Groups | Search | Server Info | Keyboard shortcuts | Login | Register [http] [https] [nntp] [nntps]

Groups > linux.debian.user > #222615 > unrolled thread

dual screen: some operations are slow (nouveau driver)

Back to article view | Back to linux.debian.user

Contents

  dual screen: some operations are slow (nouveau driver) Vincent Lefevre <vincent@vinc17.net> - 2020-05-18 16:00 +0200
    Re: dual screen: some operations are slow (nouveau driver) Vincent Lefevre <vincent@vinc17.net> - 2020-05-18 17:20 +0200
    Re: dual screen: some operations are slow (nouveau driver) Andrei POPESCU <andreimpopescu@gmail.com> - 2020-05-18 19:00 +0200
      Re: dual screen: some operations are slow (nouveau driver) Vincent Lefevre <vincent@vinc17.net> - 2020-05-18 19:20 +0200
        Re: dual screen: some operations are slow (nouveau driver) Andrei POPESCU <andreimpopescu@gmail.com> - 2020-05-18 19:50 +0200
          Re: dual screen: some operations are slow (nouveau driver) Vincent Lefevre <vincent@vinc17.net> - 2020-05-18 21:30 +0200
            Re: dual screen: some operations are slow (nouveau driver) Vincent Lefevre <vincent@vinc17.net> - 2020-05-18 21:40 +0200
          Re: dual screen: some operations are slow (nouveau driver) Vincent Lefevre <vincent@vinc17.net> - 2020-05-19 01:30 +0200

#222615 — dual screen: some operations are slow (nouveau driver)

Hi,

I have a laptop with a 3200x1800 screen, and I now have a 4K external
monitor (3840x2160). I'm currently using the following configuration
(which is the default one, at least with lightdm):

  xrandr --output DP-3 --auto --output eDP-1 --auto --right-of DP-3

The problem is that some particular operations are now slow:

  * Moving a window is very slow, except when it is small.
    My window manager is fvwm.

  * Moving N lines at a time in a terminal is slow for large terminal
    windows. I've done a test with Mutt and half-screen scrolling,
    and I could reach a factor 3 compared to the case where the
    laptop screen is off (xrandr --output eDP-1 --off).
    I've also done a test with "less" by typing "10d" and letting
    the "d" pressed to scroll 10 lines at a time; still a factor 3,
    both with xterm and GNOME Terminal.

GNOME Terminal is slower than xterm, but the ratio between 2 screens
and 1 screen remains the same under similar conditions.

I use Debian/unstable with the nouveau driver and the 5.6.0-1-amd64
kernel. Switching to the stable kernel does not seem to change
anything.

-- 
Vincent Lefèvre <vincent@vinc17.net> - Web: <https://www.vinc17.net/>
100% accessible validated (X)HTML - Blog: <https://www.vinc17.net/blog/>
Work: CR INRIA - computer arithmetic / AriC project (LIP, ENS-Lyon)

[toc] | [next] | [standalone]

#222618

Some additional information...

When I change the configuration to

  xrandr --output DP-3 --auto --output eDP-1 --auto --right-of DP-3

(usual dual screen settings), /var/log/Xorg.0.log shows:

  modeset(0): Allocate new frame buffer 7040x2160 stride

because 3840 + 3200 = 7040 and max(2160,1800) = 2160.

And when I do

  xrandr --output eDP-1 --off

to disable the laptop screen, /var/log/Xorg.0.log shows:

  modeset(0): Allocate new frame buffer 3840x2160 stride

corresponding to the external 4K monitor. I was wondering whether the
performance issue with the dual screen settings could be related to
the size of the frame buffer. So I tried

  xrandr --output DP-3 --auto --output eDP-1 --auto --same-as DP-3

in which case, the frame buffer size is still 3840x2160, and both
screens show the same content (except that the laptop screen, which
has a lower definition, does not show some bottom and right parts).
But even in this case, I notice the same performance issue.

I would add that if I switch to a Linux console (Ctrl-Alt-F1), I get
the same contents on both screens (though only 3200x1800 is used),
and everything is very fast: I can't "move windows" there, but the
scrolling tests do not show any performance issue, unlike with xorg.

Any idea?

-- 
Vincent Lefèvre <vincent@vinc17.net> - Web: <https://www.vinc17.net/>
100% accessible validated (X)HTML - Blog: <https://www.vinc17.net/blog/>
Work: CR INRIA - computer arithmetic / AriC project (LIP, ENS-Lyon)

[toc] | [prev] | [next] | [standalone]

#222620

[Multipart message — attachments visible in raw view] — view raw

On Lu, 18 mai 20, 15:57:05, Vincent Lefevre wrote:
> 
> I use Debian/unstable with the nouveau driver and the 5.6.0-1-amd64
> kernel. Switching to the stable kernel does not seem to change
> anything.

Hardware specs?

Kind regards,
Andrei
-- 
http://wiki.debian.org/FAQsFromDebianUser

[toc] | [prev] | [next] | [standalone]

#222621

On 2020-05-18 19:53:11 +0300, Andrei POPESCU wrote:
> On Lu, 18 mai 20, 15:57:05, Vincent Lefevre wrote:
> > 
> > I use Debian/unstable with the nouveau driver and the 5.6.0-1-amd64
> > kernel. Switching to the stable kernel does not seem to change
> > anything.
> 
> Hardware specs?

HP ZBook 15 G2 with 16 GB RAM and

01:00.0 VGA compatible controller [0300]: NVIDIA Corporation GK208GLM [Quadro K610M] [10de:12b9] (rev a1)

CPU and GPU related lshw information:

[...]
     *-cpu
          description: CPU
          product: Intel(R) Core(TM) i7-4810MQ CPU @ 2.80GHz
          vendor: Intel Corp.
          physical id: 9
          bus info: cpu@0
          version: Intel(R) Core(TM) i7-4810MQ CPU @ 2.80GHz
          serial: To Be Filled By O.E.M.
          slot: U3E1
          size: 1518MHz
          capacity: 3800MHz
          width: 64 bits
          clock: 100MHz
          capabilities: lm fpu fpu_exception wp vme de pse tsc msr pae mce cx8 apic sep mtrr pge mca cmov pat pse36 clflush dts acpi mmx fxsr sse sse2 ss ht tm pbe syscall nx pdpe1gb rdtscp x86-64 constant_tsc arch_perfmon pebs bts rep_good nopl xtopology nonstop_tsc cpuid aperfmperf pni pclmulqdq dtes64 monitor ds_cpl smx est tm2 ssse3 sdbg fma cx16 xtpr pdcm pcid sse4_1 sse4_2 x2apic movbe popcnt tsc_deadline_timer aes xsave avx f16c rdrand lahf_lm abm cpuid_fault epb invpcid_single pti ssbd ibrs ibpb stibp fsgsbase tsc_adjust bmi1 avx2 smep bmi2 erms invpcid xsaveopt dtherm ida arat pln pts md_clear flush_l1d cpufreq
          configuration: cores=4 enabledcores=4 threads=8
[...]
     *-pci
          description: Host bridge
          product: Xeon E3-1200 v3/4th Gen Core Processor DRAM Controller
          vendor: Intel Corporation
          physical id: 100
          bus info: pci@0000:00:00.0
          version: 06
          width: 32 bits
          clock: 33MHz
        *-pci:0
             description: PCI bridge
             product: Xeon E3-1200 v3/4th Gen Core Processor PCI Express x16 Controller
             vendor: Intel Corporation
             physical id: 1
             bus info: pci@0000:00:01.0
             version: 06
             width: 32 bits
             clock: 33MHz
             capabilities: pci pm msi pciexpress normal_decode bus_master cap_list
             configuration: driver=pcieport
             resources: irq:24 ioport:5000(size=4096) memory:d1000000-d20fffff ioport:70000000(size=301989888)
           *-display
                description: VGA compatible controller
                product: GK208GLM [Quadro K610M]
                vendor: NVIDIA Corporation
                physical id: 0
                bus info: pci@0000:01:00.0
                version: a1
                width: 64 bits
                clock: 33MHz
                capabilities: pm msi pciexpress vga_controller bus_master cap_list rom
                configuration: driver=nouveau latency=0
                resources: irq:37 memory:d1000000-d1ffffff memory:70000000-7fffffff memory:80000000-81ffffff ioport:5000(size=128) memory:c0000-dffff
           *-multimedia
                description: Audio device
                product: GK208 HDMI/DP Audio Controller
                vendor: NVIDIA Corporation
                physical id: 0.1
                bus info: pci@0000:01:00.1
                version: a1
                width: 32 bits
                clock: 33MHz
                capabilities: pm msi pciexpress bus_master cap_list
                configuration: driver=snd_hda_intel latency=0
                resources: irq:17 memory:d2000000-d2003fff
[...]

-- 
Vincent Lefèvre <vincent@vinc17.net> - Web: <https://www.vinc17.net/>
100% accessible validated (X)HTML - Blog: <https://www.vinc17.net/blog/>
Work: CR INRIA - computer arithmetic / AriC project (LIP, ENS-Lyon)

[toc] | [prev] | [next] | [standalone]

#222622

[Multipart message — attachments visible in raw view] — view raw

On Lu, 18 mai 20, 19:11:04, Vincent Lefevre wrote:
> On 2020-05-18 19:53:11 +0300, Andrei POPESCU wrote:
> > On Lu, 18 mai 20, 15:57:05, Vincent Lefevre wrote:
> > > 
> > > I use Debian/unstable with the nouveau driver and the 5.6.0-1-amd64
> > > kernel. Switching to the stable kernel does not seem to change
> > > anything.
> > 
> > Hardware specs?
> 
> HP ZBook 15 G2 with 16 GB RAM and
> 
> 01:00.0 VGA compatible controller [0300]: NVIDIA Corporation GK208GLM [Quadro K610M] [10de:12b9] (rev a1)

VRAM is also relevant.

You should probably try the Nvidia proprietary driver.

Kind regards,
Andrei
-- 
http://wiki.debian.org/FAQsFromDebianUser

[toc] | [prev] | [next] | [standalone]

#222628

On 2020-05-18 20:49:21 +0300, Andrei POPESCU wrote:
> On Lu, 18 mai 20, 19:11:04, Vincent Lefevre wrote:
> > On 2020-05-18 19:53:11 +0300, Andrei POPESCU wrote:
> > > On Lu, 18 mai 20, 15:57:05, Vincent Lefevre wrote:
> > > > 
> > > > I use Debian/unstable with the nouveau driver and the 5.6.0-1-amd64
> > > > kernel. Switching to the stable kernel does not seem to change
> > > > anything.
> > > 
> > > Hardware specs?
> > 
> > HP ZBook 15 G2 with 16 GB RAM and
> > 
> > 01:00.0 VGA compatible controller [0300]: NVIDIA Corporation GK208GLM [Quadro K610M] [10de:12b9] (rev a1)
> 
> VRAM is also relevant.

1 GB.

But note that scrolling in a full-screen Firefox is very fast.
I don't see why simpler operations would take more time.

> You should probably try the Nvidia proprietary driver.

I'll try, but I used it in the past on this machine, and it had bugs
(well, nouveau too, but less annoying).

-- 
Vincent Lefèvre <vincent@vinc17.net> - Web: <https://www.vinc17.net/>
100% accessible validated (X)HTML - Blog: <https://www.vinc17.net/blog/>
Work: CR INRIA - computer arithmetic / AriC project (LIP, ENS-Lyon)

[toc] | [prev] | [next] | [standalone]

#222629

Something interesting: using a compositor (compton) with only one
screen also makes moving windows rather slow (not as much as with
2 screens, though).

-- 
Vincent Lefèvre <vincent@vinc17.net> - Web: <https://www.vinc17.net/>
100% accessible validated (X)HTML - Blog: <https://www.vinc17.net/blog/>
Work: CR INRIA - computer arithmetic / AriC project (LIP, ENS-Lyon)

[toc] | [prev] | [next] | [standalone]

#222632

On 2020-05-18 20:49:21 +0300, Andrei POPESCU wrote:
> You should probably try the Nvidia proprietary driver.

The solves the performance issue. But
  * With light-locker, I can't blank the screen with
    "xset dpms force off". Actually it gets blanked, but there's
    something that unblank it 6 seconds later.
  * I can't control the brightness of my laptop screen.

-- 
Vincent Lefèvre <vincent@vinc17.net> - Web: <https://www.vinc17.net/>
100% accessible validated (X)HTML - Blog: <https://www.vinc17.net/blog/>
Work: CR INRIA - computer arithmetic / AriC project (LIP, ENS-Lyon)

[toc] | [prev] | [standalone]

Back to top | Article view | linux.debian.user

csiph-web