I’ve been noticing screen tearing. I had installed ppa:graphics-drivers/ppa which may have caused the problems. I have used ppa-purge to remove it, and I’m back on Nvidias tested proprietary 390 driver. I’m still having problems, however. As noted in some suggestions, I have switched to Compiz. Still I have tearing. Oddly, xdpyinfo | grep DRI
shows I am using DRI2. Is there an easy way to reset my X11 configuration? I suspect something has been changed there that is causing the tearing. Otherwise, can anyone give me some ideas or a solution? Thanks all. Cheers!
To reset X for the nvidia driver you can run
nvidia-xconfig
See http://http.download.nvidia.com/XFree86/Linux-x86/1.0-8756/README/chapter-03.html
Alternatively have you tried Marco with gpu acceleration? I use that with nvidia and it works well.
1 Like
In either case, nvidia proprietary drivers don’t support dri3, which Marco uses to remove screen tearing.
Not entirely sure about Compiz, but I suspect it would be similar. Some people have reported using Compton with Marco and that working to remove tearing, but seems more like anecdata (same with xorg.conf tweaks for nvidia)
1 Like
Thank you! I’ve reinstalled the OS. Now I’m trying to stick to the Nouveau driver. I’ll see if that fares better. Also I discovered that with a different cable I can get a higher refresh rate. That may make a big difference, too. So I’ve ordered a cable. Thanks again. Cheers!
This is slightly outdated but could still work
I seem to be getting much better results with the nouveau driver and the Marco GPU window manager even though dmesg says the nouveau driver cannot load the firmware for my gtx760. Go figure. Thank you, and cheers!
Nope. With the Nouveau driver, I started locking up – three reboots in an hour. So I’m trying the nvidia 390 again. 
At least I have Compiz for zooming, again.
I wonder if it would be worth upgrading from a gtx760 to a 1050ti? It’s a newer card. Would that be more compatible with the driver? Performance wise, I think it’s a wash.
The nouveau drivers are not great. Depending on which card it may work or it may not. On a best-case scenario they can be decent, but it’s a crapshoot.
The nvidia drivers we don’t know anything about, as they are proprietary and closed-source, but they have good performance and apart from the lack of dri3 support, they are probably what you want to use to get the best out of your video card.
Some laptops have the option of disabling the nvidia card and using an internal Intel GPU. I personally use this and it gives me very reasonable performance (I use Marco) for daily use and amazing battery life. I can always enable the nvidia driver if I need it for something specific.
1 Like
Thank you. Perhaps that’s an option for me, too. I’m working out some new components, etc. I’ve just ordered a Fire TV Cube to stream 4k video; so I’ll be watching Amazon and Hulu and possibly YouTube (Amazons battles with Google not withstanding) through that device to my 4k TV. 4K is just too much for this video card. It runs super hot and locks up. I can run my desktop through a smaller high def tv. I do all my computing from my sofa. I’ll need to work out sound and switching monitors, but, so far, that’s the plan. I was hoping to watch YouTube on my 4k. That was my reason for getting it. As my eyes get worse, a large 4k screen gives me the possibility of seeing code in videos. If I watch them through the cube, however, I might be able to use the internal graphics. Thanks for the idea. Cheers!
I’m beginning to think that my “freezes” are caused by how the kernel is negotiating with ACPI.
dmesg |grep acpi
[ 0.000000] ACPI: LAPIC_NMI (acpi_id[0xff] high edge lint[0x1])
[ 0.036778] acpiphp: ACPI Hot Plug PCI Controller Driver version: 0.5
[ 0.051305] acpi PNP0A08:00: _OSC: OS supports [ExtendedConfig ASPM ClockPM Segments MSI]
[ 0.051516] acpi PNP0A08:00: _OSC: platform does not support [PCIeHotplug PME]
[ 0.051647] acpi PNP0A08:00: _OSC: OS now controls [AER PCIeCapability]
[ 0.051648] acpi PNP0A08:00: FADT indicates ASPM is unsupported, using BIOS configuration
[ 0.072074] acpiphp: Slot [1] registered
[ 0.101156] clocksource: acpi_pm: mask: 0xffffff max_cycles: 0xffffff, max_idle_ns: 2085701024 ns
Some BIOS changes later and:
dmesg |grep acpi
[ 0.000000] ACPI: LAPIC_NMI (acpi_id[0xff] high edge lint[0x1])
[ 0.036786] acpiphp: ACPI Hot Plug PCI Controller Driver version: 0.5
[ 0.055490] acpi PNP0A08:00: _OSC: OS supports [ExtendedConfig ASPM ClockPM Segments MSI]
[ 0.055998] acpi PNP0A08:00: _OSC: OS now controls [PCIeHotplug PME AER PCIeCapability]
[ 0.097042] clocksource: acpi_pm: mask: 0xffffff max_cycles: 0xffffff, max_idle_ns: 2085701024 ns
This looks much better!