, via https://zhuanlan.zhihu.com/p/1904281445544989652 install 575.57 have same question, your method can fix!

1 Like


This Mission Center from flathub

power mode need Balanced or High Perf... can use nvidia gpu, Battery can’t!

Great sharing, save a lot of time for installing 50 series NVidia driver, thank you so much!

If newly installed:

sudo dnf config-manager addrepo --from-repofile=https://negativo17.org/repos/fedora-nvidia.repo
sudo dnf install akmod-nvidia-3:575.64-1.fc42 nvidia-driver-3:575.64-1.fc42
sudo dnf install nvidia-driver-cuda-3:575.64-1.fc42
nvidia-smi

if download rate very low, can use proxy, if you have. sudo nano /etc/dnf/dnf.conf:
format, use your proxy ip and port

[main]
proxy=http://123.123.123.123:1234

I’ve noticed that many Steam games have issues—some won’t launch, while others have multiplayer problems (either failing to find servers or freezing when joining matches). I can only play single-player or offline mode. These games used to work fine on my AMD RX 5600 XT and AMD iGPU.

on linux steam titanfall2 no server, need in titanfall2/properties/general/launch options add OPENSSL_ia32cap=~0x20000000 %command%!

now 575.64 have.

sudo dnf install https://download1.rpmfusion.org/free/fedora/rpmfusion-free-release-$(rpm -E %fedora).noarch.rpm 
sudo dnf install https://download1.rpmfusion.org/nonfree/fedora/rpmfusion-nonfree-release-$(rpm -E %fedora).noarch.rpm
sudo dnf install akmod-nvidia nvidia-driver
nvidia-smi

flatpak steam or other app need:

flatpak update

AFAIK the nvidia-driver package is not required for anything when installing from the rpmfusion repos. (though that may be an alias!)

$ sudo dnf install nvidia-driver
Updating and loading repositories:
Repositories loaded.
Package "xorg-x11-drv-nvidia-cuda-3:575.64-1.fc42.x86_64" is already installed.

Nothing to do.

The nvidia-smi command is installed when you install the xorg-x11-drv-nvidia-cuda package

$ sudo dnf provides */nvidia-smi
Updating and loading repositories:
Repositories loaded.
xorg-x11-drv-nvidia-cuda-3:575.64-1.fc42.x86_64 : CUDA driver for xorg-x11-drv-nvidia
Repo         : @System
Matched From : 
Filename     : /usr/bin/nvidia-smi

only install akmod-nvidia can’t use nvidia-smi. yes, is alias

As stated, if the xorg-x11-drv-nvidia-cuda package is not installed then the nvidia-smi command is not available.