since some days dkms-nvidia/openSUSE_13.1 folder has disappeared
at
http://download.opensuse.org/repositories/home:/Knurpht:/
what happens ?
thanks
Hi, I removed the packages for a couple of reasons:
- No time to maintain them
- The bumblebee project moved to the main repos.
- Severe issues with the driver and newer kernels for which I didn’t and don’t have time either
But, you could build your own. In fact the dkms-nvidia package was one of my first attempts.
- The bumblebee project moved to the main repos.
can you explain please
do you mean i can use bumblebee packet instead ?
thanks
Bumblebee lets you use either the Intel or NVIDIA chips as you want and you can have the NVIDIA driver installed. This only works if you have a Intel+NVIDIA machine. IMHO Optimus is a Rube Goldberg solution. The problem is people want high end graphics on Laptops but they also want long battery life. Instead of designing a single chip that will auto scale they put Intel and NVIDIA both in the hardware. And let you switch from slow energy efficient to fast power drainer mode
It is important that you install things in the right order from the right repos
i understand i can’t use dkms-nvidia packet from bumblebee repo in my case . Because in my pc there is only a nvidia graphical adapter . right ?
Right. Bumblebee is only for hybrid intel+nvidia systems. It’s main purpose is to switch between the intel and the nvidia GPU.
Why aren’t you just using the official nvidia driver packages from the official nvidia repo?
They only work with the kernel shipped with openSUSE though, so you cannot use them with Kernel:stable f.e.
Why aren’t you just using the official nvidia driver packages from the official nvidia repo?
because :
- i don’t want to install again nvidia driver when opensue updates from 3.11.x to 3.11.x+1
- cuda is well installed in the contrary of official repo driver
That’s one of the main points of using the nvidia packages from the official repo.
You do not have to install them again when a kernel update is released.
You can of course setup dkms yourself and use the driver from nvidia’s homepage though. A package is available on Packman f.e.
- cuda is well installed in the contrary of official repo driver
In the meantime CUDA should work fine in the “official repo driver”.
But I haven’t tested it myself.
If there is still a problem, somebody should reopen the bug report though:
https://bugzilla.novell.com/show_bug.cgi?id=879767
ok i will try again official repo driver
thanks to all
Ok, but please make sure that you install the corresponding nvidia-gfxG03-kmp and nvidia-uvm-gfxG03-kmp packages for your kernel (both are needed for CUDA).
Especially if you already uninstalled those packages once, YaST/zypper might select the wrong ones.
And nvidia-glG03 might not get installed automatically as well if you uninstalled it already. But that’s needed for 3D acceleration (GLX).
Another note:
The driver might not work on the first reboot after installation. If that’s the case, simply reboot and it should work then.
only dkms-nvidia was installed
i installed nvidia with one click method
then
nvidia-computeG03
nvidia-gfxG03-kmp-desktop
nvidia-glG03
nvidia-uvm-gfxG03-kmp-default
x11-video-nvidiaG03
was installed
result :
-
at first boot driver is well loaded and i open a kde session without pb
-
in nvidia settings in gui no more menu item “vdpau”
in konsole with vdpauinfo i can check vdpau is ok -
i don’t remember if there was a menu item “cuda”
in app fahcontrol (scientific app) i have “cuda not detected” -
vaapi pb : xdg error
linux-b4lz:~ # vainfo
error: XDG_RUNTIME_DIR not set in the environment.
libva info: VA-API version 0.34.0
libva info: va_getDriverName() returns 0
libva info: Trying to open /usr/lib64/dri/nvidia_drv_video.so
libva info: Found init function __vaDriverInit_0_34
libva info: va_openDriver() returns 0
vainfo: VA-API version: 0.34 (libva 1.2.1)
vainfo: Driver version: Splitted-Desktop Systems VDPAU backend for VA-API - 0.7.4
vainfo: Supported profile and entrypoints
VAProfileMPEG2Simple : VAEntrypointVLD
VAProfileMPEG2Main : VAEntrypointVLD
VAProfileMPEG4Simple : VAEntrypointVLD
VAProfileMPEG4AdvancedSimple : VAEntrypointVLD
VAProfileH264Main : VAEntrypointVLD
VAProfileH264High : VAEntrypointVLD
VAProfileVC1Simple : VAEntrypointVLD
VAProfileVC1Main : VAEntrypointVLD
VAProfileVC1Advanced : VAEntrypointVLD
linux-b4lz:~ #
plus
by default monitor frequency is set to auto thus leading to 60 Hz
the optimum for the philips 170C5 is 75 Hz
then
i must manualy set frequency to 75 Hz
You should uninstall that before installing the nvidia driver packages.
i installed nvidia with one click method
then
nvidia-computeG03
nvidia-gfxG03-kmp-desktop
nvidia-glG03
nvidia-uvm-gfxG03-kmp-default
x11-video-nvidiaG03
was installed
Do not install both nvidia-gfxG03-kmp-default and nvidia-gfxG03-kmp-desktop.
Only install the appropriate one for your kernel.
Apparently you have both kernel-desktop and kernel-default installed now, so remove one as well.
I’d suggest to remove kernel-default, but then you would have to install nvidia-uvm-gfxG03-kmp-desktop as you are missing that at the moment.
So apparently CUDA is not working?
Which kernel are you using now?
As you installed nvidia-uvm-gfxG03-kmp-default CUDA will only be working with kernel-default.
Normally the driver should automatically detect the monitor’s optimum frequency, i.e. it uses what the monitor tells it to use.
So maybe the monitor reports the wrong one?
But for a TFT display it doesn’t matter anyway whether you use 60Hz or 75Hz, as there’s no flicker.
Mine reports 60Hz as default as well, btw.
Do not install both nvidia-gfxG03-kmp-default and nvidia-gfxG03-kmp-desktop.
then something is wrong : one click script or anything else
kernel-default-devel , kernel-desktop , kernel-desktop-devel , kernel-devel , kernel-source , kernel-syms , kernel-xen-devel are installed
no kernel-default is installed
i did not installed myself any kernel package , perhaps dkms-nvidia or kde crash report or anything else ?
No. It’s perfectly legal to have both installed. And that’s needed if you want to have bnth kernel-default and kernel-desktop.
But it’s unnecessary.
kernel-default-devel , kernel-desktop , kernel-desktop-devel , kernel-devel , kernel-source , kernel-syms , kernel-xen-devel are installed
no kernel-default is installed
If you have nvidia-gfxG03-kmp-default and/or nvidia-uvm-gfxG03-kmp-default installed as you wrote, you must have kernel-default as well as they require it via package dependencies.
So which packages do you have really installed now?
rpm -qa | egrep "(nvidia|kernel)"
i did not installed myself any kernel package , perhaps dkms-nvidia or kde crash report or anything else ?
What does that mean?
Please uninstall dkms-nvidia as I said, before you install the nvidia packages.
Post the output of the above, then I’ll provide you with better instructions.
Ah, sorry.
I got confused, you do not have both installed.
But you have nvidia-uvm-gfxG03-kmp-default instead of nvidia-uvm-gfxG03-kmp-desktop. Remove that and install nvidia-uvm-gfxG03-kmp-desktop instead.
This is because nvidia-gfxG03-kmp just recommends any nvidia-uvm-gfxG03-kmp package and libzypp’s solver sometimes picks the wrong one.
This has been improved already (nvidia-gfxG03-kmp recommends nvidia-uvm-gfxG03-kmp-desktop explicitely), but the new packages are not in the repo yet.
But, as I said, you should have removed dkms-nvidia first.
It will always give problems sooner or later if you install the same driver at the same time by different means, or different versions of the driver simultanously.
So please remove dkms-nvidia, and all those nvidia packages, then enter YaST again (or use zypper) to install only those packages:
nvidia-computeG03
nvidia-gfxG03-kmp-desktop
nvidia-glG03
nvidia-uvm-gfxG03-kmp-desktop
x11-video-nvidiaG03
i uninstalled nvidia-uvm-gfxG03-kmp-default . this automatically installs nvidia-uvm-gfxG03-kmp-desktop
Please uninstall dkms-nvidia as I said, before you install the nvidia packages.
i did this before using one click installation
linux-b4lz:~ # rpm -qa | egrep “(nvidia|kernel)”
nvidia-gfxG03-kmp-desktop-331.79_k3.11.6_4-27.1.x86_64
kernel-default-devel-3.11.10-7.1.x86_64
kernel-desktop-3.11.10-11.1.x86_64
kernel-xen-devel-3.11.10-7.1.x86_64
x11-video-nvidiaG03-331.79-27.1.x86_64
kernel-xen-devel-3.11.10-11.1.x86_64
kernel-firmware-20130714git-2.17.1.noarch
kernel-devel-3.11.10-7.1.noarch
kernel-default-devel-3.11.10-11.1.x86_64
kernel-syms-3.11.10-7.1.x86_64
kernel-devel-3.11.10-11.1.noarch
kernel-desktop-3.11.6-3.1.x86_64
nvidia-glG03-331.79-27.1.x86_64
kernel-source-3.11.10-7.1.noarch
kernel-desktop-3.11.10-7.1.x86_64
nvidia-uvm-gfxG03-kmp-desktop-331.79_k3.11.6_4-27.1.x86_64
kernel-source-3.11.10-11.1.noarch
patterns-openSUSE-devel_kernel-13.1-13.6.1.x86_64
kernel-desktop-devel-3.11.10-11.1.x86_64
nvidia-computeG03-331.79-27.1.x86_64
kernel-syms-3.11.10-11.1.x86_64
kernel-desktop-devel-3.11.10-7.1.x86_64
linux-b4lz:~ #
Ok.
i did this before using one click installation
Then I must have misunderstood you. I thought you said it is installed.
linux-b4lz:~ # rpm -qa | egrep “(nvidia|kernel)”
Ok, this looks good.
So please reboot, and see if it works.
If not, you could try to add your user to the “video” group if you didn’t do so already.
sudo usermod -a -G video $USER
But this shouldn’t be needed if your system is up-to-date.
i reboot for the second time
then
cuda not detected
in nvidia setting gui no menu item “vdpau”
and still xdg pb with vaapi
my system is up to date .
i will try “sudo usermod -a -G video $USER”
i executed “sudo usermod -a -G video $USER”
then i reboot
then pb still there :
cuda not detected
in nvidia setting gui no menu item “vdpau”
and still xdg pb with vaapi