You are not logged in.
Pages: 1
Hello community,
I am the owner of a 9510 XPS laptop , which comes with both an integrated GPU (Intel) and a GTX 3050 Ti (Nvidia)
While I was checking to push a bit further the machine I realized that I cannot have the Nvidia running.
Here are some output information:
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.
xrandr --listproviders
Providers: number : 1
Provider 0: id: 0x46 cap: 0xf, Source Output, Sink Output, Source Offload, Sink Offload crtcs: 4 outputs: 4 associated providers: 0 name:modesettingI installed everything following the NVIDIA Link
I also have no detection of GPU from nvidia-xconfig
nvidia-xconfig
WARNING: Unable to locate/open X configuration file.
WARNING: Unable to parse X.Org version string.
ERROR: Unable to find any GPUs in the system.
ERROR: Unable to write to directory '/etc/X11'.I made sure I had not blacklisted anything with the name "nvidia" in /etc/modprobe.d/blacklist.conf
lspci -k | grep -A 2 -E "(VGA|3D)"
00:02.0 VGA compatible controller: Intel Corporation TigerLake-H GT1 [UHD Graphics] (rev 01)
DeviceName: Onboard - Video
Subsystem: Dell TigerLake-H GT1 [UHD Graphics]
--
01:00.0 3D controller: NVIDIA Corporation GA107M [GeForce RTX 3050 Ti Mobile] (rev a1)
Subsystem: Dell GA107M [GeForce RTX 3050 Ti Mobile]
Kernel modules: nouveau, nvidia_drm, nvidiaAs I have a NV170 model, I installed the Turing series option with the nvidia-open package
I tried installing proprietary drivers nvidia with utils and tools, but same result. I just installed also the xf86-vide-nouveau and similar result.
lspci -k | grep -A 2 -E "(VGA|3D)" SIGPIPE 1
00:02.0 VGA compatible controller: Intel Corporation TigerLake-H GT1 [UHD Graphics] (rev 01)
DeviceName: Onboard - Video
Subsystem: Dell TigerLake-H GT1 [UHD Graphics]
--
01:00.0 3D controller: NVIDIA Corporation GA107M [GeForce RTX 3050 Ti Mobile] (rev a1)
Subsystem: Dell GA107M [GeForce RTX 3050 Ti Mobile]
Kernel modules: nouveauI also tried to blacklist nouveau from the grub config file, no change.
I am not really sure where to check again, what elements are missing or to be done?
Last edited by archline (2024-02-28 19:12:33)
Offline
Did you follow all 6 steps in the link you gave? Every one of them is important.
Offline
Thanks for quick reply!
I believe so, here is a recap:
Step 1:
RTX 3050TI
Step2:
NV170
Step3:
I went fo the NV160 version model, with nvidia-open package
Step 4:
I installed lib32-nvidia-utils
Step 5:
I removed kms from the HOOKS in /etc/mkinitcpio.con
I regenerated with
mkinitcpio -PStep 6:
I rebooted.
Then I tried again with the nvidia package instead of the nvidia-open package. Same.
Last edited by archline (2024-02-27 21:14:06)
Offline
I also have no detection of GPU from nvidia-xconfig
Don't run that!
(It produces a lot of cruft and an unnecessary static server config that will come back to bite you later)
Please post your complete system journal for the boot:
sudo journalctl -b | curl -F 'file=@-' 0x0.stOffline
Noted, thank you.
Here is the result:
https://0x0.st/HR1S.txt
Offline
Feb 27 22:04:34 Archlinux-XPS15 systemd-modules-load[298]: Error running install command '/bin/false' for module nvidia: retcode 1You blacklisted the nvidia module?
https://wiki.archlinux.org/title/Kernel … probe.d/_2
Offline
Alright I realize that I have an old file in the modprobe.
The following one:
install nvidia /bin/false
I do not remember exactly why. so I removed it and ran mkinitcpio -P, but it crashed at rebooted.
I could log into Arch but could not launch X anymore, with coredump and fatal errors
here is the output of the failing boot session:
http://0x0.st/HR1x.txt
Last edited by archline (2024-02-28 00:04:59)
Offline
#7 0x00007364c26c58b0 n/a (libnvidia-gpucomp.so.545.29.06 + 0x4c58b0)
#8 0x00007364adceb9eb n/a (libnvidia-eglcore.so.550.54.14 + 0xaeb9eb)Somehow you have different versions of these installed, when both should be coming from the nvidia-utils package. Reinstall that. Did you, at some point, run the installer from the nvidia site?
Offline
Did you, at some point, run the installer from the nvidia site?
Yes, since it is one of the option presented in the wiki.
However, the installer did not execute entirely and ended in an error as well, I gave up on that one since the wiki says this solution is to avoid and we should privilege the pacman option particularly for update purposes..
#7 0x00007364c26c58b0 n/a (libnvidia-gpucomp.so.545.29.06 + 0x4c58b0) #8 0x00007364adceb9eb n/a (libnvidia-eglcore.so.550.54.14 + 0xaeb9eb)Somehow you have different versions of these installed, when both should be coming from the nvidia-utils package. Reinstall that.
What would be the cleanest way to reinstall it?
sudo pacman -Rsn nvidia-utils nvidia-prime nvidia-open These are all the "nvidia" packages installed, and re-install them?
Issue is, I am unsure it would delete all those that were installed by the nvidia*.run from the official website.
I already installed the ones above after trying to execute the official script from nvidia.
Offline
Warning: Avoid installing the NVIDIA driver through the package provided from the NVIDIA website. Installation through pacman allows upgrading the driver together with the rest of the system.
The wiki just says to use the website to look up what driver version is appropriate for your GPU if you only know the marketing name…
I already installed the ones above after trying to execute the official script from nvidia.
And you did not run into file conflicts?
Either way, the repo version updated to 550xx, tehre should be no 545xx versions around.
Try "pacman -Syu" first, pay attention to any errors and
Feb 28 00:53:05 Archlinux-XPS15 sudo[1255]: archline : TTY=tty1 ; PWD=/home/archline ; USER=root ; COMMAND=/usr/bin/mv /etc/modprobe.d/disable-nvidia.conf.bak /etc/modprobe.d/disable-nvidia.confremove that file.
Then post an updated journal
Offline
Things are now working properly, I tried an update before but I guess I messed up somewhere.
For the details:
I tried a "pacman -Syu" which returned some conflicts and failed.
The list comes from nvidia-utils.
error: failed to commit transaction (conflicting files)
nvidia-utils: /usr/lib/firmware/nvidia/550.54.14/gsp_ga10x.bin exists in filesystem
nvidia-utils: /usr/lib/firmware/nvidia/550.54.14/gsp_tu10x.bin exists in filesystem
nvidia-utils: /usr/lib/libEGL_nvidia.so.550.54.14 exists in filesystem
nvidia-utils: /usr/lib/libGLESv1_CM_nvidia.so.550.54.14 exists in filesystem
nvidia-utils: /usr/lib/libGLESv2_nvidia.so.550.54.14 exists in filesystem
nvidia-utils: /usr/lib/libGLX_nvidia.so.550.54.14 exists in filesystem
nvidia-utils: /usr/lib/libcuda.so.550.54.14 exists in filesystem
nvidia-utils: /usr/lib/libcudadebugger.so.550.54.14 exists in filesystem
nvidia-utils: /usr/lib/libnvcuvid.so.550.54.14 exists in filesystem
nvidia-utils: /usr/lib/libnvidia-allocator.so.550.54.14 exists in filesystem
nvidia-utils: /usr/lib/libnvidia-cfg.so.550.54.14 exists in filesystem
nvidia-utils: /usr/lib/libnvidia-egl-gbm.so.1.1.1 exists in filesystem
nvidia-utils: /usr/lib/libnvidia-eglcore.so.550.54.14 exists in filesystem
nvidia-utils: /usr/lib/libnvidia-encode.so.550.54.14 exists in filesystem
nvidia-utils: /usr/lib/libnvidia-fbc.so.550.54.14 exists in filesystem
nvidia-utils: /usr/lib/libnvidia-glcore.so.550.54.14 exists in filesystem
nvidia-utils: /usr/lib/libnvidia-glsi.so.550.54.14 exists in filesystem
nvidia-utils: /usr/lib/libnvidia-glvkspirv.so.550.54.14 exists in filesystem
nvidia-utils: /usr/lib/libnvidia-gpucomp.so.550.54.14 exists in filesystem
nvidia-utils: /usr/lib/libnvidia-ml.so.550.54.14 exists in filesystem
nvidia-utils: /usr/lib/libnvidia-ngx.so.550.54.14 exists in filesystem
nvidia-utils: /usr/lib/libnvidia-nvvm.so.550.54.14 exists in filesystem
nvidia-utils: /usr/lib/libnvidia-opticalflow.so.550.54.14 exists in filesystem
nvidia-utils: /usr/lib/libnvidia-pkcs11-openssl3.so.550.54.14 exists in filesystem
nvidia-utils: /usr/lib/libnvidia-pkcs11.so.550.54.14 exists in filesystem
nvidia-utils: /usr/lib/libnvidia-ptxjitcompiler.so.550.54.14 exists in filesystem
nvidia-utils: /usr/lib/libnvidia-rtcore.so.550.54.14 exists in filesystem
nvidia-utils: /usr/lib/libnvidia-tls.so.550.54.14 exists in filesystem
nvidia-utils: /usr/lib/libnvoptix.so.550.54.14 exists in filesystem
nvidia-utils: /usr/lib/vdpau/libvdpau_nvidia.so.550.54.14 exists in filesystem
nvidia-utils: /usr/share/nvidia/nvidia-application-profiles-550.54.14-key-documentation exists in filesystem
nvidia-utils: /usr/share/nvidia/nvidia-application-profiles-550.54.14-rc exists in filesystem
opencl-nvidia: /usr/lib/libnvidia-opencl.so.550.54.14 exists in filesystemI then followed the steps in the official Wiki: Pacman failed to commit transaction (conflicting files)
I checked all of the files in conflict manually with "pacman -Qo /pathToFile"
=> It returned everytime that no packages owned them, so I just renamed them all, as explained in the above link.
I then could make my update without any error.
At reboot I had a faillure at the launch of X, I followed the errors which was from a file in the /etc/X11/xorg.conf.d that was supposed to help for the dual cards (nvidia + internal)
the content was taken from Wiki's NVIDIA Optimus info:
/etc/X11/xorg.conf.d/10-nvidia-drm-outputclass.conf
Section "OutputClass"
Identifier "intel"
MatchDriver "i915"
Driver "modesetting"
EndSection
Section "OutputClass"
Identifier "nvidia"
MatchDriver "nvidia-drm"
Driver "nvidia"
Option "AllowEmptyInitialConfiguration"
Option "PrimaryGPU" "yes"
ModulePath "/usr/lib/nvidia/xorg"
ModulePath "/usr/lib/xorg/modules"
EndSectionThe output of "Nvidia-smi" is now:
nvidia-smi main
Wed Feb 28 11:15:17 2024
+-----------------------------------------------------------------------------------------+
| NVIDIA-SMI 550.54.14 Driver Version: 550.54.14 CUDA Version: 12.4 |
|-----------------------------------------+------------------------+----------------------+
| 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 3050 ... Off | 00000000:01:00.0 Off | N/A |
| N/A 60C P0 12W / 40W | 13MiB / 4096MiB | 0% Default |
| | | N/A |
+-----------------------------------------+------------------------+----------------------+
+-----------------------------------------------------------------------------------------+
| Processes: |
| GPU GI CI PID Type Process name GPU Memory |
| ID ID Usage |
|=========================================================================================|
| 0 N/A N/A 1642 G /usr/lib/Xorg 4MiB |
+-----------------------------------------------------------------------------------------+and to xrandr --listproviders
Providers: number : 2
Provider 0: id: 0x48 cap: 0xb, Source Output, Sink Output, Sink Offload crtcs: 5 outputs: 5 associated providers: 0 name:Intel
Provider 1: id: 0x1e7 cap: 0x0 crtcs: 0 outputs: 0 associated providers: 0 name:NVIDIA-G0FYI the latest journal: http://0x0.st/HRLe.txt
Thank you!
Can we consider it solved or are there a few extra checks to make?
Last edited by archline (2024-02-28 10:17:36)
Offline
No, you can mark the thread as solved.
Offline
Pages: 1