Home » Repos » General chat » CentOS 7 and updates
CentOS 7 and updates [message #3527] |
Mon, 14 March 2016 14:16  |
 |
Halfloaf
Messages: 2 Registered: March 2016 Location: Bristol, UK
|
Junior Member |
|
|
Hello all,
New user here and have been browsing this forum as guest for a bit.
So after distro hopping for a bit I've decided to stick with CentOS for my desktop / workstation machine. It seems very stable and while not feature rich by standard install, Nux repos add more than enough for my needs. Just some background, grew up on Amiga machines, the PC, then Mac, then I had 2 SGI workstations (IRIX is the best!), then back to Mac (OS X) and Linux. I build my own PC's so currently have a i7 4790k + z97 Gigabyte mobo with a NVidia GTX 970 card.
I've had a bit of a time installing CentOS, only because I've been trying, I think stupidly, to update the kernel to the latest kernel-LTS release. For some reason my first install did a MBR grub install when I wanted EFI / Grub2.
For some reason NVidia drivers 3.58> work OK with 3.10, but I'm experiencing some issues with my CPU cooler, a Corsair H80i and my GFX card. On anything below Kernel 3.16 the cooler's radiator fans go full speed all the time, if I upgrade the kernel to latest ML the fans work fine but my system won't complete a boot cycle. I end up with a blinking cursor with no gdm, so I'm suspecting gfx driver issues? A dmesg output on console suggests no activity after a virbr0 error. (I really don't need virtual machines), no other helpful stuff in the logs. I don't need the latest and greatest kernel, 3.16 or 4.1 would be just great.
The primary use of my machine is Autodesk Maya 2015, gimp, general standard desktop stuff and some light gaming (Guild Wars 2, WoW, Bastion etc). I considered going to Fedora but after 6 months on Arch Linux I need something a little more...hmm...stable!
PS - Has anybody found a way to improve the window movement speed on Gnome3 with Nvidia drivers? It's very slow? I previously used XFCE4 and Compton as compositor with silky smooth UI. I'm not married to Gnome3, at all....!
Thanks much,
-Half
|
|
|
Re: CentOS 7 and updates [message #3528 is a reply to message #3527] |
Tue, 15 March 2016 22:34   |
|
Hello,
Kernel-ml is not compatible with kmod-nvidia from ElRepo.
Install kernel-ml-devel, uninstall kmod-nvidia and install the Nvidia driver manually. See what kmod you have installed and download same from nvidia.com.
I have not experienced the slowness you mention, but could be related to the nvidia driver. I am using Intel GPU and with MATE anyway ...
HTH
|
|
|
|
Re: CentOS 7 and updates [message #3536 is a reply to message #3533] |
Mon, 21 March 2016 06:28   |
 |
alvaro
Messages: 13 Registered: March 2016
|
Junior Member |
|
|
Quote:Thanks for the response. I'll try the nvidia installer.
The gnome3 window drag issues seems to be common with nvidia cards. The video below shows it well.
https://youtu.be/ZfQ_BZ3W81o
I believe it has to do with the vsync method mutter employs which as far as I'm aware one cannot change.
Hi!
You have to enable ForceCompositionPipeline and/or ForceFullCompositionPipeline MetaMode.
Issue: https://devtalk.nvidia.com/default/topic/543305/linux/video- tearing-in-mplayer-on-a-gtx-600-with-vo-xv-and-vo-vdpau/
Solution:
- https://devtalk.nvidia.com/default/topic/543305/linux/screen -video-tearing-gtx6xx-7xx-kepler-9xx-maxwell-in-almost-all-a pplications-including-desktop/post/3802000/#3802000
- https://devtalk.nvidia.com/default/topic/543305/linux/screen -video-tearing-gtx6xx-7xx-kepler-9xx-maxwell-in-almost-all-a pplications-including-desktop/post/4031694/#4031694
- http://http.download.nvidia.com/XFree86/Linux-x86/325.15/REA DME/configtwinview.html
Example xorg.conf (only Screen section):
Section "Screen"
Identifier "Screen0"
Device "Device0"
Monitor "Monitor0"
DefaultDepth 24
Option "Stereo" "0"
Option "nvidiaXineramaInfoOrder" "DFP-1"
Option "metamodes" "HDMI-0: nvidia-auto-select +0+0 { ForceCompositionPipeline = On, ForceFullCompositionPipeline = On }, DVI-I-1: NULL { ForceCompositionPipeline = On, ForceFullCompositionPipeline = On }"
Option "TripleBuffer" "True"
Option "SLI" "Off"
Option "MultiGPU" "Off"
Option "BaseMosaic" "off"
SubSection "Display"
Depth 24
EndSubSection
EndSection
Example xorg.conf (only Screen section - two monitors):
Section "Screen"
Identifier "Screen0"
Device "Device0"
Monitor "Monitor0"
DefaultDepth 24
Option "Stereo" "0"
Option "nvidiaXineramaInfoOrder" "DFP-1"
Option "metamodes" "HDMI-0: nvidia-auto-select +0+0 { ForceCompositionPipeline = On, ForceFullCompositionPipeline = On }, DVI-I-1: nvidia-auto-select +1920+0 { ForceCompositionPipeline = On, ForceFullCompositionPipeline = On }"
Option "TripleBuffer" "True"
Option "SLI" "Off"
Option "MultiGPU" "Off"
Option "BaseMosaic" "off"
SubSection "Display"
Depth 24
EndSubSection
EndSection
Example solution with nvidia-settings:
/bin/nvidia-settings --assign CurrentMetaMode="HDMI-0: nvidia-auto-select { ForceCompositionPipeline = On, ForceFullCompositionPipeline = On }, DVI-I-1: NULL { ForceCompositionPipeline = On, ForceFullCompositionPipeline = On }"
# xrandr | grep " connect"
DVI-I-1 connected (normal left inverted right x axis y axis)
HDMI-0 connected primary 1920x1080+0+0 (normal left inverted right x axis y axis) 509mm x 286mm
# /bin/nvidia-settings -q CurrentMetaMode
Attribute 'CurrentMetaMode' (proton:0.0): id=50, switchable=no,
source=nv-control :: DPY-2: nvidia-auto-select @1920x1080 +0+0
{ViewPortIn=1920x1080, ViewPortOut=1920x1080+0+0}
or (two monitors):
/bin/nvidia-settings --assign CurrentMetaMode="HDMI-0: nvidia-auto-select { ForceCompositionPipeline = On, ForceFullCompositionPipeline = On }, DVI-I-1: nvidia-auto-select +1920+0 { ForceCompositionPipeline = On, ForceFullCompositionPipeline = On }"
# xrandr | grep " connect"
DVI-I-1 connected 1680x1050+1920+0 (normal left inverted right x axis y axis) 433mm x 270mm
HDMI-0 connected primary 1920x1080+0+0 (normal left inverted right x axis y axis) 509mm x 286mm
# /bin/nvidia-settings -q CurrentMetaMode
Attribute 'CurrentMetaMode' (proton:0.0): id=50, switchable=no,
source=nv-control :: DPY-2: nvidia-auto-select @1920x1080 +0+0
{ViewPortIn=1920x1080, ViewPortOut=1920x1080+0+0}, DPY-1: nvidia-auto-select
@1680x1050 +1920+0 {ViewPortIn=1680x1050, ViewPortOut=1680x1050+0+0}
Similar problem with Intel GPU: https://bugs.freedesktop.org/show_bug.cgi?id=37686
Solution for Intel GPU: Add the below code to the Device section (Section "Device"):
[Updated on: Mon, 21 March 2016 06:36] Report message to a moderator
|
|
|
Re: CentOS 7 and updates [message #3572 is a reply to message #3527] |
Wed, 04 May 2016 22:53  |
 |
schotty
Messages: 64 Registered: December 2012 Location: United States
|
Member |
|
|
I have been running RHEL for some time now, which is what CentOS is based on (in case you were unaware). One of the things that Red Hat does that is great, is backport from new kernels nifty features that improve stability and reliability. Not necessarily NEW features, FIXED features.
Although I have a few RL friends that do use the kernel-ml as you are toying with, I am very lazy. And with good reason. The standard kernels work quite well, and with elrepo's nvidia (or for that matter negativo's) packages, things tend to just work. That and Red Hat Insights has a fit if you use a non Red Hat kernel. And that's something I cannot make a rule to ignore either.
If you are worried about feature set, I wouldn't. The only thing that I had issues with was actually in the packaging department, and that was with Ardour and the Unreal engine (v4 if memory serves right) and it was due to glibc being .2 too far back. Not going there lol.
That said, it should be a rather simple matter if you are familiar with the nvidia installer to run that after any kernel updates.
|
|
|
Goto Forum:
Current Time: Sun Dec 15 05:47:25 GMT 2019
Total time taken to generate the page: 0.02306 seconds
|